Hospital Medicine

Alcohol is the most common substance implicated in hospitalizations for substance use disorders,¹ and as a result, hospitalists commonly diagnose and manage alcohol withdrawal syndrome (AWS) in the inpatient medical setting. The 2020 guidelines of the American Society of Addiction Medicine (ASAM) provide updated recommendations for the diagnosis, monitoring, and treatment of patients hospitalized with AWS, which we have condensed to emphasize key changes from the last update² and clarify ongoing areas of uncertainty. 

Diagnosis

Recommendation 1. All inpatients who have used alcohol recently or regularly should be risk-stratified for AWS, regardless of whether or not they have suggestive symptoms (recommendations I.3, I.4, I.5, II.10). The Alcohol Use Disorders Identification Test-(Piccinelli) Consumption (AUDIT-PC) identifies patients at risk for AWS, and the Prediction of Alcohol Withdrawal Severity Scale (PAWSS) identifies those at risk for severe or complicated AWS, which includes seizures and alcohol withdrawal delirium (formerly delirium tremens). The guideline emphasizes use of these tools rather than simply initiating Clinical Institute Withdrawal Assessment for Alcohol, revised (CIWA-Ar) monitoring on all such patients to diagnose AWS, as CIWA-Ar was developed for monitoring response to treatment, not diagnosis (recommendation I.6). 

Treating Mild/Moderate or Uncomplicated AWS

Recommendation 2. Because of their proven track record of reducing the incidence of seizure and alcohol withdrawal delirium, benzodiazepines remain the recommended first-line therapy (recommendations V.13, V.16). Symptom-triggered administration of benzodiazepines (via CIWA-Ar) is recommended over fixed-dose administration because the former is associated with shorter length of stay and lower cumulative benzodiazepine administration^3,4 (recommendation V.23). Patients with mild AWS who are at low risk for severe or complicated withdrawal should be monitored for up to 36 hours for the development of worsening symptoms (recommendation V.1). For patients with high CIWA-Ar scores or who are at increased risk for severe or complicated AWS, frequent administration of moderate to high doses of a long-acting benzodiazepine early in AWS treatment (a practice called frontloading) is recommended to quickly control symptoms and prevent clinical worsening. This approach has been shown to reduce the incidence of seizures and alcohol withdrawal delirium (recommendations V.14, V.19, V.24).

Carbamazepine or gabapentin may be used in mild or moderate AWS if benzodiazepines are contraindicated; however, neither agent is recommended as first-line therapy because a clear reduction in seizure and withdrawal delirium has not been established (recommendation V.16). Alpha-2 agonists (eg, clonidine, dexmedetomidine) may be used to treat persistent autonomic hyperactivity or anxiety when these are not adequately controlled by benzodiazepines alone (recommendation V.36).

Treating Severe or Complicated AWS

Recommendation 3. The guideline defines severe AWS as withdrawal with severe signs and symptoms, and complicated AWS as withdrawal accompanied by seizures or delirium (Appendix Table⁵). The development of complications warrants prompt treatment. Patients who experience seizure should receive a fast-acting benzodiazepine (eg, intravenous [IV] diazepam or lorazepam) (recommendation VI.4). Patients with withdrawal delirium should receive a benzodiazepine (preferably parenterally) dosed to achieve light sedation. Clinicians should be prepared for the possibility that large doses may be required and to monitor patients for oversedation and respiratory depression (recommendations VI.13, VI.17). Antipsychotics may be used as adjuncts when withdrawal delirium or other symptoms, such as hallucinosis, are not adequately controlled by benzodiazepines alone, but should not be used as monotherapy (recommendation VI.20). The guideline emphasizes that alpha-2 agonists should not be used to treat withdrawal delirium (recommendation VI.21), but they may be used as adjuncts for resistant alcohol withdrawal in the intensive care unit (ICU) (recommendations VI.27, VI.29). Phenobarbital is an acceptable alternative to benzodiazepines for severe withdrawal (recommendation V.17); however, the guideline recommends that clinicians should be experienced in its use.

Treating Wernicke Encephalopathy

Recommendation 4. Thiamine should be administered to prevent Wernicke encephalopathy (WE), with parenteral formulations recommended in patients with malnutrition, severe/complicated withdrawal, or requiring ICU-level care (recommendations V.7, V.8). In particular, all patients admitted to an ICU for AWS should receive thiamine, as diagnosis of WE is often difficult in this population. Although there is no consensus on the required dose of thiamine to treat WE, 100 mg IV or intramuscularly (IM) daily for 3 to 5 days is commonly administered (recommendation V.7). Because of a lack of evidence of harm, thiamine may be given before, after, or concurrently with glucose or dextrose (recommendation V.7). The guideline does not make a specific recommendation regarding how to risk-stratify patients for WE.

Treating Underlying Alcohol Use Disorder 

Recommendation 5. Hospitalization for AWS is an important opportunity to engage patients in treatment for alcohol use disorder (AUD), including pharmacotherapy and connection with outpatient providers (recommendation V.12). The guideline emphasizes that treatment for AUD should be initiated concomitantly with AWS management whenever possible but does not make recommendations regarding specific pharmacotherapies.

This guideline was authored by a committee of emergency medicine physicians, psychiatrists, and internists using the Department of Veterans Affairs/Department of Defense guidelines and the RAND/UCLA appropriateness method to combine the scientific literature with expert opinion. The result is a series of recommendations for physicians, physician assistants, nurse practitioners, and pharmacists that are not rated by strength; an assessment of the quality of the supporting evidence is available in an appendix. Four of the nine guideline committee members reported significant financial relationships with industry and other entities relevant to these guidelines.

Despite concern about oversedation from phenobarbital raised in small case series,⁶ observational studies comparing phenobarbital with benzodiazepines suggest phenobarbital has similar efficacy for treating AWS and that oversedation is rare.^7-9 Large randomized controlled trials in this area are lacking; however, at least one small randomized controlled trial¹⁰ among patients with AWS presenting to emergency departments supports the safety and efficacy of phenobarbital when used in combination with benzodiazepines. Given the growing body of evidence supporting the safety of phenobarbital, we believe a stronger recommendation for use in patients presenting with alcohol withdrawal delirium or treatment-resistant alcohol withdrawal is warranted. The guidelines also suggest that only “experienced clinicians” use phenobarbital for AWS, which may suppress appropriate use. Nationally, phenobarbital use for AWS remains low.¹¹Finally, although the guideline recommends initiation of treatment for AUD, specific recommendations for pharmacotherapy are not provided. Three medications currently have approval from the US Food and Drug Administration for treatment of AUD: acamprosate, naltrexone, and disulfiram. Large randomized controlled trials support the safety and efficacy of acamprosate and naltrexone, with or without counselling, in the treatment of AUD,¹² and disulfiram may be appropriate for selected highly motivated patients. We believe more specific recommendations to assist in choosing among these options would be useful.

More data are needed on the safety and efficacy of phenobarbital in patients with AWS, as well as comparative effectiveness against benzodiazepines. Recruitment is ongoing for a single clinical trial comparing the effect of phenobarbital and lorazepam on length of stay among patients in the ICU with AWS (NCT04156464); to date, no randomized trials of phenobarbital have been conducted in medical inpatients with AWS. In addition, gaps in the literature exist regarding benzodiazepine selection, and head-to-head comparisons of symptom-triggered usage of different benzodiazepines are lacking.

Alcohol is the most common substance implicated in hospitalizations for substance use disorders,¹ and as a result, hospitalists commonly diagnose and manage alcohol withdrawal syndrome (AWS) in the inpatient medical setting. The 2020 guidelines of the American Society of Addiction Medicine (ASAM) provide updated recommendations for the diagnosis, monitoring, and treatment of patients hospitalized with AWS, which we have condensed to emphasize key changes from the last update² and clarify ongoing areas of uncertainty. 

Diagnosis

Recommendation 1. All inpatients who have used alcohol recently or regularly should be risk-stratified for AWS, regardless of whether or not they have suggestive symptoms (recommendations I.3, I.4, I.5, II.10). The Alcohol Use Disorders Identification Test-(Piccinelli) Consumption (AUDIT-PC) identifies patients at risk for AWS, and the Prediction of Alcohol Withdrawal Severity Scale (PAWSS) identifies those at risk for severe or complicated AWS, which includes seizures and alcohol withdrawal delirium (formerly delirium tremens). The guideline emphasizes use of these tools rather than simply initiating Clinical Institute Withdrawal Assessment for Alcohol, revised (CIWA-Ar) monitoring on all such patients to diagnose AWS, as CIWA-Ar was developed for monitoring response to treatment, not diagnosis (recommendation I.6). 

Treating Mild/Moderate or Uncomplicated AWS

Recommendation 2. Because of their proven track record of reducing the incidence of seizure and alcohol withdrawal delirium, benzodiazepines remain the recommended first-line therapy (recommendations V.13, V.16). Symptom-triggered administration of benzodiazepines (via CIWA-Ar) is recommended over fixed-dose administration because the former is associated with shorter length of stay and lower cumulative benzodiazepine administration^3,4 (recommendation V.23). Patients with mild AWS who are at low risk for severe or complicated withdrawal should be monitored for up to 36 hours for the development of worsening symptoms (recommendation V.1). For patients with high CIWA-Ar scores or who are at increased risk for severe or complicated AWS, frequent administration of moderate to high doses of a long-acting benzodiazepine early in AWS treatment (a practice called frontloading) is recommended to quickly control symptoms and prevent clinical worsening. This approach has been shown to reduce the incidence of seizures and alcohol withdrawal delirium (recommendations V.14, V.19, V.24).

Carbamazepine or gabapentin may be used in mild or moderate AWS if benzodiazepines are contraindicated; however, neither agent is recommended as first-line therapy because a clear reduction in seizure and withdrawal delirium has not been established (recommendation V.16). Alpha-2 agonists (eg, clonidine, dexmedetomidine) may be used to treat persistent autonomic hyperactivity or anxiety when these are not adequately controlled by benzodiazepines alone (recommendation V.36).

Treating Severe or Complicated AWS

Recommendation 3. The guideline defines severe AWS as withdrawal with severe signs and symptoms, and complicated AWS as withdrawal accompanied by seizures or delirium (Appendix Table⁵). The development of complications warrants prompt treatment. Patients who experience seizure should receive a fast-acting benzodiazepine (eg, intravenous [IV] diazepam or lorazepam) (recommendation VI.4). Patients with withdrawal delirium should receive a benzodiazepine (preferably parenterally) dosed to achieve light sedation. Clinicians should be prepared for the possibility that large doses may be required and to monitor patients for oversedation and respiratory depression (recommendations VI.13, VI.17). Antipsychotics may be used as adjuncts when withdrawal delirium or other symptoms, such as hallucinosis, are not adequately controlled by benzodiazepines alone, but should not be used as monotherapy (recommendation VI.20). The guideline emphasizes that alpha-2 agonists should not be used to treat withdrawal delirium (recommendation VI.21), but they may be used as adjuncts for resistant alcohol withdrawal in the intensive care unit (ICU) (recommendations VI.27, VI.29). Phenobarbital is an acceptable alternative to benzodiazepines for severe withdrawal (recommendation V.17); however, the guideline recommends that clinicians should be experienced in its use.

Treating Wernicke Encephalopathy

Recommendation 4. Thiamine should be administered to prevent Wernicke encephalopathy (WE), with parenteral formulations recommended in patients with malnutrition, severe/complicated withdrawal, or requiring ICU-level care (recommendations V.7, V.8). In particular, all patients admitted to an ICU for AWS should receive thiamine, as diagnosis of WE is often difficult in this population. Although there is no consensus on the required dose of thiamine to treat WE, 100 mg IV or intramuscularly (IM) daily for 3 to 5 days is commonly administered (recommendation V.7). Because of a lack of evidence of harm, thiamine may be given before, after, or concurrently with glucose or dextrose (recommendation V.7). The guideline does not make a specific recommendation regarding how to risk-stratify patients for WE.

Treating Underlying Alcohol Use Disorder 

Recommendation 5. Hospitalization for AWS is an important opportunity to engage patients in treatment for alcohol use disorder (AUD), including pharmacotherapy and connection with outpatient providers (recommendation V.12). The guideline emphasizes that treatment for AUD should be initiated concomitantly with AWS management whenever possible but does not make recommendations regarding specific pharmacotherapies.

This guideline was authored by a committee of emergency medicine physicians, psychiatrists, and internists using the Department of Veterans Affairs/Department of Defense guidelines and the RAND/UCLA appropriateness method to combine the scientific literature with expert opinion. The result is a series of recommendations for physicians, physician assistants, nurse practitioners, and pharmacists that are not rated by strength; an assessment of the quality of the supporting evidence is available in an appendix. Four of the nine guideline committee members reported significant financial relationships with industry and other entities relevant to these guidelines.

Despite concern about oversedation from phenobarbital raised in small case series,⁶ observational studies comparing phenobarbital with benzodiazepines suggest phenobarbital has similar efficacy for treating AWS and that oversedation is rare.^7-9 Large randomized controlled trials in this area are lacking; however, at least one small randomized controlled trial¹⁰ among patients with AWS presenting to emergency departments supports the safety and efficacy of phenobarbital when used in combination with benzodiazepines. Given the growing body of evidence supporting the safety of phenobarbital, we believe a stronger recommendation for use in patients presenting with alcohol withdrawal delirium or treatment-resistant alcohol withdrawal is warranted. The guidelines also suggest that only “experienced clinicians” use phenobarbital for AWS, which may suppress appropriate use. Nationally, phenobarbital use for AWS remains low.¹¹Finally, although the guideline recommends initiation of treatment for AUD, specific recommendations for pharmacotherapy are not provided. Three medications currently have approval from the US Food and Drug Administration for treatment of AUD: acamprosate, naltrexone, and disulfiram. Large randomized controlled trials support the safety and efficacy of acamprosate and naltrexone, with or without counselling, in the treatment of AUD,¹² and disulfiram may be appropriate for selected highly motivated patients. We believe more specific recommendations to assist in choosing among these options would be useful.

More data are needed on the safety and efficacy of phenobarbital in patients with AWS, as well as comparative effectiveness against benzodiazepines. Recruitment is ongoing for a single clinical trial comparing the effect of phenobarbital and lorazepam on length of stay among patients in the ICU with AWS (NCT04156464); to date, no randomized trials of phenobarbital have been conducted in medical inpatients with AWS. In addition, gaps in the literature exist regarding benzodiazepine selection, and head-to-head comparisons of symptom-triggered usage of different benzodiazepines are lacking.

Alcohol is the most common substance implicated in hospitalizations for substance use disorders,¹ and as a result, hospitalists commonly diagnose and manage alcohol withdrawal syndrome (AWS) in the inpatient medical setting. The 2020 guidelines of the American Society of Addiction Medicine (ASAM) provide updated recommendations for the diagnosis, monitoring, and treatment of patients hospitalized with AWS, which we have condensed to emphasize key changes from the last update² and clarify ongoing areas of uncertainty. 

Diagnosis

Recommendation 1. All inpatients who have used alcohol recently or regularly should be risk-stratified for AWS, regardless of whether or not they have suggestive symptoms (recommendations I.3, I.4, I.5, II.10). The Alcohol Use Disorders Identification Test-(Piccinelli) Consumption (AUDIT-PC) identifies patients at risk for AWS, and the Prediction of Alcohol Withdrawal Severity Scale (PAWSS) identifies those at risk for severe or complicated AWS, which includes seizures and alcohol withdrawal delirium (formerly delirium tremens). The guideline emphasizes use of these tools rather than simply initiating Clinical Institute Withdrawal Assessment for Alcohol, revised (CIWA-Ar) monitoring on all such patients to diagnose AWS, as CIWA-Ar was developed for monitoring response to treatment, not diagnosis (recommendation I.6). 

Treating Mild/Moderate or Uncomplicated AWS

Recommendation 2. Because of their proven track record of reducing the incidence of seizure and alcohol withdrawal delirium, benzodiazepines remain the recommended first-line therapy (recommendations V.13, V.16). Symptom-triggered administration of benzodiazepines (via CIWA-Ar) is recommended over fixed-dose administration because the former is associated with shorter length of stay and lower cumulative benzodiazepine administration^3,4 (recommendation V.23). Patients with mild AWS who are at low risk for severe or complicated withdrawal should be monitored for up to 36 hours for the development of worsening symptoms (recommendation V.1). For patients with high CIWA-Ar scores or who are at increased risk for severe or complicated AWS, frequent administration of moderate to high doses of a long-acting benzodiazepine early in AWS treatment (a practice called frontloading) is recommended to quickly control symptoms and prevent clinical worsening. This approach has been shown to reduce the incidence of seizures and alcohol withdrawal delirium (recommendations V.14, V.19, V.24).

Carbamazepine or gabapentin may be used in mild or moderate AWS if benzodiazepines are contraindicated; however, neither agent is recommended as first-line therapy because a clear reduction in seizure and withdrawal delirium has not been established (recommendation V.16). Alpha-2 agonists (eg, clonidine, dexmedetomidine) may be used to treat persistent autonomic hyperactivity or anxiety when these are not adequately controlled by benzodiazepines alone (recommendation V.36).

Treating Severe or Complicated AWS

Recommendation 3. The guideline defines severe AWS as withdrawal with severe signs and symptoms, and complicated AWS as withdrawal accompanied by seizures or delirium (Appendix Table⁵). The development of complications warrants prompt treatment. Patients who experience seizure should receive a fast-acting benzodiazepine (eg, intravenous [IV] diazepam or lorazepam) (recommendation VI.4). Patients with withdrawal delirium should receive a benzodiazepine (preferably parenterally) dosed to achieve light sedation. Clinicians should be prepared for the possibility that large doses may be required and to monitor patients for oversedation and respiratory depression (recommendations VI.13, VI.17). Antipsychotics may be used as adjuncts when withdrawal delirium or other symptoms, such as hallucinosis, are not adequately controlled by benzodiazepines alone, but should not be used as monotherapy (recommendation VI.20). The guideline emphasizes that alpha-2 agonists should not be used to treat withdrawal delirium (recommendation VI.21), but they may be used as adjuncts for resistant alcohol withdrawal in the intensive care unit (ICU) (recommendations VI.27, VI.29). Phenobarbital is an acceptable alternative to benzodiazepines for severe withdrawal (recommendation V.17); however, the guideline recommends that clinicians should be experienced in its use.

Treating Wernicke Encephalopathy

Recommendation 4. Thiamine should be administered to prevent Wernicke encephalopathy (WE), with parenteral formulations recommended in patients with malnutrition, severe/complicated withdrawal, or requiring ICU-level care (recommendations V.7, V.8). In particular, all patients admitted to an ICU for AWS should receive thiamine, as diagnosis of WE is often difficult in this population. Although there is no consensus on the required dose of thiamine to treat WE, 100 mg IV or intramuscularly (IM) daily for 3 to 5 days is commonly administered (recommendation V.7). Because of a lack of evidence of harm, thiamine may be given before, after, or concurrently with glucose or dextrose (recommendation V.7). The guideline does not make a specific recommendation regarding how to risk-stratify patients for WE.

Treating Underlying Alcohol Use Disorder 

Recommendation 5. Hospitalization for AWS is an important opportunity to engage patients in treatment for alcohol use disorder (AUD), including pharmacotherapy and connection with outpatient providers (recommendation V.12). The guideline emphasizes that treatment for AUD should be initiated concomitantly with AWS management whenever possible but does not make recommendations regarding specific pharmacotherapies.

This guideline was authored by a committee of emergency medicine physicians, psychiatrists, and internists using the Department of Veterans Affairs/Department of Defense guidelines and the RAND/UCLA appropriateness method to combine the scientific literature with expert opinion. The result is a series of recommendations for physicians, physician assistants, nurse practitioners, and pharmacists that are not rated by strength; an assessment of the quality of the supporting evidence is available in an appendix. Four of the nine guideline committee members reported significant financial relationships with industry and other entities relevant to these guidelines.

Despite concern about oversedation from phenobarbital raised in small case series,⁶ observational studies comparing phenobarbital with benzodiazepines suggest phenobarbital has similar efficacy for treating AWS and that oversedation is rare.^7-9 Large randomized controlled trials in this area are lacking; however, at least one small randomized controlled trial¹⁰ among patients with AWS presenting to emergency departments supports the safety and efficacy of phenobarbital when used in combination with benzodiazepines. Given the growing body of evidence supporting the safety of phenobarbital, we believe a stronger recommendation for use in patients presenting with alcohol withdrawal delirium or treatment-resistant alcohol withdrawal is warranted. The guidelines also suggest that only “experienced clinicians” use phenobarbital for AWS, which may suppress appropriate use. Nationally, phenobarbital use for AWS remains low.¹¹Finally, although the guideline recommends initiation of treatment for AUD, specific recommendations for pharmacotherapy are not provided. Three medications currently have approval from the US Food and Drug Administration for treatment of AUD: acamprosate, naltrexone, and disulfiram. Large randomized controlled trials support the safety and efficacy of acamprosate and naltrexone, with or without counselling, in the treatment of AUD,¹² and disulfiram may be appropriate for selected highly motivated patients. We believe more specific recommendations to assist in choosing among these options would be useful.

More data are needed on the safety and efficacy of phenobarbital in patients with AWS, as well as comparative effectiveness against benzodiazepines. Recruitment is ongoing for a single clinical trial comparing the effect of phenobarbital and lorazepam on length of stay among patients in the ICU with AWS (NCT04156464); to date, no randomized trials of phenobarbital have been conducted in medical inpatients with AWS. In addition, gaps in the literature exist regarding benzodiazepine selection, and head-to-head comparisons of symptom-triggered usage of different benzodiazepines are lacking.

Narrative competence is the ability to acquire, interpret, and act on the stories of others.¹ Developing this skill through guided medical storytelling can improve health care practitioners’ (HCPs) sense of empathy and satisfaction with their work.² Narrative medicine experiences for medical students can foster a deeper understanding of their patients beyond illness-associated identities.³

Within narrative medicine, the “life story” is a specific technique that allows patients to share experiences through open-ended interviews that are entered into the health record.^4,5 By sharing life stories, patients control a narrative encompassing more than their illness and can reinforce a sense of purpose in their lives.⁶ The US Department of Veterans Affairs (VA) My Life My Story (MLMS) program gives veterans the opportunity to share their narrative with staff and volunteer interviewers. MLMS is well received by veterans, has durable positive effects for HCPs who read the stories, and has been used as a tool to teach patient-centered care to medical trainees.^7-9

We created a narrative medicine curriculum at the San Francisco VA Medical Center (SFVAMC) in which medical students interviewed veterans for the MLMS program. Medical students initially collected life stories through in-person conversation. During the COVID-19 pandemic, physical distancing regulations limited direct patient interaction for students and prompted a switch to phone and video interviews. This shift paralleled the widespread adoption of telehealth, which will persist beyond the pandemic and require teachers and learners to develop competency in forming personal connections with patients through videoconferencing.^10,11

There are no published studies describing how to guide medical students (or other historians) in generating life stories without in-person patient contact. This article details the design of a medical student curriculum incorporating MLMS and the transition to remote interaction between instructors, students, and veterans during the early COVID-19 pandemic.

MLMS Program Origins

The MLMS project began at the William S. Middleton Memorial Veterans Hospital in Madison, Wisconsin, in 2013 with staff and volunteer interviewers and has expanded to more than 60 VA facilities.⁷ In January 2020, we initiated a narrative medicine curriculum incorporating MLMS at the SFVAMC as a required component of a third-year internal medicine clerkship for medical students at the University of California San Francisco (UCSF). Fifty-four medical students in 10 cohorts participated in the curriculum in 2020. The primary program objectives were for medical students to develop skills for eliciting and recording a life story and to appreciate the impact of this activity on a veteran’s experience of receiving health care. Secondary objectives were for students to understand the mission of the VA health care system and veteran demographics.

The first cohort of 6 UCSF medical students participated in MLMS during their 8-week VA clerkship. Students attended a 1-hour small group session to introduce the program and build narrative medicine skills. Preparation for this session involved listening to 2 podcast episodes introducing the VA health care system and MLMS.^12,13 The session began with a short interactive discussion of veteran demographics with an emphasis on addressing assumptions students might have about the veteran population. Students were taught strategies for engaging in open-ended conversations without emphasizing illness. Each student practiced collecting a life story with a simulated patient portrayed by an instructor and received feedback from classmates and instructors.

Over the following weeks, students selected a hospitalized veteran, typically a patient they were caring for, introduced MLMS, and obtained verbal consent to participate. They conducted a 60- to 90-minute interview, wrote and organized the life story, read it to the veteran, and solicited edits. Once a final version was generated, the student provided the veteran with printed copies and offered to place the story in the Computerized Patient Record System (CPRS).

Near the end of their rotation, students attended a 1-hour small group session in which they shared reflections on the experience of collecting a life story, the impact of veterans’ life experiences on their health and illness, and moments when students confronted their own stereotypes and implicit biases. Students then reviewed narrative medicine skills that are generalizable to all patient interactions.

COVID-19-Related Adaptation

In March 2020, shortly after the second student cohort began, medical students were removed from the clinical setting in response to the COVID-19 pandemic. The 8-week clerkship was converted to a 3-week remote learning rotation. The MLMS experience was preserved by converting small group sessions to videoconferences and expanding the pool of eligible patients to include veterans who students had met on prior rotations, current inpatients, and outpatients from VA primary care clinics. Students contacted veterans after an instructor had introduced MLMS to the veteran and confirmed that the veteran was interested in participating.

Students in the second and third cohorts completed a telephone-based iteration of MLMS in which interviews and life story reviews were conducted over the telephone and printed copies mailed to the veteran. For the fourth, fifth, and sixth cohorts, MLMS was transitioned to a video-based program with inpatients. Instructors collaborated with a volunteer group supplying tablet devices to inpatients to make video calls to their families during the pandemic.¹⁴ Clerkship students coordinated with that volunteer group to interview veterans and review their stories through the tablet devices.

From July to December 2020 medical students returned to 4-week on-site clinical rotations at the SFVAMC. The program returned to the original format for cohorts 7 to 10, with students attending in-person small group sessions and conducting in-person interviews with inpatients.

Curriculum Evaluation

Students completed surveys in the week after the curriculum concluded. Survey completion was voluntary, anonymous, and had no bearing on their evaluation or grade (pass/fail only). Likert scale questions (1, strongly disagree; 5, strongly agree) were used to assess the program (eAppendix 1). One-way analysis of variance testing was used to compare means stratified by method of interview (in person, telephone, or video). Surveys also included free-response questions asking students to highlight aspects of the program they valued or would change; responses were summarized by theme. This program evaluation was deemed exempt from review by the UCSF Human Research Protection Program Institutional Review Board.

Sixty-two veteran stories were collected by 54 participating students (one student was unable to complete an interview, while several students completed multiple interviews). Fifty-four (87%) veterans requested their stories be entered into the medical record.

All 54 students completed the survey. Students reported that the MLMS curriculum helped them develop new skills for eliciting and recording a life story (mean [SD] 4.5 [0.7]). Most students strongly agreed that MLMS helped them understand how sharing a life story can impact a veteran’s experience of receiving health care, with a mean (SD) score of 4.8 (0.4). After completing MLMS, students also reported a better understanding of the mission of the VA and veteran demographics with a mean (SD) score of 4.4 (0.7) and 4.3 (0.7), respectively. Stratification of survey responses by method of interview (in person, telephone, or video) revealed no statistically significant differences in evaluations (Table 1).

Discussion

Previous research has demonstrated that a narrative medicine curriculum can help medicine clerkship students develop narrative competence through patient storytelling with a focus on a patient’s illness narrative.¹⁵ The VA MLMS program extends the patient narrative beyond health care–related experiences and encompasses their broader life story. This article adds to the MLMS and narrative medicine literature by demonstrating that the efficacy of teaching patient-centered care to medical trainees through direct interviews can be maintained in remote formats.⁹ The article also provides guidance for MLMS programs that wish to conduct remote veteran interviews.

The widespread adoption of telemedicine will require trainees to develop communication skills to establish therapeutic relationships with patients both face-to-face and through videoconferencing. In order to promote this important skill across varying levels of physical distancing, narrative medicine programs should be adaptable to a virtual learning environment. As we redesigned MLMS for the remote setting, we learned several key lessons that can guide similar curricular and programmatic innovations at other institutions. For example, videoconferencing created stronger connections between the students and veterans than telephone calls. However, tablet-based video interviews also introduced many technological challenges and required on-site personnel (nurses and volunteers) to connect students, veterans, and technology. Solutions for technology and communication challenges related to the basic personnel and infrastructure needed to start and maintain a remote MLMS program are outlined in Table 2.

Limitations

Limitations of this study include the participation of trainees from a single institution and a lack of assessment of the impact of MLMS on veterans. Future research could assess whether life story skills and practices are maintained after the medicine clerkship. In addition, future studies could examine veterans’ perspectives through interviews with qualitative analysis to learn how MLMS affected their experience of receiving health care.

Conclusions

This is the first report of a remote-capable life story curriculum for medical students. Shifting to a virtual MLMS curriculum requires protocols and people to link interviewers, veterans, and technology. Training for in-person interactions while being prepared for remote interviewing is essential to ensure that the MLMS experience remains available to interviewers and veterans who otherwise may never have the chance to connect. The restrictions and isolation of the COVID-19 pandemic will fade, but using MLMS to virtually connect patients, providers, and students will remain an important capability and opportunity as health care shifts to more virtual interaction.

Acknowledgments

The authors thank Emma Levine, MD, for her assistance coordinating video interviews; Thor Ringler, MS, MFA, for his assistance with manuscript review; and the veterans of the San Francisco VA Health Care System for sharing their stories.

Narrative competence is the ability to acquire, interpret, and act on the stories of others.¹ Developing this skill through guided medical storytelling can improve health care practitioners’ (HCPs) sense of empathy and satisfaction with their work.² Narrative medicine experiences for medical students can foster a deeper understanding of their patients beyond illness-associated identities.³

Within narrative medicine, the “life story” is a specific technique that allows patients to share experiences through open-ended interviews that are entered into the health record.^4,5 By sharing life stories, patients control a narrative encompassing more than their illness and can reinforce a sense of purpose in their lives.⁶ The US Department of Veterans Affairs (VA) My Life My Story (MLMS) program gives veterans the opportunity to share their narrative with staff and volunteer interviewers. MLMS is well received by veterans, has durable positive effects for HCPs who read the stories, and has been used as a tool to teach patient-centered care to medical trainees.^7-9

We created a narrative medicine curriculum at the San Francisco VA Medical Center (SFVAMC) in which medical students interviewed veterans for the MLMS program. Medical students initially collected life stories through in-person conversation. During the COVID-19 pandemic, physical distancing regulations limited direct patient interaction for students and prompted a switch to phone and video interviews. This shift paralleled the widespread adoption of telehealth, which will persist beyond the pandemic and require teachers and learners to develop competency in forming personal connections with patients through videoconferencing.^10,11

There are no published studies describing how to guide medical students (or other historians) in generating life stories without in-person patient contact. This article details the design of a medical student curriculum incorporating MLMS and the transition to remote interaction between instructors, students, and veterans during the early COVID-19 pandemic.

MLMS Program Origins

The MLMS project began at the William S. Middleton Memorial Veterans Hospital in Madison, Wisconsin, in 2013 with staff and volunteer interviewers and has expanded to more than 60 VA facilities.⁷ In January 2020, we initiated a narrative medicine curriculum incorporating MLMS at the SFVAMC as a required component of a third-year internal medicine clerkship for medical students at the University of California San Francisco (UCSF). Fifty-four medical students in 10 cohorts participated in the curriculum in 2020. The primary program objectives were for medical students to develop skills for eliciting and recording a life story and to appreciate the impact of this activity on a veteran’s experience of receiving health care. Secondary objectives were for students to understand the mission of the VA health care system and veteran demographics.

The first cohort of 6 UCSF medical students participated in MLMS during their 8-week VA clerkship. Students attended a 1-hour small group session to introduce the program and build narrative medicine skills. Preparation for this session involved listening to 2 podcast episodes introducing the VA health care system and MLMS.^12,13 The session began with a short interactive discussion of veteran demographics with an emphasis on addressing assumptions students might have about the veteran population. Students were taught strategies for engaging in open-ended conversations without emphasizing illness. Each student practiced collecting a life story with a simulated patient portrayed by an instructor and received feedback from classmates and instructors.

Over the following weeks, students selected a hospitalized veteran, typically a patient they were caring for, introduced MLMS, and obtained verbal consent to participate. They conducted a 60- to 90-minute interview, wrote and organized the life story, read it to the veteran, and solicited edits. Once a final version was generated, the student provided the veteran with printed copies and offered to place the story in the Computerized Patient Record System (CPRS).

Near the end of their rotation, students attended a 1-hour small group session in which they shared reflections on the experience of collecting a life story, the impact of veterans’ life experiences on their health and illness, and moments when students confronted their own stereotypes and implicit biases. Students then reviewed narrative medicine skills that are generalizable to all patient interactions.

COVID-19-Related Adaptation

In March 2020, shortly after the second student cohort began, medical students were removed from the clinical setting in response to the COVID-19 pandemic. The 8-week clerkship was converted to a 3-week remote learning rotation. The MLMS experience was preserved by converting small group sessions to videoconferences and expanding the pool of eligible patients to include veterans who students had met on prior rotations, current inpatients, and outpatients from VA primary care clinics. Students contacted veterans after an instructor had introduced MLMS to the veteran and confirmed that the veteran was interested in participating.

Students in the second and third cohorts completed a telephone-based iteration of MLMS in which interviews and life story reviews were conducted over the telephone and printed copies mailed to the veteran. For the fourth, fifth, and sixth cohorts, MLMS was transitioned to a video-based program with inpatients. Instructors collaborated with a volunteer group supplying tablet devices to inpatients to make video calls to their families during the pandemic.¹⁴ Clerkship students coordinated with that volunteer group to interview veterans and review their stories through the tablet devices.

From July to December 2020 medical students returned to 4-week on-site clinical rotations at the SFVAMC. The program returned to the original format for cohorts 7 to 10, with students attending in-person small group sessions and conducting in-person interviews with inpatients.

Curriculum Evaluation

Students completed surveys in the week after the curriculum concluded. Survey completion was voluntary, anonymous, and had no bearing on their evaluation or grade (pass/fail only). Likert scale questions (1, strongly disagree; 5, strongly agree) were used to assess the program (eAppendix 1). One-way analysis of variance testing was used to compare means stratified by method of interview (in person, telephone, or video). Surveys also included free-response questions asking students to highlight aspects of the program they valued or would change; responses were summarized by theme. This program evaluation was deemed exempt from review by the UCSF Human Research Protection Program Institutional Review Board.

Sixty-two veteran stories were collected by 54 participating students (one student was unable to complete an interview, while several students completed multiple interviews). Fifty-four (87%) veterans requested their stories be entered into the medical record.

All 54 students completed the survey. Students reported that the MLMS curriculum helped them develop new skills for eliciting and recording a life story (mean [SD] 4.5 [0.7]). Most students strongly agreed that MLMS helped them understand how sharing a life story can impact a veteran’s experience of receiving health care, with a mean (SD) score of 4.8 (0.4). After completing MLMS, students also reported a better understanding of the mission of the VA and veteran demographics with a mean (SD) score of 4.4 (0.7) and 4.3 (0.7), respectively. Stratification of survey responses by method of interview (in person, telephone, or video) revealed no statistically significant differences in evaluations (Table 1).

Discussion

Previous research has demonstrated that a narrative medicine curriculum can help medicine clerkship students develop narrative competence through patient storytelling with a focus on a patient’s illness narrative.¹⁵ The VA MLMS program extends the patient narrative beyond health care–related experiences and encompasses their broader life story. This article adds to the MLMS and narrative medicine literature by demonstrating that the efficacy of teaching patient-centered care to medical trainees through direct interviews can be maintained in remote formats.⁹ The article also provides guidance for MLMS programs that wish to conduct remote veteran interviews.

The widespread adoption of telemedicine will require trainees to develop communication skills to establish therapeutic relationships with patients both face-to-face and through videoconferencing. In order to promote this important skill across varying levels of physical distancing, narrative medicine programs should be adaptable to a virtual learning environment. As we redesigned MLMS for the remote setting, we learned several key lessons that can guide similar curricular and programmatic innovations at other institutions. For example, videoconferencing created stronger connections between the students and veterans than telephone calls. However, tablet-based video interviews also introduced many technological challenges and required on-site personnel (nurses and volunteers) to connect students, veterans, and technology. Solutions for technology and communication challenges related to the basic personnel and infrastructure needed to start and maintain a remote MLMS program are outlined in Table 2.

Limitations

Limitations of this study include the participation of trainees from a single institution and a lack of assessment of the impact of MLMS on veterans. Future research could assess whether life story skills and practices are maintained after the medicine clerkship. In addition, future studies could examine veterans’ perspectives through interviews with qualitative analysis to learn how MLMS affected their experience of receiving health care.

Conclusions

This is the first report of a remote-capable life story curriculum for medical students. Shifting to a virtual MLMS curriculum requires protocols and people to link interviewers, veterans, and technology. Training for in-person interactions while being prepared for remote interviewing is essential to ensure that the MLMS experience remains available to interviewers and veterans who otherwise may never have the chance to connect. The restrictions and isolation of the COVID-19 pandemic will fade, but using MLMS to virtually connect patients, providers, and students will remain an important capability and opportunity as health care shifts to more virtual interaction.

Acknowledgments

The authors thank Emma Levine, MD, for her assistance coordinating video interviews; Thor Ringler, MS, MFA, for his assistance with manuscript review; and the veterans of the San Francisco VA Health Care System for sharing their stories.

Narrative competence is the ability to acquire, interpret, and act on the stories of others.¹ Developing this skill through guided medical storytelling can improve health care practitioners’ (HCPs) sense of empathy and satisfaction with their work.² Narrative medicine experiences for medical students can foster a deeper understanding of their patients beyond illness-associated identities.³

Within narrative medicine, the “life story” is a specific technique that allows patients to share experiences through open-ended interviews that are entered into the health record.^4,5 By sharing life stories, patients control a narrative encompassing more than their illness and can reinforce a sense of purpose in their lives.⁶ The US Department of Veterans Affairs (VA) My Life My Story (MLMS) program gives veterans the opportunity to share their narrative with staff and volunteer interviewers. MLMS is well received by veterans, has durable positive effects for HCPs who read the stories, and has been used as a tool to teach patient-centered care to medical trainees.^7-9

We created a narrative medicine curriculum at the San Francisco VA Medical Center (SFVAMC) in which medical students interviewed veterans for the MLMS program. Medical students initially collected life stories through in-person conversation. During the COVID-19 pandemic, physical distancing regulations limited direct patient interaction for students and prompted a switch to phone and video interviews. This shift paralleled the widespread adoption of telehealth, which will persist beyond the pandemic and require teachers and learners to develop competency in forming personal connections with patients through videoconferencing.^10,11

There are no published studies describing how to guide medical students (or other historians) in generating life stories without in-person patient contact. This article details the design of a medical student curriculum incorporating MLMS and the transition to remote interaction between instructors, students, and veterans during the early COVID-19 pandemic.

MLMS Program Origins

The MLMS project began at the William S. Middleton Memorial Veterans Hospital in Madison, Wisconsin, in 2013 with staff and volunteer interviewers and has expanded to more than 60 VA facilities.⁷ In January 2020, we initiated a narrative medicine curriculum incorporating MLMS at the SFVAMC as a required component of a third-year internal medicine clerkship for medical students at the University of California San Francisco (UCSF). Fifty-four medical students in 10 cohorts participated in the curriculum in 2020. The primary program objectives were for medical students to develop skills for eliciting and recording a life story and to appreciate the impact of this activity on a veteran’s experience of receiving health care. Secondary objectives were for students to understand the mission of the VA health care system and veteran demographics.

The first cohort of 6 UCSF medical students participated in MLMS during their 8-week VA clerkship. Students attended a 1-hour small group session to introduce the program and build narrative medicine skills. Preparation for this session involved listening to 2 podcast episodes introducing the VA health care system and MLMS.^12,13 The session began with a short interactive discussion of veteran demographics with an emphasis on addressing assumptions students might have about the veteran population. Students were taught strategies for engaging in open-ended conversations without emphasizing illness. Each student practiced collecting a life story with a simulated patient portrayed by an instructor and received feedback from classmates and instructors.

Over the following weeks, students selected a hospitalized veteran, typically a patient they were caring for, introduced MLMS, and obtained verbal consent to participate. They conducted a 60- to 90-minute interview, wrote and organized the life story, read it to the veteran, and solicited edits. Once a final version was generated, the student provided the veteran with printed copies and offered to place the story in the Computerized Patient Record System (CPRS).

Near the end of their rotation, students attended a 1-hour small group session in which they shared reflections on the experience of collecting a life story, the impact of veterans’ life experiences on their health and illness, and moments when students confronted their own stereotypes and implicit biases. Students then reviewed narrative medicine skills that are generalizable to all patient interactions.

COVID-19-Related Adaptation

In March 2020, shortly after the second student cohort began, medical students were removed from the clinical setting in response to the COVID-19 pandemic. The 8-week clerkship was converted to a 3-week remote learning rotation. The MLMS experience was preserved by converting small group sessions to videoconferences and expanding the pool of eligible patients to include veterans who students had met on prior rotations, current inpatients, and outpatients from VA primary care clinics. Students contacted veterans after an instructor had introduced MLMS to the veteran and confirmed that the veteran was interested in participating.

Students in the second and third cohorts completed a telephone-based iteration of MLMS in which interviews and life story reviews were conducted over the telephone and printed copies mailed to the veteran. For the fourth, fifth, and sixth cohorts, MLMS was transitioned to a video-based program with inpatients. Instructors collaborated with a volunteer group supplying tablet devices to inpatients to make video calls to their families during the pandemic.¹⁴ Clerkship students coordinated with that volunteer group to interview veterans and review their stories through the tablet devices.

From July to December 2020 medical students returned to 4-week on-site clinical rotations at the SFVAMC. The program returned to the original format for cohorts 7 to 10, with students attending in-person small group sessions and conducting in-person interviews with inpatients.

Curriculum Evaluation

Students completed surveys in the week after the curriculum concluded. Survey completion was voluntary, anonymous, and had no bearing on their evaluation or grade (pass/fail only). Likert scale questions (1, strongly disagree; 5, strongly agree) were used to assess the program (eAppendix 1). One-way analysis of variance testing was used to compare means stratified by method of interview (in person, telephone, or video). Surveys also included free-response questions asking students to highlight aspects of the program they valued or would change; responses were summarized by theme. This program evaluation was deemed exempt from review by the UCSF Human Research Protection Program Institutional Review Board.

Sixty-two veteran stories were collected by 54 participating students (one student was unable to complete an interview, while several students completed multiple interviews). Fifty-four (87%) veterans requested their stories be entered into the medical record.

All 54 students completed the survey. Students reported that the MLMS curriculum helped them develop new skills for eliciting and recording a life story (mean [SD] 4.5 [0.7]). Most students strongly agreed that MLMS helped them understand how sharing a life story can impact a veteran’s experience of receiving health care, with a mean (SD) score of 4.8 (0.4). After completing MLMS, students also reported a better understanding of the mission of the VA and veteran demographics with a mean (SD) score of 4.4 (0.7) and 4.3 (0.7), respectively. Stratification of survey responses by method of interview (in person, telephone, or video) revealed no statistically significant differences in evaluations (Table 1).

Discussion

Previous research has demonstrated that a narrative medicine curriculum can help medicine clerkship students develop narrative competence through patient storytelling with a focus on a patient’s illness narrative.¹⁵ The VA MLMS program extends the patient narrative beyond health care–related experiences and encompasses their broader life story. This article adds to the MLMS and narrative medicine literature by demonstrating that the efficacy of teaching patient-centered care to medical trainees through direct interviews can be maintained in remote formats.⁹ The article also provides guidance for MLMS programs that wish to conduct remote veteran interviews.

The widespread adoption of telemedicine will require trainees to develop communication skills to establish therapeutic relationships with patients both face-to-face and through videoconferencing. In order to promote this important skill across varying levels of physical distancing, narrative medicine programs should be adaptable to a virtual learning environment. As we redesigned MLMS for the remote setting, we learned several key lessons that can guide similar curricular and programmatic innovations at other institutions. For example, videoconferencing created stronger connections between the students and veterans than telephone calls. However, tablet-based video interviews also introduced many technological challenges and required on-site personnel (nurses and volunteers) to connect students, veterans, and technology. Solutions for technology and communication challenges related to the basic personnel and infrastructure needed to start and maintain a remote MLMS program are outlined in Table 2.

Limitations

Limitations of this study include the participation of trainees from a single institution and a lack of assessment of the impact of MLMS on veterans. Future research could assess whether life story skills and practices are maintained after the medicine clerkship. In addition, future studies could examine veterans’ perspectives through interviews with qualitative analysis to learn how MLMS affected their experience of receiving health care.

Conclusions

This is the first report of a remote-capable life story curriculum for medical students. Shifting to a virtual MLMS curriculum requires protocols and people to link interviewers, veterans, and technology. Training for in-person interactions while being prepared for remote interviewing is essential to ensure that the MLMS experience remains available to interviewers and veterans who otherwise may never have the chance to connect. The restrictions and isolation of the COVID-19 pandemic will fade, but using MLMS to virtually connect patients, providers, and students will remain an important capability and opportunity as health care shifts to more virtual interaction.

Acknowledgments

The authors thank Emma Levine, MD, for her assistance coordinating video interviews; Thor Ringler, MS, MFA, for his assistance with manuscript review; and the veterans of the San Francisco VA Health Care System for sharing their stories.

Medication nonadherence is common with oral antipsychotic formulations, resulting in relapse, increased morbidity, and more frequent psychiatric hospitalization.^1-7 Psychiatric hospitalization and illness decompensation is costly to health care systems and leads to reduced quality of life for veterans and families.^6,7 Long-acting injectable antipsychotics (LAIAs) were developed to enhance antipsychotic adherence and improve patient outcomes, including reduced psychiatric hospitalization.^8-12

Little outcomes data exist comparing LAIAs, including biweekly risperidone microspheres and monthly paliperidone palmitate.^10-13 Risperidone microspheres require a 3-week oral crossover and are administered every 2 weeks, whereas paliperidone palmitate does not require an oral crossover and is administered every 4 weeks. The paliperidone palmitate loading regimen replaces an oral crossover.

The primary objective of this study was to compare the number of psychiatric hospitalizations between veterans administered risperidone microspheres and those on paliperidone palmitate pre- and post-LAIA initiation. Secondary objectives were to assess rehospitalization rates between patients taking risperidone microspheres and paliperidone palmitate, reduction in pre- and posthospitalization rates with LAIAs, and medication adherence.

Methods

This observational study with a retrospective cohort design was conducted at the Veterans Affairs Loma Linda Healthcare System (VALLHS) in California. We examined veterans who were initiated on LAIAs risperidone microspheres or paliperidone palmitate from January 01, 2016 through December 31, 2018. Veterans who were aged ≥ 18 years and received ≥ 2 injections of either risperidone microspheres or paliperidone palmitate during the study period were included. Veterans were excluded if they had received < 2 doses of either LAIA, received the LAIA outside of the review period, were nonadherent to risperidone crossover if they received risperidone microspheres, or transferred their care to another facility. At VALLHS, LAIA injections are administered by a nurse, and veterans must travel to the facility to receive the injections.

Extracted patient chart elements included participant demographics; diagnoses; comorbid alcohol, nicotine, opioid, or other substance use; duration on LAIA; psychiatric hospitalizations pre- and postinitiation of the LAIA; medication adherence; and medication discontinuation based on clinician documentation and clinic orders (Table 1).

Data Analysis

Patient demographics were analyzed using descriptive statistics and experimental comparisons between the risperidone microspheres and paliperidone palmitate groups to assess baseline differences between groups. Psychiatric hospitalizations pre- and post-LAIA were analyzed with parallel group (between veterans–independent groups) and pre-post (within veterans–dependent groups) designs. Index hospitalizations were examined for a period equivalent to the length of time veterans were on the LAIA. Psychiatric rehospitalization rates were analyzed for patients who had index hospitalizations and were rehospitalized for any period when they were receiving the LAIA. Incidences of pre- and post-LAIA hospitalizations were calculated in 100 person-years.

Parallel-group analysis was analyzed using the χ² and Mann-Whitney U tests. Pre-post analyses were analyzed using the Wilcoxon rank sum test. P was set at < .05 for statistical significance.

Results

We screened 111 veterans, and 97 were included in this study (risperidone microspheres, 44; paliperidone palmitate, 53). Mean (SD) age was 46 (13.8) years, 92% were male, 38% were White, 94% were diagnosed with schizophrenia or schizoaffective disorder, and 11% were homeless. Substance use was documented as 52% for nicotine products, 40% for alcohol, 31% for cannabis, 27% for methamphetamine, 7% for cocaine, and 3% for opioids. Cannabis, methamphetamine, cocaine, and opioid use were based on clinician documentation and listed as active diagnoses at the time of LAIA initiation. Statistical significance was found in index hospitalizations P = .009) and history of cocaine use disorder (6.8% vs 7.5%, P < .001).

Veterans administered risperidone microspheres had fewer mean (SD) post-LAIA hospitalizations (0.4 [1.0] vs 0.9 [1.5]; P = .02) and were less likely to be rehospitalized (22.7% vs 47.2%, P = .01) compared with paliperidone palmitate. However, veterans taking risperidone microspheres had a shorter mean (SD) treatment duration (41.6 [40.2] vs 58.2 [45.7] weeks, P = .04) compared with paliperidone palmitate, mainly because patients switched to a different LAIA or oral antipsychotic. No differences were detected in nonadherence and discontinuation between risperidone microspheres and paliperidone palmitate. All veterans in the risperidone microspheres group adhered to oral risperidone crossover with an average 87.8% days covered (Table 2).

Discussion

The main finding of this study was that initiation of LAIAs significantly reduced hospitalizations. Veterans taking risperidone microspheres had higher index hospitalizations and lower posttreatment hospitalizations compared with paliperidone palmitate. We found that patients initiated on risperidone microspheres had more hospitalizations before use of a LAIA than those initiated on paliperidone palmitate. Risperidone microspheres reduced the number of hospitalization post-LAIA significantly more than paliperidone palmitate. We also found that veterans taking risperidone microspheres were on the medication for less mean (SD) time than those on paliperidone palmitate (41.6 [40.2] vs 58.2 [45.7] weeks; P = .04).

To our knowledge, this is one of the few studies that compared outcomes of psychiatric hospitalizations, medication adherence, and treatment discontinuation between risperidone microspheres and paliperidone palmitate, specifically in a veteran population.^16-19 Limosin and colleagues aimed to compare length of stay during the initial hospitalization, rehospitalization risk, and treatment duration between risperidone microspheres and paliperidone palmitate in patients with schizophrenia.¹⁶ These researchers detected no differences in initial hospitalization duration and time to rehospitalization between risperidone microspheres and paliperidone palmitate.¹⁶ The study revealed a more favorable trend in time to discontinuation for paliperidone palmitate, but switching between LAIAs might have confounded the data.¹⁶ The authors note that their study lacked power, and patients on paliperidone palmitate had significantly more nonpsychiatric comorbidities.¹⁶ Joshi and colleagues looked at adherence, medication discontinuation, hospitalization rates, emergency department visits, and hospitalization costs between risperidone microspheres and paliperidone palmitate in patients identified in Truven MarketScan Commercial, Medicare Supplemental, and Medicaid Multi-State insurance databases.¹⁷ The authors found paliperidone palmitate to be superior in all objectives with better adherence, lower discontinuation rates, less likelihood of hospitalization, fewer emergency department visits, and lower hospitalization costs compared with risperidone microspheres.¹⁷ Korell and colleagues aimed to establish reference ranges for plasma concentrations of risperidone and paliperidone among adherent patients.¹⁸

The results in our study are somewhat unexpected in part because of the close relationship between risperidone and paliperidone. Risperidone is converted to paliperidone (9-OH-risperidone) via hepatic cytochrome P450 2D6. Although the molecules do not have identical pharmacologic profiles, it is accepted that they are similar enough that risperidone can establish oral tolerability when transitioning therapy to paliperidone palmitate and vice versa.²⁴ Although the active moiety in risperidone microspheres and paliperidone palmitate is similar, the dosing interval for risperidone microspheres is 2 weeks compared with 4 weeks with paliperidone palmitate. One potential explanation as to why veterans started on risperidone microspheres experienced better outcomes is because they had twice as many office visits with the health care team. Facility procedures dictate veterans receive the LAIA at an on-site clinic. During the visits, a licensed vocational nurse administers the injection and monitors the patient for 15 to 30 minutes afterward.

Despite new LAIAs coming to market, high-quality data examining potential differences in treatment outcomes among agents are limited. This is problematic for clinicians who want to optimize care by understanding how administration schedules or other aspects of LAIA use could modify treatment outcomes. Our results suggest that an advantage might exist in selecting an agent with a more frequent administration schedule, at least initially. This could allow for close monitoring and regular therapeutic contact, which could improve short-term outcomes. This conclusion is supported by meta-analyses, randomized controlled trials, and conceptual articles conducted by Wehring and colleagues, Berwaerts and colleagues, and Parellada and colleagues, respectively, who examined patients on different LAIAs and contact with health care professionals as part of their research.^26-28 These researchers concluded that patients who had regular contact with a health care professional had better outcomes when initiated on a LAIA.^26-28

Limitations

There are several limitations in this study. Retrospective and observational methods introduce risks of bias and confounding variables. Sample size might have limited statistical power to detect differences. Veterans might have had undocumented pre- or posthospitalizations at other institutions, which was not accounted for and lack of rehospitalization is not conclusive of a positive outcome. Institutions could improve on our study and help to fill gaps in comparative data by conducting larger analyses over longer periods and including more LAIA agents.

Conclusions

Although veterans that were administered risperidone microspheres had a shorter treatment duration, they were less likely to be rehospitalized, had a fewer mean number of post-LAIA hospitalizations, and had a larger difference in incidence in 100 person-years compared with veterans on paliperidone palmitate. Nonadherence and discontinuation rates were comparable between risperidone microspheres and paliperidone palmitate. Future studies could aim to further clarify differences in outcomes among agents or administration schedules.

Medication nonadherence is common with oral antipsychotic formulations, resulting in relapse, increased morbidity, and more frequent psychiatric hospitalization.^1-7 Psychiatric hospitalization and illness decompensation is costly to health care systems and leads to reduced quality of life for veterans and families.^6,7 Long-acting injectable antipsychotics (LAIAs) were developed to enhance antipsychotic adherence and improve patient outcomes, including reduced psychiatric hospitalization.^8-12

Little outcomes data exist comparing LAIAs, including biweekly risperidone microspheres and monthly paliperidone palmitate.^10-13 Risperidone microspheres require a 3-week oral crossover and are administered every 2 weeks, whereas paliperidone palmitate does not require an oral crossover and is administered every 4 weeks. The paliperidone palmitate loading regimen replaces an oral crossover.

The primary objective of this study was to compare the number of psychiatric hospitalizations between veterans administered risperidone microspheres and those on paliperidone palmitate pre- and post-LAIA initiation. Secondary objectives were to assess rehospitalization rates between patients taking risperidone microspheres and paliperidone palmitate, reduction in pre- and posthospitalization rates with LAIAs, and medication adherence.

Methods

This observational study with a retrospective cohort design was conducted at the Veterans Affairs Loma Linda Healthcare System (VALLHS) in California. We examined veterans who were initiated on LAIAs risperidone microspheres or paliperidone palmitate from January 01, 2016 through December 31, 2018. Veterans who were aged ≥ 18 years and received ≥ 2 injections of either risperidone microspheres or paliperidone palmitate during the study period were included. Veterans were excluded if they had received < 2 doses of either LAIA, received the LAIA outside of the review period, were nonadherent to risperidone crossover if they received risperidone microspheres, or transferred their care to another facility. At VALLHS, LAIA injections are administered by a nurse, and veterans must travel to the facility to receive the injections.

Extracted patient chart elements included participant demographics; diagnoses; comorbid alcohol, nicotine, opioid, or other substance use; duration on LAIA; psychiatric hospitalizations pre- and postinitiation of the LAIA; medication adherence; and medication discontinuation based on clinician documentation and clinic orders (Table 1).

Data Analysis

Patient demographics were analyzed using descriptive statistics and experimental comparisons between the risperidone microspheres and paliperidone palmitate groups to assess baseline differences between groups. Psychiatric hospitalizations pre- and post-LAIA were analyzed with parallel group (between veterans–independent groups) and pre-post (within veterans–dependent groups) designs. Index hospitalizations were examined for a period equivalent to the length of time veterans were on the LAIA. Psychiatric rehospitalization rates were analyzed for patients who had index hospitalizations and were rehospitalized for any period when they were receiving the LAIA. Incidences of pre- and post-LAIA hospitalizations were calculated in 100 person-years.

Parallel-group analysis was analyzed using the χ² and Mann-Whitney U tests. Pre-post analyses were analyzed using the Wilcoxon rank sum test. P was set at < .05 for statistical significance.

Results

We screened 111 veterans, and 97 were included in this study (risperidone microspheres, 44; paliperidone palmitate, 53). Mean (SD) age was 46 (13.8) years, 92% were male, 38% were White, 94% were diagnosed with schizophrenia or schizoaffective disorder, and 11% were homeless. Substance use was documented as 52% for nicotine products, 40% for alcohol, 31% for cannabis, 27% for methamphetamine, 7% for cocaine, and 3% for opioids. Cannabis, methamphetamine, cocaine, and opioid use were based on clinician documentation and listed as active diagnoses at the time of LAIA initiation. Statistical significance was found in index hospitalizations P = .009) and history of cocaine use disorder (6.8% vs 7.5%, P < .001).

Veterans administered risperidone microspheres had fewer mean (SD) post-LAIA hospitalizations (0.4 [1.0] vs 0.9 [1.5]; P = .02) and were less likely to be rehospitalized (22.7% vs 47.2%, P = .01) compared with paliperidone palmitate. However, veterans taking risperidone microspheres had a shorter mean (SD) treatment duration (41.6 [40.2] vs 58.2 [45.7] weeks, P = .04) compared with paliperidone palmitate, mainly because patients switched to a different LAIA or oral antipsychotic. No differences were detected in nonadherence and discontinuation between risperidone microspheres and paliperidone palmitate. All veterans in the risperidone microspheres group adhered to oral risperidone crossover with an average 87.8% days covered (Table 2).

Discussion

The main finding of this study was that initiation of LAIAs significantly reduced hospitalizations. Veterans taking risperidone microspheres had higher index hospitalizations and lower posttreatment hospitalizations compared with paliperidone palmitate. We found that patients initiated on risperidone microspheres had more hospitalizations before use of a LAIA than those initiated on paliperidone palmitate. Risperidone microspheres reduced the number of hospitalization post-LAIA significantly more than paliperidone palmitate. We also found that veterans taking risperidone microspheres were on the medication for less mean (SD) time than those on paliperidone palmitate (41.6 [40.2] vs 58.2 [45.7] weeks; P = .04).

To our knowledge, this is one of the few studies that compared outcomes of psychiatric hospitalizations, medication adherence, and treatment discontinuation between risperidone microspheres and paliperidone palmitate, specifically in a veteran population.^16-19 Limosin and colleagues aimed to compare length of stay during the initial hospitalization, rehospitalization risk, and treatment duration between risperidone microspheres and paliperidone palmitate in patients with schizophrenia.¹⁶ These researchers detected no differences in initial hospitalization duration and time to rehospitalization between risperidone microspheres and paliperidone palmitate.¹⁶ The study revealed a more favorable trend in time to discontinuation for paliperidone palmitate, but switching between LAIAs might have confounded the data.¹⁶ The authors note that their study lacked power, and patients on paliperidone palmitate had significantly more nonpsychiatric comorbidities.¹⁶ Joshi and colleagues looked at adherence, medication discontinuation, hospitalization rates, emergency department visits, and hospitalization costs between risperidone microspheres and paliperidone palmitate in patients identified in Truven MarketScan Commercial, Medicare Supplemental, and Medicaid Multi-State insurance databases.¹⁷ The authors found paliperidone palmitate to be superior in all objectives with better adherence, lower discontinuation rates, less likelihood of hospitalization, fewer emergency department visits, and lower hospitalization costs compared with risperidone microspheres.¹⁷ Korell and colleagues aimed to establish reference ranges for plasma concentrations of risperidone and paliperidone among adherent patients.¹⁸

The results in our study are somewhat unexpected in part because of the close relationship between risperidone and paliperidone. Risperidone is converted to paliperidone (9-OH-risperidone) via hepatic cytochrome P450 2D6. Although the molecules do not have identical pharmacologic profiles, it is accepted that they are similar enough that risperidone can establish oral tolerability when transitioning therapy to paliperidone palmitate and vice versa.²⁴ Although the active moiety in risperidone microspheres and paliperidone palmitate is similar, the dosing interval for risperidone microspheres is 2 weeks compared with 4 weeks with paliperidone palmitate. One potential explanation as to why veterans started on risperidone microspheres experienced better outcomes is because they had twice as many office visits with the health care team. Facility procedures dictate veterans receive the LAIA at an on-site clinic. During the visits, a licensed vocational nurse administers the injection and monitors the patient for 15 to 30 minutes afterward.

Despite new LAIAs coming to market, high-quality data examining potential differences in treatment outcomes among agents are limited. This is problematic for clinicians who want to optimize care by understanding how administration schedules or other aspects of LAIA use could modify treatment outcomes. Our results suggest that an advantage might exist in selecting an agent with a more frequent administration schedule, at least initially. This could allow for close monitoring and regular therapeutic contact, which could improve short-term outcomes. This conclusion is supported by meta-analyses, randomized controlled trials, and conceptual articles conducted by Wehring and colleagues, Berwaerts and colleagues, and Parellada and colleagues, respectively, who examined patients on different LAIAs and contact with health care professionals as part of their research.^26-28 These researchers concluded that patients who had regular contact with a health care professional had better outcomes when initiated on a LAIA.^26-28

Limitations

There are several limitations in this study. Retrospective and observational methods introduce risks of bias and confounding variables. Sample size might have limited statistical power to detect differences. Veterans might have had undocumented pre- or posthospitalizations at other institutions, which was not accounted for and lack of rehospitalization is not conclusive of a positive outcome. Institutions could improve on our study and help to fill gaps in comparative data by conducting larger analyses over longer periods and including more LAIA agents.

Conclusions

Although veterans that were administered risperidone microspheres had a shorter treatment duration, they were less likely to be rehospitalized, had a fewer mean number of post-LAIA hospitalizations, and had a larger difference in incidence in 100 person-years compared with veterans on paliperidone palmitate. Nonadherence and discontinuation rates were comparable between risperidone microspheres and paliperidone palmitate. Future studies could aim to further clarify differences in outcomes among agents or administration schedules.

Medication nonadherence is common with oral antipsychotic formulations, resulting in relapse, increased morbidity, and more frequent psychiatric hospitalization.^1-7 Psychiatric hospitalization and illness decompensation is costly to health care systems and leads to reduced quality of life for veterans and families.^6,7 Long-acting injectable antipsychotics (LAIAs) were developed to enhance antipsychotic adherence and improve patient outcomes, including reduced psychiatric hospitalization.^8-12

Little outcomes data exist comparing LAIAs, including biweekly risperidone microspheres and monthly paliperidone palmitate.^10-13 Risperidone microspheres require a 3-week oral crossover and are administered every 2 weeks, whereas paliperidone palmitate does not require an oral crossover and is administered every 4 weeks. The paliperidone palmitate loading regimen replaces an oral crossover.

The primary objective of this study was to compare the number of psychiatric hospitalizations between veterans administered risperidone microspheres and those on paliperidone palmitate pre- and post-LAIA initiation. Secondary objectives were to assess rehospitalization rates between patients taking risperidone microspheres and paliperidone palmitate, reduction in pre- and posthospitalization rates with LAIAs, and medication adherence.

Methods

This observational study with a retrospective cohort design was conducted at the Veterans Affairs Loma Linda Healthcare System (VALLHS) in California. We examined veterans who were initiated on LAIAs risperidone microspheres or paliperidone palmitate from January 01, 2016 through December 31, 2018. Veterans who were aged ≥ 18 years and received ≥ 2 injections of either risperidone microspheres or paliperidone palmitate during the study period were included. Veterans were excluded if they had received < 2 doses of either LAIA, received the LAIA outside of the review period, were nonadherent to risperidone crossover if they received risperidone microspheres, or transferred their care to another facility. At VALLHS, LAIA injections are administered by a nurse, and veterans must travel to the facility to receive the injections.

Extracted patient chart elements included participant demographics; diagnoses; comorbid alcohol, nicotine, opioid, or other substance use; duration on LAIA; psychiatric hospitalizations pre- and postinitiation of the LAIA; medication adherence; and medication discontinuation based on clinician documentation and clinic orders (Table 1).

Data Analysis

Patient demographics were analyzed using descriptive statistics and experimental comparisons between the risperidone microspheres and paliperidone palmitate groups to assess baseline differences between groups. Psychiatric hospitalizations pre- and post-LAIA were analyzed with parallel group (between veterans–independent groups) and pre-post (within veterans–dependent groups) designs. Index hospitalizations were examined for a period equivalent to the length of time veterans were on the LAIA. Psychiatric rehospitalization rates were analyzed for patients who had index hospitalizations and were rehospitalized for any period when they were receiving the LAIA. Incidences of pre- and post-LAIA hospitalizations were calculated in 100 person-years.

Parallel-group analysis was analyzed using the χ² and Mann-Whitney U tests. Pre-post analyses were analyzed using the Wilcoxon rank sum test. P was set at < .05 for statistical significance.

Results

We screened 111 veterans, and 97 were included in this study (risperidone microspheres, 44; paliperidone palmitate, 53). Mean (SD) age was 46 (13.8) years, 92% were male, 38% were White, 94% were diagnosed with schizophrenia or schizoaffective disorder, and 11% were homeless. Substance use was documented as 52% for nicotine products, 40% for alcohol, 31% for cannabis, 27% for methamphetamine, 7% for cocaine, and 3% for opioids. Cannabis, methamphetamine, cocaine, and opioid use were based on clinician documentation and listed as active diagnoses at the time of LAIA initiation. Statistical significance was found in index hospitalizations P = .009) and history of cocaine use disorder (6.8% vs 7.5%, P < .001).

Veterans administered risperidone microspheres had fewer mean (SD) post-LAIA hospitalizations (0.4 [1.0] vs 0.9 [1.5]; P = .02) and were less likely to be rehospitalized (22.7% vs 47.2%, P = .01) compared with paliperidone palmitate. However, veterans taking risperidone microspheres had a shorter mean (SD) treatment duration (41.6 [40.2] vs 58.2 [45.7] weeks, P = .04) compared with paliperidone palmitate, mainly because patients switched to a different LAIA or oral antipsychotic. No differences were detected in nonadherence and discontinuation between risperidone microspheres and paliperidone palmitate. All veterans in the risperidone microspheres group adhered to oral risperidone crossover with an average 87.8% days covered (Table 2).

Discussion

The main finding of this study was that initiation of LAIAs significantly reduced hospitalizations. Veterans taking risperidone microspheres had higher index hospitalizations and lower posttreatment hospitalizations compared with paliperidone palmitate. We found that patients initiated on risperidone microspheres had more hospitalizations before use of a LAIA than those initiated on paliperidone palmitate. Risperidone microspheres reduced the number of hospitalization post-LAIA significantly more than paliperidone palmitate. We also found that veterans taking risperidone microspheres were on the medication for less mean (SD) time than those on paliperidone palmitate (41.6 [40.2] vs 58.2 [45.7] weeks; P = .04).

To our knowledge, this is one of the few studies that compared outcomes of psychiatric hospitalizations, medication adherence, and treatment discontinuation between risperidone microspheres and paliperidone palmitate, specifically in a veteran population.^16-19 Limosin and colleagues aimed to compare length of stay during the initial hospitalization, rehospitalization risk, and treatment duration between risperidone microspheres and paliperidone palmitate in patients with schizophrenia.¹⁶ These researchers detected no differences in initial hospitalization duration and time to rehospitalization between risperidone microspheres and paliperidone palmitate.¹⁶ The study revealed a more favorable trend in time to discontinuation for paliperidone palmitate, but switching between LAIAs might have confounded the data.¹⁶ The authors note that their study lacked power, and patients on paliperidone palmitate had significantly more nonpsychiatric comorbidities.¹⁶ Joshi and colleagues looked at adherence, medication discontinuation, hospitalization rates, emergency department visits, and hospitalization costs between risperidone microspheres and paliperidone palmitate in patients identified in Truven MarketScan Commercial, Medicare Supplemental, and Medicaid Multi-State insurance databases.¹⁷ The authors found paliperidone palmitate to be superior in all objectives with better adherence, lower discontinuation rates, less likelihood of hospitalization, fewer emergency department visits, and lower hospitalization costs compared with risperidone microspheres.¹⁷ Korell and colleagues aimed to establish reference ranges for plasma concentrations of risperidone and paliperidone among adherent patients.¹⁸

The results in our study are somewhat unexpected in part because of the close relationship between risperidone and paliperidone. Risperidone is converted to paliperidone (9-OH-risperidone) via hepatic cytochrome P450 2D6. Although the molecules do not have identical pharmacologic profiles, it is accepted that they are similar enough that risperidone can establish oral tolerability when transitioning therapy to paliperidone palmitate and vice versa.²⁴ Although the active moiety in risperidone microspheres and paliperidone palmitate is similar, the dosing interval for risperidone microspheres is 2 weeks compared with 4 weeks with paliperidone palmitate. One potential explanation as to why veterans started on risperidone microspheres experienced better outcomes is because they had twice as many office visits with the health care team. Facility procedures dictate veterans receive the LAIA at an on-site clinic. During the visits, a licensed vocational nurse administers the injection and monitors the patient for 15 to 30 minutes afterward.

Despite new LAIAs coming to market, high-quality data examining potential differences in treatment outcomes among agents are limited. This is problematic for clinicians who want to optimize care by understanding how administration schedules or other aspects of LAIA use could modify treatment outcomes. Our results suggest that an advantage might exist in selecting an agent with a more frequent administration schedule, at least initially. This could allow for close monitoring and regular therapeutic contact, which could improve short-term outcomes. This conclusion is supported by meta-analyses, randomized controlled trials, and conceptual articles conducted by Wehring and colleagues, Berwaerts and colleagues, and Parellada and colleagues, respectively, who examined patients on different LAIAs and contact with health care professionals as part of their research.^26-28 These researchers concluded that patients who had regular contact with a health care professional had better outcomes when initiated on a LAIA.^26-28

Limitations

There are several limitations in this study. Retrospective and observational methods introduce risks of bias and confounding variables. Sample size might have limited statistical power to detect differences. Veterans might have had undocumented pre- or posthospitalizations at other institutions, which was not accounted for and lack of rehospitalization is not conclusive of a positive outcome. Institutions could improve on our study and help to fill gaps in comparative data by conducting larger analyses over longer periods and including more LAIA agents.

Conclusions

Although veterans that were administered risperidone microspheres had a shorter treatment duration, they were less likely to be rehospitalized, had a fewer mean number of post-LAIA hospitalizations, and had a larger difference in incidence in 100 person-years compared with veterans on paliperidone palmitate. Nonadherence and discontinuation rates were comparable between risperidone microspheres and paliperidone palmitate. Future studies could aim to further clarify differences in outcomes among agents or administration schedules.

Readers of this column may recall that since I have been the Editor-in-Chief of Federal Practitioner, my December editorial focuses on the best and worst of the year in federal medicine. In 2021, these evaluative terms fail to capture the sadness and global devastation that mark this grim epoch of the continuing pandemic, increasing climate disasters, rising political tensions, and racial violence. Thus, this year my editorial is framed in terms of the philosophical or theological categories of good and evil as the only concepts that can even begin to express the horrendous events that occurred in West Virginia.

On June 28, 2018, then US Department of Veterans Affairs (VA) Veterans Health Administration (VHA) Executive-in-Charge, Carolyn Clancy, MD, contacted Inspector General Michael Missal to alert him that “there may be an ‘Angel of Death’ in Clarksburg [West Virginia].”¹ Two years later Reta Mays, a 46-year-old VA nursing assistant, entered a guilty plea in federal court to the deaths of 7 vulnerable veterans. The legal charges were second-degree murder and 1 count of assault with intent to commit murder by injecting insulin. The victims were all patients on Ward 3 at the Louis A. Johnson VA, Medical Center in Clarksburg, where Mays worked the night shift from 2015 to 2018.² Mays was sentenced in May of this year to 7 consecutive life terms for each of the veterans whose lives she cruelly ended and an additional 240 months for the eighth patient who survived her murder attempt.3

The term angel of death has religious roots in Judaism, although not strictly in the Hebrew scriptures. Neither the Jewish nor Christian Bible identifies a specific figure who is the angel of death. The idea first appears in Rabbinic literature and Jewish tradition.⁴ The angel God sends as a messenger of death is known as malakh ha-mavet in Hebrew. The revered Jewish physician and philosopher Moses Maimonides taught in his Guide for the Perplexed the angel of death is synonymous with the devil, and the evil inclination that dwells in the mind of all human beings.⁵In modern times, the concept of an angel of death has come to designate a serial killer who is a health care professional (HCP). A group of forensic scientists, HCPs, and attorneys, including former VA Under Secretary for Health Dr. Kenneth Kizer, published a study of HCPs who had been prosecuted or convicted of serial murder. Nurses constituted the largest group of offenders (60%) with nursing aides like Mays responsible for 18% of murders, and physicians 12%. The review found that though health care serial killers are rare, they operate in nations across the Western world, in many different states in this country, and in almost all health care settings, including previous VA angels of death.⁶Nursing aides who are not supposed to have access to medications—a major problem in Mays’ case—nor permitted to administer them more often resort to noncontrolled substances to kill their victims.¹ Mays chose insulin as her murder weapon as did 13% of serial killers. Just as insulin may be difficult to detect in toxicology, so Mays and others like her committed their crimes on the night shift when they were less likely to be discovered.⁶

Many of us feel compelled to seek a rational motivation for why healers would mutate into killers: If we can find a reason for this heinous behavior it somehow helps us feel the world is more intelligible and controllable. Unfortunately, despite intensive forensic investigations of multiple angels of death, there is little definitive understanding of the motives of these murders.⁶ Mays disclosed more than most. As part of a plea bargain, she provided investigators with 2 rationales for her killing: She wanted to ease the patients’ suffering. Such claims of being an angel of mercy are common among HCP serial murders, which the patterns of the killings generally disprove. The patients Mays lethally injected, while mostly old and ill, were all expected to recover and leave the hospital. The Inspector General report uncovered a cautionary detail that has at least indirect bearing on the nursing assistant’s contention that she “wanted to let the patient’s die gently”: Contrary to VHA requirements, the facility had no functioning palliative care team. This finding in no way excuses or even explains Mays’ actions; it does, however, reinforce the essential value of palliative care expertise in an aging veteran population with many life-limiting conditions.⁷

Mays’ second motivation seems more plausible, based on her life narrative and the literature on HCP serial killers. Mays disclosed to investigators that she “had a lot of stress and chaos in [her] personal and professional life and these actions gave [her] a sense of control.”¹ Her prior use of excessive force when employed at a prison as well as forensic science indicating that feelings of wielding power over life and death often drive health care murders, suggest this may have been a factor in Mays’ horrific conduct.⁸

It seems blasphemous to associate the word good in the same pages with this terrible evil. Nothing can compensate or justify the betrayal of the sacred oath of an HCP and the public trust of a VHA employee. Yet that very impossibility carries with it an obligation to ask, as did the author of an article about a recent Canadian nurse serial killer, “What can we learn from the [Mays] story?”⁹

Mays could never have taken the lives of 8 patients without clinical and administrative lapses and shortcuts at all levels of the health care system. Indeed, the 100-plus page Inspector General report makes 15 recommendations for the VHA, the Veterans Integrated Service Network, and the facility, encompassing areas of personnel hiring and performance evaluation, medication management and security, reporting and responding to unexplained events, quality and safety programs oversight, leaders’ responses, corrective actions, and even computer systems data analysis.

I want to suggest 2 ethical additions to this list addressed to all of us as VHA staff and especially to those of us who are HCPs. From the perspective of virtue ethics, Reta Mays is a tragedy about complacency and compromise in everyday work that the pandemic has made even more frequent and challenging to avoid and resist. This is what the Roman Virgil means in the epigraph that the road down to hell is easy and the road back very difficult.

I propose the need for discernment in trying to listen to our moral intuitions that tell us something is amiss and diligence in adhering to best practices even when we are fearful, exhausted, demoralized, or apathetic. These 2 habits of commitment to veterans, one of compassion and the other of competence, can help us follow the good inclinations of our hearts and together with system changes can bar the doors of our hospitals to the visits of future angels of death. This dedication is the least we owe to the families of the patients at Clarksburg whose loved ones never came home and whose questions likely can never be fully answered.

Readers of this column may recall that since I have been the Editor-in-Chief of Federal Practitioner, my December editorial focuses on the best and worst of the year in federal medicine. In 2021, these evaluative terms fail to capture the sadness and global devastation that mark this grim epoch of the continuing pandemic, increasing climate disasters, rising political tensions, and racial violence. Thus, this year my editorial is framed in terms of the philosophical or theological categories of good and evil as the only concepts that can even begin to express the horrendous events that occurred in West Virginia.

On June 28, 2018, then US Department of Veterans Affairs (VA) Veterans Health Administration (VHA) Executive-in-Charge, Carolyn Clancy, MD, contacted Inspector General Michael Missal to alert him that “there may be an ‘Angel of Death’ in Clarksburg [West Virginia].”¹ Two years later Reta Mays, a 46-year-old VA nursing assistant, entered a guilty plea in federal court to the deaths of 7 vulnerable veterans. The legal charges were second-degree murder and 1 count of assault with intent to commit murder by injecting insulin. The victims were all patients on Ward 3 at the Louis A. Johnson VA, Medical Center in Clarksburg, where Mays worked the night shift from 2015 to 2018.² Mays was sentenced in May of this year to 7 consecutive life terms for each of the veterans whose lives she cruelly ended and an additional 240 months for the eighth patient who survived her murder attempt.3

The term angel of death has religious roots in Judaism, although not strictly in the Hebrew scriptures. Neither the Jewish nor Christian Bible identifies a specific figure who is the angel of death. The idea first appears in Rabbinic literature and Jewish tradition.⁴ The angel God sends as a messenger of death is known as malakh ha-mavet in Hebrew. The revered Jewish physician and philosopher Moses Maimonides taught in his Guide for the Perplexed the angel of death is synonymous with the devil, and the evil inclination that dwells in the mind of all human beings.⁵In modern times, the concept of an angel of death has come to designate a serial killer who is a health care professional (HCP). A group of forensic scientists, HCPs, and attorneys, including former VA Under Secretary for Health Dr. Kenneth Kizer, published a study of HCPs who had been prosecuted or convicted of serial murder. Nurses constituted the largest group of offenders (60%) with nursing aides like Mays responsible for 18% of murders, and physicians 12%. The review found that though health care serial killers are rare, they operate in nations across the Western world, in many different states in this country, and in almost all health care settings, including previous VA angels of death.⁶Nursing aides who are not supposed to have access to medications—a major problem in Mays’ case—nor permitted to administer them more often resort to noncontrolled substances to kill their victims.¹ Mays chose insulin as her murder weapon as did 13% of serial killers. Just as insulin may be difficult to detect in toxicology, so Mays and others like her committed their crimes on the night shift when they were less likely to be discovered.⁶

Many of us feel compelled to seek a rational motivation for why healers would mutate into killers: If we can find a reason for this heinous behavior it somehow helps us feel the world is more intelligible and controllable. Unfortunately, despite intensive forensic investigations of multiple angels of death, there is little definitive understanding of the motives of these murders.⁶ Mays disclosed more than most. As part of a plea bargain, she provided investigators with 2 rationales for her killing: She wanted to ease the patients’ suffering. Such claims of being an angel of mercy are common among HCP serial murders, which the patterns of the killings generally disprove. The patients Mays lethally injected, while mostly old and ill, were all expected to recover and leave the hospital. The Inspector General report uncovered a cautionary detail that has at least indirect bearing on the nursing assistant’s contention that she “wanted to let the patient’s die gently”: Contrary to VHA requirements, the facility had no functioning palliative care team. This finding in no way excuses or even explains Mays’ actions; it does, however, reinforce the essential value of palliative care expertise in an aging veteran population with many life-limiting conditions.⁷

Mays’ second motivation seems more plausible, based on her life narrative and the literature on HCP serial killers. Mays disclosed to investigators that she “had a lot of stress and chaos in [her] personal and professional life and these actions gave [her] a sense of control.”¹ Her prior use of excessive force when employed at a prison as well as forensic science indicating that feelings of wielding power over life and death often drive health care murders, suggest this may have been a factor in Mays’ horrific conduct.⁸

It seems blasphemous to associate the word good in the same pages with this terrible evil. Nothing can compensate or justify the betrayal of the sacred oath of an HCP and the public trust of a VHA employee. Yet that very impossibility carries with it an obligation to ask, as did the author of an article about a recent Canadian nurse serial killer, “What can we learn from the [Mays] story?”⁹

Mays could never have taken the lives of 8 patients without clinical and administrative lapses and shortcuts at all levels of the health care system. Indeed, the 100-plus page Inspector General report makes 15 recommendations for the VHA, the Veterans Integrated Service Network, and the facility, encompassing areas of personnel hiring and performance evaluation, medication management and security, reporting and responding to unexplained events, quality and safety programs oversight, leaders’ responses, corrective actions, and even computer systems data analysis.

I want to suggest 2 ethical additions to this list addressed to all of us as VHA staff and especially to those of us who are HCPs. From the perspective of virtue ethics, Reta Mays is a tragedy about complacency and compromise in everyday work that the pandemic has made even more frequent and challenging to avoid and resist. This is what the Roman Virgil means in the epigraph that the road down to hell is easy and the road back very difficult.

I propose the need for discernment in trying to listen to our moral intuitions that tell us something is amiss and diligence in adhering to best practices even when we are fearful, exhausted, demoralized, or apathetic. These 2 habits of commitment to veterans, one of compassion and the other of competence, can help us follow the good inclinations of our hearts and together with system changes can bar the doors of our hospitals to the visits of future angels of death. This dedication is the least we owe to the families of the patients at Clarksburg whose loved ones never came home and whose questions likely can never be fully answered.

Readers of this column may recall that since I have been the Editor-in-Chief of Federal Practitioner, my December editorial focuses on the best and worst of the year in federal medicine. In 2021, these evaluative terms fail to capture the sadness and global devastation that mark this grim epoch of the continuing pandemic, increasing climate disasters, rising political tensions, and racial violence. Thus, this year my editorial is framed in terms of the philosophical or theological categories of good and evil as the only concepts that can even begin to express the horrendous events that occurred in West Virginia.

On June 28, 2018, then US Department of Veterans Affairs (VA) Veterans Health Administration (VHA) Executive-in-Charge, Carolyn Clancy, MD, contacted Inspector General Michael Missal to alert him that “there may be an ‘Angel of Death’ in Clarksburg [West Virginia].”¹ Two years later Reta Mays, a 46-year-old VA nursing assistant, entered a guilty plea in federal court to the deaths of 7 vulnerable veterans. The legal charges were second-degree murder and 1 count of assault with intent to commit murder by injecting insulin. The victims were all patients on Ward 3 at the Louis A. Johnson VA, Medical Center in Clarksburg, where Mays worked the night shift from 2015 to 2018.² Mays was sentenced in May of this year to 7 consecutive life terms for each of the veterans whose lives she cruelly ended and an additional 240 months for the eighth patient who survived her murder attempt.3

The term angel of death has religious roots in Judaism, although not strictly in the Hebrew scriptures. Neither the Jewish nor Christian Bible identifies a specific figure who is the angel of death. The idea first appears in Rabbinic literature and Jewish tradition.⁴ The angel God sends as a messenger of death is known as malakh ha-mavet in Hebrew. The revered Jewish physician and philosopher Moses Maimonides taught in his Guide for the Perplexed the angel of death is synonymous with the devil, and the evil inclination that dwells in the mind of all human beings.⁵In modern times, the concept of an angel of death has come to designate a serial killer who is a health care professional (HCP). A group of forensic scientists, HCPs, and attorneys, including former VA Under Secretary for Health Dr. Kenneth Kizer, published a study of HCPs who had been prosecuted or convicted of serial murder. Nurses constituted the largest group of offenders (60%) with nursing aides like Mays responsible for 18% of murders, and physicians 12%. The review found that though health care serial killers are rare, they operate in nations across the Western world, in many different states in this country, and in almost all health care settings, including previous VA angels of death.⁶Nursing aides who are not supposed to have access to medications—a major problem in Mays’ case—nor permitted to administer them more often resort to noncontrolled substances to kill their victims.¹ Mays chose insulin as her murder weapon as did 13% of serial killers. Just as insulin may be difficult to detect in toxicology, so Mays and others like her committed their crimes on the night shift when they were less likely to be discovered.⁶

Many of us feel compelled to seek a rational motivation for why healers would mutate into killers: If we can find a reason for this heinous behavior it somehow helps us feel the world is more intelligible and controllable. Unfortunately, despite intensive forensic investigations of multiple angels of death, there is little definitive understanding of the motives of these murders.⁶ Mays disclosed more than most. As part of a plea bargain, she provided investigators with 2 rationales for her killing: She wanted to ease the patients’ suffering. Such claims of being an angel of mercy are common among HCP serial murders, which the patterns of the killings generally disprove. The patients Mays lethally injected, while mostly old and ill, were all expected to recover and leave the hospital. The Inspector General report uncovered a cautionary detail that has at least indirect bearing on the nursing assistant’s contention that she “wanted to let the patient’s die gently”: Contrary to VHA requirements, the facility had no functioning palliative care team. This finding in no way excuses or even explains Mays’ actions; it does, however, reinforce the essential value of palliative care expertise in an aging veteran population with many life-limiting conditions.⁷

Mays’ second motivation seems more plausible, based on her life narrative and the literature on HCP serial killers. Mays disclosed to investigators that she “had a lot of stress and chaos in [her] personal and professional life and these actions gave [her] a sense of control.”¹ Her prior use of excessive force when employed at a prison as well as forensic science indicating that feelings of wielding power over life and death often drive health care murders, suggest this may have been a factor in Mays’ horrific conduct.⁸

It seems blasphemous to associate the word good in the same pages with this terrible evil. Nothing can compensate or justify the betrayal of the sacred oath of an HCP and the public trust of a VHA employee. Yet that very impossibility carries with it an obligation to ask, as did the author of an article about a recent Canadian nurse serial killer, “What can we learn from the [Mays] story?”⁹

Mays could never have taken the lives of 8 patients without clinical and administrative lapses and shortcuts at all levels of the health care system. Indeed, the 100-plus page Inspector General report makes 15 recommendations for the VHA, the Veterans Integrated Service Network, and the facility, encompassing areas of personnel hiring and performance evaluation, medication management and security, reporting and responding to unexplained events, quality and safety programs oversight, leaders’ responses, corrective actions, and even computer systems data analysis.

I want to suggest 2 ethical additions to this list addressed to all of us as VHA staff and especially to those of us who are HCPs. From the perspective of virtue ethics, Reta Mays is a tragedy about complacency and compromise in everyday work that the pandemic has made even more frequent and challenging to avoid and resist. This is what the Roman Virgil means in the epigraph that the road down to hell is easy and the road back very difficult.

I propose the need for discernment in trying to listen to our moral intuitions that tell us something is amiss and diligence in adhering to best practices even when we are fearful, exhausted, demoralized, or apathetic. These 2 habits of commitment to veterans, one of compassion and the other of competence, can help us follow the good inclinations of our hearts and together with system changes can bar the doors of our hospitals to the visits of future angels of death. This dedication is the least we owe to the families of the patients at Clarksburg whose loved ones never came home and whose questions likely can never be fully answered.

Study Overview

Objective. To compare the effects of intermittent energy restriction (IER) with those of continuous energy restriction (CER) on blood pressure control and weight loss in overweight and obese patients with hypertension during a 6-month period.

Design. Randomized controlled trial.

Settings and participants. The trial was conducted at the Affiliated Hospital of Jiaxing University from June 1, 2020, to April 30, 2021. Chinese adults were recruited using advertisements and flyers posted in the hospital and local communities. Prior to participation in study activities, all participants gave informed consent prior to recruitment and were provided compensation in the form of a $38 voucher at 3 and 6 months for their time for participating in the study.

The main inclusion criteria were patients between the ages of 18 and 70 years, hypertension, and body mass index (BMI) ranging from 24 to 40 kg/m². The exclusion criteria were systolic blood pressure (SBP) ≥ 180 mmHg or diastolic blood pressure (DBP) ≥ 120 mmHg, type 1 or 2 diabetes with a history of severe hypoglycemic episodes, pregnancy or breastfeeding, usage of glucagon-like peptide 1 receptor agonists, weight loss > 5 kg within the past 3 months or previous weight loss surgery, and inability to adhere to the dietary protocol.

Of the 294 participants screened for eligibility, 205 were randomized in a 1:1 ratio to the IER group (n = 102) or the CER group (n = 103), stratified by sex and BMI (as overweight or obese). All participants were required to have a stable medication regimen and weight in the 3 months prior to enrollment and not to use weight-loss drugs or vitamin supplements for the duration of the study. Researchers and participants were not blinded to the study group assignment.

Interventions. Participants randomly assigned to the IER group followed a 5:2 eating pattern: a very-low-energy diet of 500-600 kcal for 2 days of the week along with their usual diet for the other 5 days. The 2 days of calorie restriction could be consecutive or nonconsecutive, with a minimum of 0.8 g supplemental protein per kg of body weight per day, in accordance with the 2016 Dietary Guidelines for Chinese Residents. The CER group was advised to consume 1000 kcal/day for women and 1200 kcal/day for men on a 7-day energy restriction. That is, they were prescribed a daily 25% restriction based on the general principles of a Mediterranean-type diet (30% fat, 45-50% carbohydrate, and 20-25% protein).

Both groups received dietary education from a qualified dietitian and were recommended to maintain their current daily activity levels throughout the trial. Written dietary information brochures with portion advice and sample meal plans were provided to improve compliance in each group. All participants received a digital cooking scale to weigh foods to ensure accuracy of intake and were required to keep a food diary while following the recommended recipe on 2 days/week during calorie restriction to help with adherence. No food was provided. All participants were followed up by regular outpatient visits to both cardiologists and dietitians once a month. Diet checklists, activity schedules, and weight were reviewed to assess compliance with dietary advice at each visit.

Of note, participants were encouraged to measure and record their BP twice daily, and if 2 consecutive BP readings were < 110/70 mmHg and/or accompanied by hypotensive episodes with symptoms (dizziness, nausea, headache, and fatigue), they were asked to contact the investigators directly. Antihypertensive medication changes were then made in consultation with cardiologists. In addition, a medication management protocol (ie, doses of antidiabetic medications, including insulin and sulfonylurea) was designed to avoid hypoglycemia. Medication could be reduced in the CER group based on the basal dose at the endocrinologist’s discretion. In the IER group, insulin and sulfonylureas were discontinued on calorie restriction days only, and long-acting insulin was discontinued the night before the IER day. Insulin was not to be resumed until a full day’s caloric intake was achieved.

Measures and analysis. The primary outcomes of this study were changes in BP and weight (measured using an automatic digital sphygmomanometer and an electronic scale), and the secondary outcomes were changes in body composition (assessed by dual-energy x-ray absorptiometry scanning), as well as glycosylated hemoglobin A_1c (HbA_1c) levels and blood lipids after 6 months. All outcome measures were recorded at baseline and at each monthly visit. Incidence rates of hypoglycemia were based on blood glucose (defined as blood glucose < 70 mg/dL) and/or symptomatic hypoglycemia (symptoms of sweating, paleness, dizziness, and confusion). Two cardiologists who were blind to the patients’ diet condition measured and recorded all pertinent clinical parameters and adjudicated serious adverse events.

Data were compared using independent-samples t-tests or the Mann–Whitney U test for continuous variables, and Pearson’s χ² test or Fisher’s exact test for categorial variables as appropriate. Repeated-measures ANOVA via a linear mixed model was employed to test the effects of diet, time, and their interaction. In subgroup analyses, differential effects of the intervention on the primary outcomes were evaluated with respect to patients’ level of education, domicile, and sex based on the statistical significance of the interaction term for the subgroup of interest in the multivariate model. Analyses were performed based on completers and on an intention-to-treat principle.

Main results. Among the 205 randomized participants, 118 were women and 87 were men; mean (SD) age was 50.5 (8.8) years; mean (SD) BMI was 28.7 (2.6); mean (SD) SBP was 143 (10) mmHg; and mean (SD) DBP was 91 (9) mmHg. At the end of the 6-month intervention, 173 (84.4%) completed the study (IER group: n = 88; CER group: n = 85). Both groups had similar dropout rates at 6 months (IER group: 14 participants [13.7%]; CER group: 18 participants [17.5%]; P = .83) and were well matched for baseline characteristics except for triglyceride levels.

In the completers analysis, both groups experienced significant reductions in weight (mean [SEM]), but there was no difference between treatment groups (−7.2 [0.6] kg in the IER group vs −7.1 [0.6] kg in the CER group; diet by time P = .72). Similarly, the change in SBP and DBP achieved was statistically significant over time, but there was also no difference between the dietary interventions (−8 [0.7] mmHg in the IER group vs −8 [0.6] mmHg in the CER group, diet by time P = .68; −6 [0.6] mmHg in the IER group vs −6 [0.5] mmHg in the CER group, diet by time P = .53]. Subgroup analyses of the association of the intervention with weight, SBP and DBP by sex, education, and domicile showed no significant between-group differences.

All measures of body composition decreased significantly at 6 months with both groups experiencing comparable reductions in total fat mass (−5.5 [0.6] kg in the IER group vs −4.8 [0.5] kg in the CER group, diet by time P = .08) and android fat mass (−1.1 [0.2] kg in the IER group vs −0.8 [0.2] kg in the CER group, diet by time P = .16). Of note, participants in the CER group lost significantly more total fat-free mass than did participants in the IER group (mean [SEM], −2.3 [0.2] kg vs −1.7 [0.2] kg; P = .03], and there was a trend toward a greater change in total fat mass in the IER group (P = .08). The secondary outcome of mean (SEM) HbA_1c (−0.2% [0.1%]) and blood lipid levels (triglyceride level, −1.0 [0.3] mmol/L; total cholesterol level, −0.9 [0.2] mmol/L; low-density lipoprotein cholesterol level, −0.9 [0.2 mmol/L; high-density lipoprotein cholesterol level, 0.7 [0.3] mmol/L] improved with weight loss (P < .05), with no differences between groups (diet by time P > .05).

The intention-to-treat analysis demonstrated that IER and CER are equally effective for weight loss and blood pressure control: both groups experienced significant reductions in weight, SBP, and DBP, but with no difference between treatment groups – mean (SEM) weight change with IER was −7.0 (0.6) kg vs −6.8 (0.6) kg with CER; the mean (SEM) SBP with IER was −7 (0.7) mmHg vs −7 (0.6) mmHg with CER; and the mean (SEM) DBP with IER was −6 (0.5) mmHg vs −5 (0.5) mmHg with CER, (diet by time P = .62, .39, and .41, respectively). There were favorable improvements in body composition, HbA_1c, and blood lipid levels, with no differences between groups.

Conclusion. A 2-day severe energy restriction with 5 days of habitual eating compared to 7 days of CER provides an acceptable alternative for BP control and weight loss in overweight and obese individuals with hypertension after 6 months. IER may offer a useful alternative strategy for this population, who find continuous weight-loss diets too difficult to maintain.

Commentary

Globally, obesity represents a major health challenge as it substantially increases the risk of diseases such as hypertension, type 2 diabetes, and coronary heart disease.¹ Lifestyle modifications, including weight loss and increased physical activity, are recommended in major guidelines as a first-step intervention in the treatment of hypertensive patients.² However, lifestyle and behavioral interventions aimed at reducing calorie intake through low-calorie dieting is challenging as it is dependent on individual motivation and adherence to a strict, continuous protocol. Further, CER strategies have limited effectiveness because complex and persistent hormonal, metabolic, and neurochemical adaptations defend against weight loss and promote weight regain.^3-4 IER has drawn attention in the popular media as an alternative to CER due to its feasibility and even potential for higher rates of compliance.⁵

This study adds to the literature as it is the first randomized controlled trial (to the knowledge of the authors at the time of publication) to explore 2 forms of energy restriction – CER and IER – and their impact on weight loss, BP, body composition, HbA_1c, and blood lipid levels in overweight and obese patients with high blood pressure. Results from this study showed that IER is as effective as, but not superior to, CER (in terms of the outcomes measures assessed). Specifically, findings highlighted that the 5:2 diet is an effective strategy and noninferior to that of daily calorie restriction for BP and weight control. In addition, both weight loss and BP reduction were greater in a subgroup of obese compared with overweight participants, which indicates that obese populations may benefit more from energy restriction. As the authors highlight, this study both aligns with and expands on current related literature.

This study has both strengths and limitations, especially with regard to the design and data analysis strategy. A key strength is the randomized controlled trial design which enables increased internal validity and decreases several sources of bias, including selection bias and confounding. In addition, it was also designed as a pragmatic trial, with the protocol reflecting efforts to replicate the real-world environment by not supplying meal replacements or food. Notably, only 9 patients could not comply with the protocol, indicating that acceptability of the diet protocol was high. However, as this was only a 6-month long study, further studies are needed to determine whether a 5:2 diet is sustainable (and effective) in the long-term compared with CER, which the authors highlight. The study was also adequately powered to detect clinically meaningful differences in weight loss and SBP, and appropriate analyses were performed on both the basis of completers and on an intention-to-treat principle. However, further studies are needed that are adequately powered to also detect clinically meaningful differences in the other measures, ie, body composition, HbA_1c, and blood lipid levels. Importantly, generalizability of findings from this study is limited as the study population comprises only Chinese adults, predominately middle-aged, overweight, and had mildly to moderately elevated SBP and DBP, and excluded diabetic patients. Thus, findings are not necessarily applicable to individuals with highly elevated blood pressure or poorly controlled diabetes.

Applications for Clinical Practice

Results of this study demonstrated that IER is an effective alternative diet strategy for weight loss and blood pressure control in overweight and obese patients with hypertension and is comparable to CER. This is relevant for clinical practice as IER may be easier to maintain in this population compared to continuous weight-loss diets. Importantly, both types of calorie restriction require clinical oversight as medication changes and periodic monitoring of hypotensive and hypoglycemic episodes are needed. Clinicians should consider what is feasible and sustainable for their patients when recommending intermittent energy restriction.

Financial disclosures: None.

Study Overview

Objective. To compare the effects of intermittent energy restriction (IER) with those of continuous energy restriction (CER) on blood pressure control and weight loss in overweight and obese patients with hypertension during a 6-month period.

Design. Randomized controlled trial.

Settings and participants. The trial was conducted at the Affiliated Hospital of Jiaxing University from June 1, 2020, to April 30, 2021. Chinese adults were recruited using advertisements and flyers posted in the hospital and local communities. Prior to participation in study activities, all participants gave informed consent prior to recruitment and were provided compensation in the form of a $38 voucher at 3 and 6 months for their time for participating in the study.

The main inclusion criteria were patients between the ages of 18 and 70 years, hypertension, and body mass index (BMI) ranging from 24 to 40 kg/m². The exclusion criteria were systolic blood pressure (SBP) ≥ 180 mmHg or diastolic blood pressure (DBP) ≥ 120 mmHg, type 1 or 2 diabetes with a history of severe hypoglycemic episodes, pregnancy or breastfeeding, usage of glucagon-like peptide 1 receptor agonists, weight loss > 5 kg within the past 3 months or previous weight loss surgery, and inability to adhere to the dietary protocol.

Of the 294 participants screened for eligibility, 205 were randomized in a 1:1 ratio to the IER group (n = 102) or the CER group (n = 103), stratified by sex and BMI (as overweight or obese). All participants were required to have a stable medication regimen and weight in the 3 months prior to enrollment and not to use weight-loss drugs or vitamin supplements for the duration of the study. Researchers and participants were not blinded to the study group assignment.

Interventions. Participants randomly assigned to the IER group followed a 5:2 eating pattern: a very-low-energy diet of 500-600 kcal for 2 days of the week along with their usual diet for the other 5 days. The 2 days of calorie restriction could be consecutive or nonconsecutive, with a minimum of 0.8 g supplemental protein per kg of body weight per day, in accordance with the 2016 Dietary Guidelines for Chinese Residents. The CER group was advised to consume 1000 kcal/day for women and 1200 kcal/day for men on a 7-day energy restriction. That is, they were prescribed a daily 25% restriction based on the general principles of a Mediterranean-type diet (30% fat, 45-50% carbohydrate, and 20-25% protein).

Both groups received dietary education from a qualified dietitian and were recommended to maintain their current daily activity levels throughout the trial. Written dietary information brochures with portion advice and sample meal plans were provided to improve compliance in each group. All participants received a digital cooking scale to weigh foods to ensure accuracy of intake and were required to keep a food diary while following the recommended recipe on 2 days/week during calorie restriction to help with adherence. No food was provided. All participants were followed up by regular outpatient visits to both cardiologists and dietitians once a month. Diet checklists, activity schedules, and weight were reviewed to assess compliance with dietary advice at each visit.

Of note, participants were encouraged to measure and record their BP twice daily, and if 2 consecutive BP readings were < 110/70 mmHg and/or accompanied by hypotensive episodes with symptoms (dizziness, nausea, headache, and fatigue), they were asked to contact the investigators directly. Antihypertensive medication changes were then made in consultation with cardiologists. In addition, a medication management protocol (ie, doses of antidiabetic medications, including insulin and sulfonylurea) was designed to avoid hypoglycemia. Medication could be reduced in the CER group based on the basal dose at the endocrinologist’s discretion. In the IER group, insulin and sulfonylureas were discontinued on calorie restriction days only, and long-acting insulin was discontinued the night before the IER day. Insulin was not to be resumed until a full day’s caloric intake was achieved.

Measures and analysis. The primary outcomes of this study were changes in BP and weight (measured using an automatic digital sphygmomanometer and an electronic scale), and the secondary outcomes were changes in body composition (assessed by dual-energy x-ray absorptiometry scanning), as well as glycosylated hemoglobin A_1c (HbA_1c) levels and blood lipids after 6 months. All outcome measures were recorded at baseline and at each monthly visit. Incidence rates of hypoglycemia were based on blood glucose (defined as blood glucose < 70 mg/dL) and/or symptomatic hypoglycemia (symptoms of sweating, paleness, dizziness, and confusion). Two cardiologists who were blind to the patients’ diet condition measured and recorded all pertinent clinical parameters and adjudicated serious adverse events.

Data were compared using independent-samples t-tests or the Mann–Whitney U test for continuous variables, and Pearson’s χ² test or Fisher’s exact test for categorial variables as appropriate. Repeated-measures ANOVA via a linear mixed model was employed to test the effects of diet, time, and their interaction. In subgroup analyses, differential effects of the intervention on the primary outcomes were evaluated with respect to patients’ level of education, domicile, and sex based on the statistical significance of the interaction term for the subgroup of interest in the multivariate model. Analyses were performed based on completers and on an intention-to-treat principle.

Main results. Among the 205 randomized participants, 118 were women and 87 were men; mean (SD) age was 50.5 (8.8) years; mean (SD) BMI was 28.7 (2.6); mean (SD) SBP was 143 (10) mmHg; and mean (SD) DBP was 91 (9) mmHg. At the end of the 6-month intervention, 173 (84.4%) completed the study (IER group: n = 88; CER group: n = 85). Both groups had similar dropout rates at 6 months (IER group: 14 participants [13.7%]; CER group: 18 participants [17.5%]; P = .83) and were well matched for baseline characteristics except for triglyceride levels.

In the completers analysis, both groups experienced significant reductions in weight (mean [SEM]), but there was no difference between treatment groups (−7.2 [0.6] kg in the IER group vs −7.1 [0.6] kg in the CER group; diet by time P = .72). Similarly, the change in SBP and DBP achieved was statistically significant over time, but there was also no difference between the dietary interventions (−8 [0.7] mmHg in the IER group vs −8 [0.6] mmHg in the CER group, diet by time P = .68; −6 [0.6] mmHg in the IER group vs −6 [0.5] mmHg in the CER group, diet by time P = .53]. Subgroup analyses of the association of the intervention with weight, SBP and DBP by sex, education, and domicile showed no significant between-group differences.

All measures of body composition decreased significantly at 6 months with both groups experiencing comparable reductions in total fat mass (−5.5 [0.6] kg in the IER group vs −4.8 [0.5] kg in the CER group, diet by time P = .08) and android fat mass (−1.1 [0.2] kg in the IER group vs −0.8 [0.2] kg in the CER group, diet by time P = .16). Of note, participants in the CER group lost significantly more total fat-free mass than did participants in the IER group (mean [SEM], −2.3 [0.2] kg vs −1.7 [0.2] kg; P = .03], and there was a trend toward a greater change in total fat mass in the IER group (P = .08). The secondary outcome of mean (SEM) HbA_1c (−0.2% [0.1%]) and blood lipid levels (triglyceride level, −1.0 [0.3] mmol/L; total cholesterol level, −0.9 [0.2] mmol/L; low-density lipoprotein cholesterol level, −0.9 [0.2 mmol/L; high-density lipoprotein cholesterol level, 0.7 [0.3] mmol/L] improved with weight loss (P < .05), with no differences between groups (diet by time P > .05).

The intention-to-treat analysis demonstrated that IER and CER are equally effective for weight loss and blood pressure control: both groups experienced significant reductions in weight, SBP, and DBP, but with no difference between treatment groups – mean (SEM) weight change with IER was −7.0 (0.6) kg vs −6.8 (0.6) kg with CER; the mean (SEM) SBP with IER was −7 (0.7) mmHg vs −7 (0.6) mmHg with CER; and the mean (SEM) DBP with IER was −6 (0.5) mmHg vs −5 (0.5) mmHg with CER, (diet by time P = .62, .39, and .41, respectively). There were favorable improvements in body composition, HbA_1c, and blood lipid levels, with no differences between groups.

Conclusion. A 2-day severe energy restriction with 5 days of habitual eating compared to 7 days of CER provides an acceptable alternative for BP control and weight loss in overweight and obese individuals with hypertension after 6 months. IER may offer a useful alternative strategy for this population, who find continuous weight-loss diets too difficult to maintain.

Commentary

Globally, obesity represents a major health challenge as it substantially increases the risk of diseases such as hypertension, type 2 diabetes, and coronary heart disease.¹ Lifestyle modifications, including weight loss and increased physical activity, are recommended in major guidelines as a first-step intervention in the treatment of hypertensive patients.² However, lifestyle and behavioral interventions aimed at reducing calorie intake through low-calorie dieting is challenging as it is dependent on individual motivation and adherence to a strict, continuous protocol. Further, CER strategies have limited effectiveness because complex and persistent hormonal, metabolic, and neurochemical adaptations defend against weight loss and promote weight regain.^3-4 IER has drawn attention in the popular media as an alternative to CER due to its feasibility and even potential for higher rates of compliance.⁵

This study adds to the literature as it is the first randomized controlled trial (to the knowledge of the authors at the time of publication) to explore 2 forms of energy restriction – CER and IER – and their impact on weight loss, BP, body composition, HbA_1c, and blood lipid levels in overweight and obese patients with high blood pressure. Results from this study showed that IER is as effective as, but not superior to, CER (in terms of the outcomes measures assessed). Specifically, findings highlighted that the 5:2 diet is an effective strategy and noninferior to that of daily calorie restriction for BP and weight control. In addition, both weight loss and BP reduction were greater in a subgroup of obese compared with overweight participants, which indicates that obese populations may benefit more from energy restriction. As the authors highlight, this study both aligns with and expands on current related literature.

This study has both strengths and limitations, especially with regard to the design and data analysis strategy. A key strength is the randomized controlled trial design which enables increased internal validity and decreases several sources of bias, including selection bias and confounding. In addition, it was also designed as a pragmatic trial, with the protocol reflecting efforts to replicate the real-world environment by not supplying meal replacements or food. Notably, only 9 patients could not comply with the protocol, indicating that acceptability of the diet protocol was high. However, as this was only a 6-month long study, further studies are needed to determine whether a 5:2 diet is sustainable (and effective) in the long-term compared with CER, which the authors highlight. The study was also adequately powered to detect clinically meaningful differences in weight loss and SBP, and appropriate analyses were performed on both the basis of completers and on an intention-to-treat principle. However, further studies are needed that are adequately powered to also detect clinically meaningful differences in the other measures, ie, body composition, HbA_1c, and blood lipid levels. Importantly, generalizability of findings from this study is limited as the study population comprises only Chinese adults, predominately middle-aged, overweight, and had mildly to moderately elevated SBP and DBP, and excluded diabetic patients. Thus, findings are not necessarily applicable to individuals with highly elevated blood pressure or poorly controlled diabetes.

Applications for Clinical Practice

Results of this study demonstrated that IER is an effective alternative diet strategy for weight loss and blood pressure control in overweight and obese patients with hypertension and is comparable to CER. This is relevant for clinical practice as IER may be easier to maintain in this population compared to continuous weight-loss diets. Importantly, both types of calorie restriction require clinical oversight as medication changes and periodic monitoring of hypotensive and hypoglycemic episodes are needed. Clinicians should consider what is feasible and sustainable for their patients when recommending intermittent energy restriction.

Financial disclosures: None.

Study Overview

Objective. To compare the effects of intermittent energy restriction (IER) with those of continuous energy restriction (CER) on blood pressure control and weight loss in overweight and obese patients with hypertension during a 6-month period.

Design. Randomized controlled trial.

Settings and participants. The trial was conducted at the Affiliated Hospital of Jiaxing University from June 1, 2020, to April 30, 2021. Chinese adults were recruited using advertisements and flyers posted in the hospital and local communities. Prior to participation in study activities, all participants gave informed consent prior to recruitment and were provided compensation in the form of a $38 voucher at 3 and 6 months for their time for participating in the study.

The main inclusion criteria were patients between the ages of 18 and 70 years, hypertension, and body mass index (BMI) ranging from 24 to 40 kg/m². The exclusion criteria were systolic blood pressure (SBP) ≥ 180 mmHg or diastolic blood pressure (DBP) ≥ 120 mmHg, type 1 or 2 diabetes with a history of severe hypoglycemic episodes, pregnancy or breastfeeding, usage of glucagon-like peptide 1 receptor agonists, weight loss > 5 kg within the past 3 months or previous weight loss surgery, and inability to adhere to the dietary protocol.

Of the 294 participants screened for eligibility, 205 were randomized in a 1:1 ratio to the IER group (n = 102) or the CER group (n = 103), stratified by sex and BMI (as overweight or obese). All participants were required to have a stable medication regimen and weight in the 3 months prior to enrollment and not to use weight-loss drugs or vitamin supplements for the duration of the study. Researchers and participants were not blinded to the study group assignment.

Interventions. Participants randomly assigned to the IER group followed a 5:2 eating pattern: a very-low-energy diet of 500-600 kcal for 2 days of the week along with their usual diet for the other 5 days. The 2 days of calorie restriction could be consecutive or nonconsecutive, with a minimum of 0.8 g supplemental protein per kg of body weight per day, in accordance with the 2016 Dietary Guidelines for Chinese Residents. The CER group was advised to consume 1000 kcal/day for women and 1200 kcal/day for men on a 7-day energy restriction. That is, they were prescribed a daily 25% restriction based on the general principles of a Mediterranean-type diet (30% fat, 45-50% carbohydrate, and 20-25% protein).

Both groups received dietary education from a qualified dietitian and were recommended to maintain their current daily activity levels throughout the trial. Written dietary information brochures with portion advice and sample meal plans were provided to improve compliance in each group. All participants received a digital cooking scale to weigh foods to ensure accuracy of intake and were required to keep a food diary while following the recommended recipe on 2 days/week during calorie restriction to help with adherence. No food was provided. All participants were followed up by regular outpatient visits to both cardiologists and dietitians once a month. Diet checklists, activity schedules, and weight were reviewed to assess compliance with dietary advice at each visit.

Of note, participants were encouraged to measure and record their BP twice daily, and if 2 consecutive BP readings were < 110/70 mmHg and/or accompanied by hypotensive episodes with symptoms (dizziness, nausea, headache, and fatigue), they were asked to contact the investigators directly. Antihypertensive medication changes were then made in consultation with cardiologists. In addition, a medication management protocol (ie, doses of antidiabetic medications, including insulin and sulfonylurea) was designed to avoid hypoglycemia. Medication could be reduced in the CER group based on the basal dose at the endocrinologist’s discretion. In the IER group, insulin and sulfonylureas were discontinued on calorie restriction days only, and long-acting insulin was discontinued the night before the IER day. Insulin was not to be resumed until a full day’s caloric intake was achieved.

Measures and analysis. The primary outcomes of this study were changes in BP and weight (measured using an automatic digital sphygmomanometer and an electronic scale), and the secondary outcomes were changes in body composition (assessed by dual-energy x-ray absorptiometry scanning), as well as glycosylated hemoglobin A_1c (HbA_1c) levels and blood lipids after 6 months. All outcome measures were recorded at baseline and at each monthly visit. Incidence rates of hypoglycemia were based on blood glucose (defined as blood glucose < 70 mg/dL) and/or symptomatic hypoglycemia (symptoms of sweating, paleness, dizziness, and confusion). Two cardiologists who were blind to the patients’ diet condition measured and recorded all pertinent clinical parameters and adjudicated serious adverse events.

Data were compared using independent-samples t-tests or the Mann–Whitney U test for continuous variables, and Pearson’s χ² test or Fisher’s exact test for categorial variables as appropriate. Repeated-measures ANOVA via a linear mixed model was employed to test the effects of diet, time, and their interaction. In subgroup analyses, differential effects of the intervention on the primary outcomes were evaluated with respect to patients’ level of education, domicile, and sex based on the statistical significance of the interaction term for the subgroup of interest in the multivariate model. Analyses were performed based on completers and on an intention-to-treat principle.

Main results. Among the 205 randomized participants, 118 were women and 87 were men; mean (SD) age was 50.5 (8.8) years; mean (SD) BMI was 28.7 (2.6); mean (SD) SBP was 143 (10) mmHg; and mean (SD) DBP was 91 (9) mmHg. At the end of the 6-month intervention, 173 (84.4%) completed the study (IER group: n = 88; CER group: n = 85). Both groups had similar dropout rates at 6 months (IER group: 14 participants [13.7%]; CER group: 18 participants [17.5%]; P = .83) and were well matched for baseline characteristics except for triglyceride levels.

In the completers analysis, both groups experienced significant reductions in weight (mean [SEM]), but there was no difference between treatment groups (−7.2 [0.6] kg in the IER group vs −7.1 [0.6] kg in the CER group; diet by time P = .72). Similarly, the change in SBP and DBP achieved was statistically significant over time, but there was also no difference between the dietary interventions (−8 [0.7] mmHg in the IER group vs −8 [0.6] mmHg in the CER group, diet by time P = .68; −6 [0.6] mmHg in the IER group vs −6 [0.5] mmHg in the CER group, diet by time P = .53]. Subgroup analyses of the association of the intervention with weight, SBP and DBP by sex, education, and domicile showed no significant between-group differences.

All measures of body composition decreased significantly at 6 months with both groups experiencing comparable reductions in total fat mass (−5.5 [0.6] kg in the IER group vs −4.8 [0.5] kg in the CER group, diet by time P = .08) and android fat mass (−1.1 [0.2] kg in the IER group vs −0.8 [0.2] kg in the CER group, diet by time P = .16). Of note, participants in the CER group lost significantly more total fat-free mass than did participants in the IER group (mean [SEM], −2.3 [0.2] kg vs −1.7 [0.2] kg; P = .03], and there was a trend toward a greater change in total fat mass in the IER group (P = .08). The secondary outcome of mean (SEM) HbA_1c (−0.2% [0.1%]) and blood lipid levels (triglyceride level, −1.0 [0.3] mmol/L; total cholesterol level, −0.9 [0.2] mmol/L; low-density lipoprotein cholesterol level, −0.9 [0.2 mmol/L; high-density lipoprotein cholesterol level, 0.7 [0.3] mmol/L] improved with weight loss (P < .05), with no differences between groups (diet by time P > .05).

The intention-to-treat analysis demonstrated that IER and CER are equally effective for weight loss and blood pressure control: both groups experienced significant reductions in weight, SBP, and DBP, but with no difference between treatment groups – mean (SEM) weight change with IER was −7.0 (0.6) kg vs −6.8 (0.6) kg with CER; the mean (SEM) SBP with IER was −7 (0.7) mmHg vs −7 (0.6) mmHg with CER; and the mean (SEM) DBP with IER was −6 (0.5) mmHg vs −5 (0.5) mmHg with CER, (diet by time P = .62, .39, and .41, respectively). There were favorable improvements in body composition, HbA_1c, and blood lipid levels, with no differences between groups.

Conclusion. A 2-day severe energy restriction with 5 days of habitual eating compared to 7 days of CER provides an acceptable alternative for BP control and weight loss in overweight and obese individuals with hypertension after 6 months. IER may offer a useful alternative strategy for this population, who find continuous weight-loss diets too difficult to maintain.

Commentary

Globally, obesity represents a major health challenge as it substantially increases the risk of diseases such as hypertension, type 2 diabetes, and coronary heart disease.¹ Lifestyle modifications, including weight loss and increased physical activity, are recommended in major guidelines as a first-step intervention in the treatment of hypertensive patients.² However, lifestyle and behavioral interventions aimed at reducing calorie intake through low-calorie dieting is challenging as it is dependent on individual motivation and adherence to a strict, continuous protocol. Further, CER strategies have limited effectiveness because complex and persistent hormonal, metabolic, and neurochemical adaptations defend against weight loss and promote weight regain.^3-4 IER has drawn attention in the popular media as an alternative to CER due to its feasibility and even potential for higher rates of compliance.⁵

This study adds to the literature as it is the first randomized controlled trial (to the knowledge of the authors at the time of publication) to explore 2 forms of energy restriction – CER and IER – and their impact on weight loss, BP, body composition, HbA_1c, and blood lipid levels in overweight and obese patients with high blood pressure. Results from this study showed that IER is as effective as, but not superior to, CER (in terms of the outcomes measures assessed). Specifically, findings highlighted that the 5:2 diet is an effective strategy and noninferior to that of daily calorie restriction for BP and weight control. In addition, both weight loss and BP reduction were greater in a subgroup of obese compared with overweight participants, which indicates that obese populations may benefit more from energy restriction. As the authors highlight, this study both aligns with and expands on current related literature.

This study has both strengths and limitations, especially with regard to the design and data analysis strategy. A key strength is the randomized controlled trial design which enables increased internal validity and decreases several sources of bias, including selection bias and confounding. In addition, it was also designed as a pragmatic trial, with the protocol reflecting efforts to replicate the real-world environment by not supplying meal replacements or food. Notably, only 9 patients could not comply with the protocol, indicating that acceptability of the diet protocol was high. However, as this was only a 6-month long study, further studies are needed to determine whether a 5:2 diet is sustainable (and effective) in the long-term compared with CER, which the authors highlight. The study was also adequately powered to detect clinically meaningful differences in weight loss and SBP, and appropriate analyses were performed on both the basis of completers and on an intention-to-treat principle. However, further studies are needed that are adequately powered to also detect clinically meaningful differences in the other measures, ie, body composition, HbA_1c, and blood lipid levels. Importantly, generalizability of findings from this study is limited as the study population comprises only Chinese adults, predominately middle-aged, overweight, and had mildly to moderately elevated SBP and DBP, and excluded diabetic patients. Thus, findings are not necessarily applicable to individuals with highly elevated blood pressure or poorly controlled diabetes.

Applications for Clinical Practice

Results of this study demonstrated that IER is an effective alternative diet strategy for weight loss and blood pressure control in overweight and obese patients with hypertension and is comparable to CER. This is relevant for clinical practice as IER may be easier to maintain in this population compared to continuous weight-loss diets. Importantly, both types of calorie restriction require clinical oversight as medication changes and periodic monitoring of hypotensive and hypoglycemic episodes are needed. Clinicians should consider what is feasible and sustainable for their patients when recommending intermittent energy restriction.

Financial disclosures: None.

Study Overview

Objective. The objective of this study was to evaluate orders and documentation describing perioperative management of code status in adults.

Design. A retrospective case series of all adult inpatients admitted to hospitals at 1 academic health system in the US.

Setting and participants. This retrospective case series was conducted at 5 hospitals within the University of Pennsylvania Health System. Cases included all adult inpatients admitted to hospitals between March 2017 and September 2018 who had a Do-Not-Resuscitate (DNR) order placed in their medical record during admission and subsequently underwent a surgical procedure that required anesthesia care.

Main outcome measures. Medical records of included cases were manually reviewed by the authors to verify whether a DNR order was in place at the time surgical intervention was discussed with a patient. Clinical notes and DNR orders of eligible cases were reviewed to identify documentation and outcome of goals of care discussions that were conducted within 48 hours prior to the surgical procedure. Collected data included patient demographics (age, sex, race); case characteristics (American Society of Anesthesiologists [ASA] physical status score, anesthesia type [general vs others such as regional], emergency status [emergent vs elective surgery], procedures by service [surgical including hip fracture repair, gastrostomy or jejunostomy, or exploratory laparotomy vs medical including endoscopy, bronchoscopy, or transesophageal echocardiogram]); and hospital policy for perioperative management of DNR orders (written policy encouraging discussion vs written policy plus additional initiatives, including procedure-specific DNR form). The primary outcome was the presence of a preoperative order or note documenting code status discussion or change. Data were analyzed using χ² and Fisher exact tests and the threshold for statistical significance was P < .05.

Main results. Of the 27 665 inpatient procedures identified across 5 hospitals, 444 (1.6%) cases met the inclusion criteria. Patients from these cases aged 75 (SD 13) years (95% CI, 72-77 years); 247 (56%, 95% CI, 55%-57%) were women; and 300 (68%, 95% CI, 65%-71%) were White. A total of 426 patients (96%, 95% CI, 90%-100%) had an ASA physical status score of 3 or higher and 237 (53%, 95% CI, 51%-56%) received general anesthesia. The most common procedures performed were endoscopy (148 [33%]), hip fracture repair (43 [10%]), and gastrostomy or jejunostomy (28 [6%]). Reevaluation of code status was documented in 126 cases (28%, 95% CI, 25%-31%); code status orders were changed in 20 of 126 cases (16%, 95% CI, 7%-24%); and a note was filed without a corresponding order for 106 of 126 cases (84%, 95% CI, 75%-95%). In the majority of cases (109 of 126 [87%], 95% CI, 78%-95%) in which documented discussion occurred, DNR orders were suspended. Of 126 cases in which a discussion was documented, participants of these discussions included surgeons 10% of the time (13 cases, 95% CI, 8%-13%), members of the anesthesia team 51% of the time (64 cases, 95% CI, 49%-53%), and medicine or palliative care clinicians 39% of the time (49 cases, 95% CI, 37%-41%).

The rate of documented preoperative code status discussion was higher in patients with higher ASA physical status score (35% in patients with an ASA physical status score ≥ 4 [55 of 155] vs 25% in those with an ASA physical status score ≤ 3 [71 of 289]; P = .02). The rates of documented preoperative code status discussion were similar by anesthesia type (29% for general anesthesia [69 of 237 cases] vs 28% [57 of 207 cases] for other modalities; P = .70). The hospitals involved in this study all had a written policy encouraging rediscussion of code status before surgery. However, only 1 hospital reported added measures (eg, provision of a procedure-specific DNR form) to increase documentation of preoperative code status discussions. In this specific hospital, documentation of preoperative code status discussions was higher compared to other hospitals (67% [37 of 55 cases] vs 23% [89 of 389 cases]; P < .01).

Conclusion. In a retrospective case series conducted at 5 hospitals within 1 academic health system in the US, fewer than 1 in 5 patients with preexisting DNR orders had a documented discussion of code status prior to undergoing surgery. Additional strategies including the development of institutional protocols that facilitate perioperative management of advance directives, identification of local champions, and patient education, should be explored as means to improve preoperative code status reevaulation per guideline recommendations.

It is not unusual that patients with a DNR order may require and undergo surgical interventions to treat reversible conditions, prevent progression of underlying disease, or mitigate distressing symptoms such as pain. For instance, intubation, mechanical ventilation, and administration of vasoactive drugs are resuscitative measures that may be needed to safely anesthetize and sedate a patient. As such, the American College of Surgeons¹ has provided a statement on advance directives by patients with an existing DNR order to guide management. Specifically, the statement indicates that the best approach for these patients is a policy of “required reconsideration” of the existing DNR order. Required reconsideration means that “the patient or designated surrogate and the physicians who will be responsible for the patient’s care should, when possible, discuss the new intraoperative and perioperative risks associated with the surgical procedure, the patient’s treatment goals, and an approach for potentially life-threatening problems consistent with the patient’s values and preferences.” Moreover, the required reconsideration discussion needs to occur as early as it is practical once a decision is made to have surgery because the discussion “may result in the patient agreeing to suspend the DNR order during surgery and the perioperative period, retaining the original DNR order, or modifying the DNR order.” Given that surgical patients with DNR orders have significant comorbidities, many sustain postoperative complications, and nearly 1 in 4 die within 30 days of surgery, preoperative advance care planning (ACP) and code status discussions are particularly essential to delivering high quality surgical care.²

In the current study, Hadler et al³ conducted a retrospective analysis to evaluate orders and documentation describing perioperative management of code status in patients with existing DNR order at an academic health system in the US. The authors reported that fewer than 20% of patients with existing DNR orders had a documented discussion of code status prior to undergoing surgery. These findings add to the notion that compliance with such guidance on required reconsideration discussion is suboptimal in perioperative care in the US.^4,5 A recently published study focused on patients aged more than 60 years undergoing high-risk oncologic or vascular surgeries similarly showed that the frequency of ACP discussions or advance directive documentations among older patients was low.⁶ This growing body of evidence is highly clinically relevant in that preoperative discussion on code status is highly relevant to the care of older adults, a population group that accounts for the majority of surgeries and is most vulnerable to poor surgical outcomes. Additionally, it highlights a disconnect between the shared recognition by surgeons and patients that ACP discussion is important in perioperative care and its low implementation rates.

Unsurprisingly, Hadler et al³ reported that added measures such as the provision of a procedure-specific DNR form led to an increase in the documentation of preoperative code status discussions in 1 of the hospitals studied. The authors suggested that strategies such as the development of institutional protocols aimed to facilitate perioperative advance directive discussions, identify local champions, and educate patients may be ways to improve preoperative code status reevaulation. The idea that institutional value and culture are key factors impacting surgeon behavior and may influence the practice of ACP discussion is not new. Thus, creative and adaptable strategies, resources, and trainings that are required by medical institutions and hospitals to support preoperative ACP discussions with patients undergoing surgeries need to be identified, validated, and implemented to optimize perioperative care in vulnerable patients.

Applications for Clinical Practice

The findings from the current study indicate that less than 20% of patients with preexisting DNR orders have a documented discussion of code status prior to undergoing surgery. Physicians and health care institutions need to identify barriers to, and implement strategies that, facilitate and optimize preoperative ACP discussions in order to provide patient-centered care in vulnerable surgical patients.

Financial disclosures: None.

Study Overview

Objective. The objective of this study was to evaluate orders and documentation describing perioperative management of code status in adults.

Design. A retrospective case series of all adult inpatients admitted to hospitals at 1 academic health system in the US.

Setting and participants. This retrospective case series was conducted at 5 hospitals within the University of Pennsylvania Health System. Cases included all adult inpatients admitted to hospitals between March 2017 and September 2018 who had a Do-Not-Resuscitate (DNR) order placed in their medical record during admission and subsequently underwent a surgical procedure that required anesthesia care.

Main outcome measures. Medical records of included cases were manually reviewed by the authors to verify whether a DNR order was in place at the time surgical intervention was discussed with a patient. Clinical notes and DNR orders of eligible cases were reviewed to identify documentation and outcome of goals of care discussions that were conducted within 48 hours prior to the surgical procedure. Collected data included patient demographics (age, sex, race); case characteristics (American Society of Anesthesiologists [ASA] physical status score, anesthesia type [general vs others such as regional], emergency status [emergent vs elective surgery], procedures by service [surgical including hip fracture repair, gastrostomy or jejunostomy, or exploratory laparotomy vs medical including endoscopy, bronchoscopy, or transesophageal echocardiogram]); and hospital policy for perioperative management of DNR orders (written policy encouraging discussion vs written policy plus additional initiatives, including procedure-specific DNR form). The primary outcome was the presence of a preoperative order or note documenting code status discussion or change. Data were analyzed using χ² and Fisher exact tests and the threshold for statistical significance was P < .05.

Main results. Of the 27 665 inpatient procedures identified across 5 hospitals, 444 (1.6%) cases met the inclusion criteria. Patients from these cases aged 75 (SD 13) years (95% CI, 72-77 years); 247 (56%, 95% CI, 55%-57%) were women; and 300 (68%, 95% CI, 65%-71%) were White. A total of 426 patients (96%, 95% CI, 90%-100%) had an ASA physical status score of 3 or higher and 237 (53%, 95% CI, 51%-56%) received general anesthesia. The most common procedures performed were endoscopy (148 [33%]), hip fracture repair (43 [10%]), and gastrostomy or jejunostomy (28 [6%]). Reevaluation of code status was documented in 126 cases (28%, 95% CI, 25%-31%); code status orders were changed in 20 of 126 cases (16%, 95% CI, 7%-24%); and a note was filed without a corresponding order for 106 of 126 cases (84%, 95% CI, 75%-95%). In the majority of cases (109 of 126 [87%], 95% CI, 78%-95%) in which documented discussion occurred, DNR orders were suspended. Of 126 cases in which a discussion was documented, participants of these discussions included surgeons 10% of the time (13 cases, 95% CI, 8%-13%), members of the anesthesia team 51% of the time (64 cases, 95% CI, 49%-53%), and medicine or palliative care clinicians 39% of the time (49 cases, 95% CI, 37%-41%).

The rate of documented preoperative code status discussion was higher in patients with higher ASA physical status score (35% in patients with an ASA physical status score ≥ 4 [55 of 155] vs 25% in those with an ASA physical status score ≤ 3 [71 of 289]; P = .02). The rates of documented preoperative code status discussion were similar by anesthesia type (29% for general anesthesia [69 of 237 cases] vs 28% [57 of 207 cases] for other modalities; P = .70). The hospitals involved in this study all had a written policy encouraging rediscussion of code status before surgery. However, only 1 hospital reported added measures (eg, provision of a procedure-specific DNR form) to increase documentation of preoperative code status discussions. In this specific hospital, documentation of preoperative code status discussions was higher compared to other hospitals (67% [37 of 55 cases] vs 23% [89 of 389 cases]; P < .01).

Conclusion. In a retrospective case series conducted at 5 hospitals within 1 academic health system in the US, fewer than 1 in 5 patients with preexisting DNR orders had a documented discussion of code status prior to undergoing surgery. Additional strategies including the development of institutional protocols that facilitate perioperative management of advance directives, identification of local champions, and patient education, should be explored as means to improve preoperative code status reevaulation per guideline recommendations.

It is not unusual that patients with a DNR order may require and undergo surgical interventions to treat reversible conditions, prevent progression of underlying disease, or mitigate distressing symptoms such as pain. For instance, intubation, mechanical ventilation, and administration of vasoactive drugs are resuscitative measures that may be needed to safely anesthetize and sedate a patient. As such, the American College of Surgeons¹ has provided a statement on advance directives by patients with an existing DNR order to guide management. Specifically, the statement indicates that the best approach for these patients is a policy of “required reconsideration” of the existing DNR order. Required reconsideration means that “the patient or designated surrogate and the physicians who will be responsible for the patient’s care should, when possible, discuss the new intraoperative and perioperative risks associated with the surgical procedure, the patient’s treatment goals, and an approach for potentially life-threatening problems consistent with the patient’s values and preferences.” Moreover, the required reconsideration discussion needs to occur as early as it is practical once a decision is made to have surgery because the discussion “may result in the patient agreeing to suspend the DNR order during surgery and the perioperative period, retaining the original DNR order, or modifying the DNR order.” Given that surgical patients with DNR orders have significant comorbidities, many sustain postoperative complications, and nearly 1 in 4 die within 30 days of surgery, preoperative advance care planning (ACP) and code status discussions are particularly essential to delivering high quality surgical care.²

In the current study, Hadler et al³ conducted a retrospective analysis to evaluate orders and documentation describing perioperative management of code status in patients with existing DNR order at an academic health system in the US. The authors reported that fewer than 20% of patients with existing DNR orders had a documented discussion of code status prior to undergoing surgery. These findings add to the notion that compliance with such guidance on required reconsideration discussion is suboptimal in perioperative care in the US.^4,5 A recently published study focused on patients aged more than 60 years undergoing high-risk oncologic or vascular surgeries similarly showed that the frequency of ACP discussions or advance directive documentations among older patients was low.⁶ This growing body of evidence is highly clinically relevant in that preoperative discussion on code status is highly relevant to the care of older adults, a population group that accounts for the majority of surgeries and is most vulnerable to poor surgical outcomes. Additionally, it highlights a disconnect between the shared recognition by surgeons and patients that ACP discussion is important in perioperative care and its low implementation rates.

Unsurprisingly, Hadler et al³ reported that added measures such as the provision of a procedure-specific DNR form led to an increase in the documentation of preoperative code status discussions in 1 of the hospitals studied. The authors suggested that strategies such as the development of institutional protocols aimed to facilitate perioperative advance directive discussions, identify local champions, and educate patients may be ways to improve preoperative code status reevaulation. The idea that institutional value and culture are key factors impacting surgeon behavior and may influence the practice of ACP discussion is not new. Thus, creative and adaptable strategies, resources, and trainings that are required by medical institutions and hospitals to support preoperative ACP discussions with patients undergoing surgeries need to be identified, validated, and implemented to optimize perioperative care in vulnerable patients.

Applications for Clinical Practice

The findings from the current study indicate that less than 20% of patients with preexisting DNR orders have a documented discussion of code status prior to undergoing surgery. Physicians and health care institutions need to identify barriers to, and implement strategies that, facilitate and optimize preoperative ACP discussions in order to provide patient-centered care in vulnerable surgical patients.

Financial disclosures: None.

Study Overview

Objective. The objective of this study was to evaluate orders and documentation describing perioperative management of code status in adults.

Design. A retrospective case series of all adult inpatients admitted to hospitals at 1 academic health system in the US.

Setting and participants. This retrospective case series was conducted at 5 hospitals within the University of Pennsylvania Health System. Cases included all adult inpatients admitted to hospitals between March 2017 and September 2018 who had a Do-Not-Resuscitate (DNR) order placed in their medical record during admission and subsequently underwent a surgical procedure that required anesthesia care.

Main outcome measures. Medical records of included cases were manually reviewed by the authors to verify whether a DNR order was in place at the time surgical intervention was discussed with a patient. Clinical notes and DNR orders of eligible cases were reviewed to identify documentation and outcome of goals of care discussions that were conducted within 48 hours prior to the surgical procedure. Collected data included patient demographics (age, sex, race); case characteristics (American Society of Anesthesiologists [ASA] physical status score, anesthesia type [general vs others such as regional], emergency status [emergent vs elective surgery], procedures by service [surgical including hip fracture repair, gastrostomy or jejunostomy, or exploratory laparotomy vs medical including endoscopy, bronchoscopy, or transesophageal echocardiogram]); and hospital policy for perioperative management of DNR orders (written policy encouraging discussion vs written policy plus additional initiatives, including procedure-specific DNR form). The primary outcome was the presence of a preoperative order or note documenting code status discussion or change. Data were analyzed using χ² and Fisher exact tests and the threshold for statistical significance was P < .05.

Main results. Of the 27 665 inpatient procedures identified across 5 hospitals, 444 (1.6%) cases met the inclusion criteria. Patients from these cases aged 75 (SD 13) years (95% CI, 72-77 years); 247 (56%, 95% CI, 55%-57%) were women; and 300 (68%, 95% CI, 65%-71%) were White. A total of 426 patients (96%, 95% CI, 90%-100%) had an ASA physical status score of 3 or higher and 237 (53%, 95% CI, 51%-56%) received general anesthesia. The most common procedures performed were endoscopy (148 [33%]), hip fracture repair (43 [10%]), and gastrostomy or jejunostomy (28 [6%]). Reevaluation of code status was documented in 126 cases (28%, 95% CI, 25%-31%); code status orders were changed in 20 of 126 cases (16%, 95% CI, 7%-24%); and a note was filed without a corresponding order for 106 of 126 cases (84%, 95% CI, 75%-95%). In the majority of cases (109 of 126 [87%], 95% CI, 78%-95%) in which documented discussion occurred, DNR orders were suspended. Of 126 cases in which a discussion was documented, participants of these discussions included surgeons 10% of the time (13 cases, 95% CI, 8%-13%), members of the anesthesia team 51% of the time (64 cases, 95% CI, 49%-53%), and medicine or palliative care clinicians 39% of the time (49 cases, 95% CI, 37%-41%).

The rate of documented preoperative code status discussion was higher in patients with higher ASA physical status score (35% in patients with an ASA physical status score ≥ 4 [55 of 155] vs 25% in those with an ASA physical status score ≤ 3 [71 of 289]; P = .02). The rates of documented preoperative code status discussion were similar by anesthesia type (29% for general anesthesia [69 of 237 cases] vs 28% [57 of 207 cases] for other modalities; P = .70). The hospitals involved in this study all had a written policy encouraging rediscussion of code status before surgery. However, only 1 hospital reported added measures (eg, provision of a procedure-specific DNR form) to increase documentation of preoperative code status discussions. In this specific hospital, documentation of preoperative code status discussions was higher compared to other hospitals (67% [37 of 55 cases] vs 23% [89 of 389 cases]; P < .01).

Conclusion. In a retrospective case series conducted at 5 hospitals within 1 academic health system in the US, fewer than 1 in 5 patients with preexisting DNR orders had a documented discussion of code status prior to undergoing surgery. Additional strategies including the development of institutional protocols that facilitate perioperative management of advance directives, identification of local champions, and patient education, should be explored as means to improve preoperative code status reevaulation per guideline recommendations.

It is not unusual that patients with a DNR order may require and undergo surgical interventions to treat reversible conditions, prevent progression of underlying disease, or mitigate distressing symptoms such as pain. For instance, intubation, mechanical ventilation, and administration of vasoactive drugs are resuscitative measures that may be needed to safely anesthetize and sedate a patient. As such, the American College of Surgeons¹ has provided a statement on advance directives by patients with an existing DNR order to guide management. Specifically, the statement indicates that the best approach for these patients is a policy of “required reconsideration” of the existing DNR order. Required reconsideration means that “the patient or designated surrogate and the physicians who will be responsible for the patient’s care should, when possible, discuss the new intraoperative and perioperative risks associated with the surgical procedure, the patient’s treatment goals, and an approach for potentially life-threatening problems consistent with the patient’s values and preferences.” Moreover, the required reconsideration discussion needs to occur as early as it is practical once a decision is made to have surgery because the discussion “may result in the patient agreeing to suspend the DNR order during surgery and the perioperative period, retaining the original DNR order, or modifying the DNR order.” Given that surgical patients with DNR orders have significant comorbidities, many sustain postoperative complications, and nearly 1 in 4 die within 30 days of surgery, preoperative advance care planning (ACP) and code status discussions are particularly essential to delivering high quality surgical care.²

In the current study, Hadler et al³ conducted a retrospective analysis to evaluate orders and documentation describing perioperative management of code status in patients with existing DNR order at an academic health system in the US. The authors reported that fewer than 20% of patients with existing DNR orders had a documented discussion of code status prior to undergoing surgery. These findings add to the notion that compliance with such guidance on required reconsideration discussion is suboptimal in perioperative care in the US.^4,5 A recently published study focused on patients aged more than 60 years undergoing high-risk oncologic or vascular surgeries similarly showed that the frequency of ACP discussions or advance directive documentations among older patients was low.⁶ This growing body of evidence is highly clinically relevant in that preoperative discussion on code status is highly relevant to the care of older adults, a population group that accounts for the majority of surgeries and is most vulnerable to poor surgical outcomes. Additionally, it highlights a disconnect between the shared recognition by surgeons and patients that ACP discussion is important in perioperative care and its low implementation rates.

Unsurprisingly, Hadler et al³ reported that added measures such as the provision of a procedure-specific DNR form led to an increase in the documentation of preoperative code status discussions in 1 of the hospitals studied. The authors suggested that strategies such as the development of institutional protocols aimed to facilitate perioperative advance directive discussions, identify local champions, and educate patients may be ways to improve preoperative code status reevaulation. The idea that institutional value and culture are key factors impacting surgeon behavior and may influence the practice of ACP discussion is not new. Thus, creative and adaptable strategies, resources, and trainings that are required by medical institutions and hospitals to support preoperative ACP discussions with patients undergoing surgeries need to be identified, validated, and implemented to optimize perioperative care in vulnerable patients.

Applications for Clinical Practice

The findings from the current study indicate that less than 20% of patients with preexisting DNR orders have a documented discussion of code status prior to undergoing surgery. Physicians and health care institutions need to identify barriers to, and implement strategies that, facilitate and optimize preoperative ACP discussions in order to provide patient-centered care in vulnerable surgical patients.

Financial disclosures: None.

Study Overview

Objective. To determine whether fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) of nonculprit lesion in patients with ST-segment elevation myocardial infarction (STEMI) is superior to angiography-guided PCI.

Design. Multicenter randomized control trial blinded to outcome, conducted in 41 sites in France.

Setting and participants. A total of 1163 patients with STEMI and multivessel coronary disease, who had undergone successful PCI to the culprit lesion were randomized to either FFR-guided PCI or angiography-guided PCI for nonculprit lesions. Randomization was stratified according to the trial site and timing of the procedure (immediate or staged).

Main outcome measures. The primary outcome was a composite of death from any cause, nonfatal myocardial infarction (MI) or unplanned hospitalization leading to urgent revascularization at 1 year.

Main results. At 1 year, the primary outcome occurred in 32 of 586 patients (5.5%) in the FFR-guided group and in 24 of 577 (4.2%) in the angiography-guided group (hazard ratio [HR], 1.32; 95% CI, 0.78-2.23; P = .31). The rate of death (1.5% vs 1.7%), nonfatal MI (3.1% vs 1.7%), and unplanned hospitalization leading to urgent revascularization (3.1% vs 1.7%) were also similar between FFR-guided and angiography-guided groups.

Conclusion. Among patients with STEMI and multivessel disease who had undergone successful PCI of the culprit vessel, an FFR-guided strategy for complete revascularization was not superior to angiography-guided strategy for reducing death, MI, or urgent revascularization at 1 year.

Commentary

Patients presenting with STEMI often have multivessel disease.¹ Recently, multiple studies have reported the benefit of nonculprit vessel revascularization in patients presenting with hemodynamically stable STEMI compared to culprit-only strategy including the most recent COMPLETE trial which showed reduction in death and MI.^2-6 However, the previous studies have variable design in evaluating the nonculprit vessel, some utilized FFR guidance, while others used angiography guidance. Whether FFR-guided PCI of nonculprit vessel can improve outcome in patients presenting STEMI remains unknown.

In the FLOWER-MI study, Puymirat et al investigated the use of FFR compared to angiography-guided nonculprit vessel PCI. A total of 1163 patients presenting with STEMI and multivessel disease who had undergone successful PCI to the culprit vessel, were randomized to either FFR guidance or angiography guidance among 41 centers in France. The authors found that after 1 year, there was no difference in composite endpoint of death, nonfatal MI or unplanned hospitalization leading to urgent revascularization in the FFR-guided group compared to angiography-guided group (5.5% vs 4.2%, HR, 1.32; 95% CI, 0.678-2.23; P = .31). There was also no difference in individual components of primary outcomes or secondary outcomes such as rate of stent thrombosis, any revascularization, or hospitalization.

There are a few interesting points to consider in this study. Ever since the Fractional Flow Reserve vs Angiography for Multivessel Evaluation (FAME) trial reported the lower incidence of major adverse events in routine FFR measurement during PCI compared to angiography-guided PCI, physiological assessment has become the gold standard for treatment of stable ischemic heart disease.⁷ However, the results of the current FLOWER-MI trial were not consistent with the FAME trial and there are few possible reasons to consider.

First, the use of FFR in the setting of STEMI is less validated compared to stable ischemic heart disease.⁸ Microvascular dysfunction during the acute phase can affect the FFR reading and the lesion severity can be underestimated.⁸ Second, the rate of composite endpoint was much lower in this study compared to FAME despite using the same composite endpoint of death, nonfatal MI, and unplanned hospitalization leading to urgent revascularization. At 1 year, the incidence of primary outcome was 13.5% in the FFR-guided group compared to 18.6% in the angiography-guided group in the FAME study compared to 5.5% and 4.2% in the FLOWER-MI study, despite having a sicker population presenting with STEMI. This is likely due to improvement in the PCI techniques such as radial approach, imaging guidance, and advancement in medical therapy such as use of more potent antiplatelet therapy. With lower incidence of primary outcome, larger number of patients are needed to detect the difference in the composite outcome. Finally, the operators’ visual assessment may have been calibrated to the physiologic assessment as the operators are routinely using FFR assessment which may have diminished the benefit of FFR guidance seen in the early FAME study.

Another interesting finding from this study was that although the study protocol encouraged the operators to perform the nonculprit PCI in the same setting, only 4% had nonculprit PCI in the same setting and 96% of the patients underwent a staged PCI. The advantage of performing the nonculprit PCI on the same setting is to have 1 fewer procedure for the patient. On the other hand, the disadvantage of this approach includes prolongation of the index procedure, theoretically higher risk of complication during the acute phase and vasospasm leading to overestimation of the lesion severity. A recent analysis from the COMPLETE study did not show any difference when comparing staged PCI during the index hospitalization vs after discharge.⁹ The optimal timing of the staged PCI needs to be investigated in future studies.

A limitation of this study is the lower than expected incidence of clinical events decreasing the statistical power of the study. However, there was no signal that FFR-guided PCI is better compared to the angiography-guided group. In fact, the curve started to diverge at 6 months favoring the angiography-guided group. In addition, there was no core-lab analysis for completeness of revascularization.

Applications for Clinical Practice

In patients presenting with hemodynamically stable STEMI for undergoing nonculprit vessel PCI, both FFR-guided or angiography-guided strategies can be considered.

Financial disclosures: None.

Study Overview

Objective. To determine whether fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) of nonculprit lesion in patients with ST-segment elevation myocardial infarction (STEMI) is superior to angiography-guided PCI.

Design. Multicenter randomized control trial blinded to outcome, conducted in 41 sites in France.

Setting and participants. A total of 1163 patients with STEMI and multivessel coronary disease, who had undergone successful PCI to the culprit lesion were randomized to either FFR-guided PCI or angiography-guided PCI for nonculprit lesions. Randomization was stratified according to the trial site and timing of the procedure (immediate or staged).

Main outcome measures. The primary outcome was a composite of death from any cause, nonfatal myocardial infarction (MI) or unplanned hospitalization leading to urgent revascularization at 1 year.

Main results. At 1 year, the primary outcome occurred in 32 of 586 patients (5.5%) in the FFR-guided group and in 24 of 577 (4.2%) in the angiography-guided group (hazard ratio [HR], 1.32; 95% CI, 0.78-2.23; P = .31). The rate of death (1.5% vs 1.7%), nonfatal MI (3.1% vs 1.7%), and unplanned hospitalization leading to urgent revascularization (3.1% vs 1.7%) were also similar between FFR-guided and angiography-guided groups.

Conclusion. Among patients with STEMI and multivessel disease who had undergone successful PCI of the culprit vessel, an FFR-guided strategy for complete revascularization was not superior to angiography-guided strategy for reducing death, MI, or urgent revascularization at 1 year.

Commentary

Patients presenting with STEMI often have multivessel disease.¹ Recently, multiple studies have reported the benefit of nonculprit vessel revascularization in patients presenting with hemodynamically stable STEMI compared to culprit-only strategy including the most recent COMPLETE trial which showed reduction in death and MI.^2-6 However, the previous studies have variable design in evaluating the nonculprit vessel, some utilized FFR guidance, while others used angiography guidance. Whether FFR-guided PCI of nonculprit vessel can improve outcome in patients presenting STEMI remains unknown.

In the FLOWER-MI study, Puymirat et al investigated the use of FFR compared to angiography-guided nonculprit vessel PCI. A total of 1163 patients presenting with STEMI and multivessel disease who had undergone successful PCI to the culprit vessel, were randomized to either FFR guidance or angiography guidance among 41 centers in France. The authors found that after 1 year, there was no difference in composite endpoint of death, nonfatal MI or unplanned hospitalization leading to urgent revascularization in the FFR-guided group compared to angiography-guided group (5.5% vs 4.2%, HR, 1.32; 95% CI, 0.678-2.23; P = .31). There was also no difference in individual components of primary outcomes or secondary outcomes such as rate of stent thrombosis, any revascularization, or hospitalization.

There are a few interesting points to consider in this study. Ever since the Fractional Flow Reserve vs Angiography for Multivessel Evaluation (FAME) trial reported the lower incidence of major adverse events in routine FFR measurement during PCI compared to angiography-guided PCI, physiological assessment has become the gold standard for treatment of stable ischemic heart disease.⁷ However, the results of the current FLOWER-MI trial were not consistent with the FAME trial and there are few possible reasons to consider.

First, the use of FFR in the setting of STEMI is less validated compared to stable ischemic heart disease.⁸ Microvascular dysfunction during the acute phase can affect the FFR reading and the lesion severity can be underestimated.⁸ Second, the rate of composite endpoint was much lower in this study compared to FAME despite using the same composite endpoint of death, nonfatal MI, and unplanned hospitalization leading to urgent revascularization. At 1 year, the incidence of primary outcome was 13.5% in the FFR-guided group compared to 18.6% in the angiography-guided group in the FAME study compared to 5.5% and 4.2% in the FLOWER-MI study, despite having a sicker population presenting with STEMI. This is likely due to improvement in the PCI techniques such as radial approach, imaging guidance, and advancement in medical therapy such as use of more potent antiplatelet therapy. With lower incidence of primary outcome, larger number of patients are needed to detect the difference in the composite outcome. Finally, the operators’ visual assessment may have been calibrated to the physiologic assessment as the operators are routinely using FFR assessment which may have diminished the benefit of FFR guidance seen in the early FAME study.

Another interesting finding from this study was that although the study protocol encouraged the operators to perform the nonculprit PCI in the same setting, only 4% had nonculprit PCI in the same setting and 96% of the patients underwent a staged PCI. The advantage of performing the nonculprit PCI on the same setting is to have 1 fewer procedure for the patient. On the other hand, the disadvantage of this approach includes prolongation of the index procedure, theoretically higher risk of complication during the acute phase and vasospasm leading to overestimation of the lesion severity. A recent analysis from the COMPLETE study did not show any difference when comparing staged PCI during the index hospitalization vs after discharge.⁹ The optimal timing of the staged PCI needs to be investigated in future studies.

A limitation of this study is the lower than expected incidence of clinical events decreasing the statistical power of the study. However, there was no signal that FFR-guided PCI is better compared to the angiography-guided group. In fact, the curve started to diverge at 6 months favoring the angiography-guided group. In addition, there was no core-lab analysis for completeness of revascularization.

Applications for Clinical Practice

In patients presenting with hemodynamically stable STEMI for undergoing nonculprit vessel PCI, both FFR-guided or angiography-guided strategies can be considered.

Financial disclosures: None.

Study Overview

Objective. To determine whether fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) of nonculprit lesion in patients with ST-segment elevation myocardial infarction (STEMI) is superior to angiography-guided PCI.

Design. Multicenter randomized control trial blinded to outcome, conducted in 41 sites in France.

Setting and participants. A total of 1163 patients with STEMI and multivessel coronary disease, who had undergone successful PCI to the culprit lesion were randomized to either FFR-guided PCI or angiography-guided PCI for nonculprit lesions. Randomization was stratified according to the trial site and timing of the procedure (immediate or staged).

Main outcome measures. The primary outcome was a composite of death from any cause, nonfatal myocardial infarction (MI) or unplanned hospitalization leading to urgent revascularization at 1 year.

Main results. At 1 year, the primary outcome occurred in 32 of 586 patients (5.5%) in the FFR-guided group and in 24 of 577 (4.2%) in the angiography-guided group (hazard ratio [HR], 1.32; 95% CI, 0.78-2.23; P = .31). The rate of death (1.5% vs 1.7%), nonfatal MI (3.1% vs 1.7%), and unplanned hospitalization leading to urgent revascularization (3.1% vs 1.7%) were also similar between FFR-guided and angiography-guided groups.

Conclusion. Among patients with STEMI and multivessel disease who had undergone successful PCI of the culprit vessel, an FFR-guided strategy for complete revascularization was not superior to angiography-guided strategy for reducing death, MI, or urgent revascularization at 1 year.

Commentary

Patients presenting with STEMI often have multivessel disease.¹ Recently, multiple studies have reported the benefit of nonculprit vessel revascularization in patients presenting with hemodynamically stable STEMI compared to culprit-only strategy including the most recent COMPLETE trial which showed reduction in death and MI.^2-6 However, the previous studies have variable design in evaluating the nonculprit vessel, some utilized FFR guidance, while others used angiography guidance. Whether FFR-guided PCI of nonculprit vessel can improve outcome in patients presenting STEMI remains unknown.

In the FLOWER-MI study, Puymirat et al investigated the use of FFR compared to angiography-guided nonculprit vessel PCI. A total of 1163 patients presenting with STEMI and multivessel disease who had undergone successful PCI to the culprit vessel, were randomized to either FFR guidance or angiography guidance among 41 centers in France. The authors found that after 1 year, there was no difference in composite endpoint of death, nonfatal MI or unplanned hospitalization leading to urgent revascularization in the FFR-guided group compared to angiography-guided group (5.5% vs 4.2%, HR, 1.32; 95% CI, 0.678-2.23; P = .31). There was also no difference in individual components of primary outcomes or secondary outcomes such as rate of stent thrombosis, any revascularization, or hospitalization.

There are a few interesting points to consider in this study. Ever since the Fractional Flow Reserve vs Angiography for Multivessel Evaluation (FAME) trial reported the lower incidence of major adverse events in routine FFR measurement during PCI compared to angiography-guided PCI, physiological assessment has become the gold standard for treatment of stable ischemic heart disease.⁷ However, the results of the current FLOWER-MI trial were not consistent with the FAME trial and there are few possible reasons to consider.

First, the use of FFR in the setting of STEMI is less validated compared to stable ischemic heart disease.⁸ Microvascular dysfunction during the acute phase can affect the FFR reading and the lesion severity can be underestimated.⁸ Second, the rate of composite endpoint was much lower in this study compared to FAME despite using the same composite endpoint of death, nonfatal MI, and unplanned hospitalization leading to urgent revascularization. At 1 year, the incidence of primary outcome was 13.5% in the FFR-guided group compared to 18.6% in the angiography-guided group in the FAME study compared to 5.5% and 4.2% in the FLOWER-MI study, despite having a sicker population presenting with STEMI. This is likely due to improvement in the PCI techniques such as radial approach, imaging guidance, and advancement in medical therapy such as use of more potent antiplatelet therapy. With lower incidence of primary outcome, larger number of patients are needed to detect the difference in the composite outcome. Finally, the operators’ visual assessment may have been calibrated to the physiologic assessment as the operators are routinely using FFR assessment which may have diminished the benefit of FFR guidance seen in the early FAME study.

Another interesting finding from this study was that although the study protocol encouraged the operators to perform the nonculprit PCI in the same setting, only 4% had nonculprit PCI in the same setting and 96% of the patients underwent a staged PCI. The advantage of performing the nonculprit PCI on the same setting is to have 1 fewer procedure for the patient. On the other hand, the disadvantage of this approach includes prolongation of the index procedure, theoretically higher risk of complication during the acute phase and vasospasm leading to overestimation of the lesion severity. A recent analysis from the COMPLETE study did not show any difference when comparing staged PCI during the index hospitalization vs after discharge.⁹ The optimal timing of the staged PCI needs to be investigated in future studies.

A limitation of this study is the lower than expected incidence of clinical events decreasing the statistical power of the study. However, there was no signal that FFR-guided PCI is better compared to the angiography-guided group. In fact, the curve started to diverge at 6 months favoring the angiography-guided group. In addition, there was no core-lab analysis for completeness of revascularization.

Applications for Clinical Practice

In patients presenting with hemodynamically stable STEMI for undergoing nonculprit vessel PCI, both FFR-guided or angiography-guided strategies can be considered.

Financial disclosures: None.

From Illawarra Shoalhaven Local Health District, New South Wales, Australia (Maren Jones, Dr. Hewitt, Philippa King, Rhiannon Thorn, Edward Davidson, and Tiana-Lee Elphick), and Illawarra Health and Medical Research Institute, University of Wollongong, Wollongong, New South Wales, Australia (Dr. Hewitt)

Objective: To assess the association between change scores in the Functional Independence Measure (FIM) with evaluative measures used in physiotherapy to objectively show that use of the FIM in isolation is limited.

Design: Retrospective observational study.

Setting: Five rehabilitation inpatient wards from 1 public local health district in NSW Australia.

Participants: Patient data over a 5-year time frame (2015 to 2019) were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups (Australasian Rehabilitation Outcome Centre classification) were identified for inclusion in this study: Reconditioning (n = 742, mean age 76.88 years); Orthopedic Fracture (n = 585, mean age 77.46 years); and Orthopedic Replacement (n = 377, mean age 73.84 years).

Measurements: The difference between the admission and discharge scores were calculated for each measure. Kruskal-Wallis and χ² tests were used to analyze the data.

Results: Pearson correlation (r) coefficients between FIM Motor change to the de Morton’s Mobility Index (DEMMI) change was r = 0.396, FIM Motor change to the Timed Up and Go (TUG) change was r = -0.217, and the FIM Motor change to the Ten Meter Walk Test (10MWT) change was .194.

Conclusion: The FIM Motor change scores showed a weak positive association to the DEMMI change and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, functional mobility, and dynamic balance.

Keywords: physiotherapy; rehabilitation; clinical outcome measures.

Patients receive interdisciplinary inpatient rehabilitation treatment after they have sustained a lower limb fracture, a lower limb joint replacement, or have generalized deconditioning (muscle wasting and disuse atrophy) following hospitalization for surgery or illness. The degree of a patient’s impairment or loss of functional capacity, as well as their ability to manage at home safely, is assessed using standardized outcome measures during their recovery and rehabilitation.^1,2

Physiotherapists routinely use validated outcome measures to assess patient progress and to measure goal attainment through assessment of functional independence, dynamic balance performance, and ambulatory ability. These objective assessments provide clinicians with information about the effectiveness of the rehabilitation program, as well as the patient’s ability to manage in their home environment, to determine the need for assistive devices, level of caregiver support, future level of autonomy, and strategies for falls prevention.^3-7

There is a view among service providers that rehabilitation decisions can be based on a singular measure of function known as the Functional Independence Measure (FIM). This is an understandable position because not only is the FIM an internationally recognized, valid, and reliable tool, but, as a singular measure, it also means measurement consistency across rehabilitation sites is more likely. However, rehabilitation is complex, and it is risky to base decisions on a single measure, which might not capture the results of rehabilitation treatment ingredients on individual patient targets.^8,9

The patient’s progress is objectively assessed using functional outcome measures such as the FIM. Other measures used typically in our service include the de Morton’s Mobility Index (DEMMI), Timed Up and Go (TUG), and the Ten Meter Walk Test (10MWT), which measure patient mobility, balance during directional changes, and walking ability, respectively. Additional measures include patient progression to a less supportive level of assistance (ie, number of persons required to assist or level of supervision) or the selection of a walking aid (eg, forearm support frame, crutches). This progression—or lack thereof—assists in decision-making regarding the individual’s future once they are discharged from rehabilitation. Such considerations would include the need to modify the home environment, selection of assistive devices, community access (walking indoors, outdoors, and shopping), personal care needs, and age-appropriate care facility recommendations (ie, level of care). The use of outcome measures also indicates the need for further referrals to other care providers upon discharge from the rehabilitation facility.

There is widespread support in the literature for the use of the FIM, DEMMI, TUG, and 10MWT in rehabilitation population groups. For example, DEMMI has been validated in hip fracture patients during rehabilitation,¹⁰ as well as among older people hospitalized for medical illness.^11-13 It has also been shown to be a predictor of discharge destination for patients living with frailty in geriatric rehabilitation settings,¹⁴ and to have moderate predictive validity for functional independence after 4 weeks of rehabilitation.¹⁵ Similarly, TUG has been validated for use among hospitalized and community-dwelling individuals,^16-18 and for patients after joint arthroplasty^19,20 or hip fracture.²¹ It has also been shown to be an indicator of fall risk,^22-24 as well as a predictor of fracture incidence.²⁵ Furthermore, TUG has been identified as an indicator of a patient’s ability to walk in the community without the need for a walking device.²⁶ It has also been shown to be an early identifier of patients in need of rehabilitation.²⁷ Normative values for TUG have been reported, and the association with gait time established.²⁸

Gait speed has been shown to predict adverse outcomes in community-dwelling older people.²⁹ In fact, the 10MWT has been established as a powerful tool to benchmark rehabilitation recovery after a medical event.³⁰ Results of the test relate to overall quality of walking, health status, morbidity, and the rate of mortality.^31-33 Meaningful improvement, minimum detectable change (0.19-0.34 m/s), and responsiveness in common physical performance in older adults has been reported.^26,34,36

Structural and functional impairment has been used to define rehabilitation classes by the Australasian Rehabilitation Outcome Centre (AROC) in the Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP) Version 4.^37-43 Variables used for grouping are age, care type, function, and impairment for rehabilitation. FIM was developed in order to assess patients’ outcomes after inpatient multidisciplinary care, and is an internationally accepted measure of functioning.⁴⁴ It is a holistic outcome measure, which can be used to determine the patient’s level of disability and burden of care, and is widely used in both public and private inpatient rehabilitation settings. Each patient classification is reported separately within the case mix structure.⁴⁵ Inpatient rehabilitation centers are evaluated and compared by the AROC,⁴⁶ with an emphasis on length of stay and the FIM change. The most successful centers demonstrate shorter length of stay and greater FIM improvement. Although the FIM is a valuable measure, it does not provide a complete picture of the individual patient’s rehabilitation gain: ie, the specific attributes of patients’ mobility, walking ability, or balance during directional changes.

A large-scale analysis of the association between the holistic disability measure of the FIM and the more mobility- and ambulation-focused physiotherapy outcomes has not been documented.

The well-documented DEMMI accumulates points for the patient’s mobility in a similar fashion to the FIM, but with more mobility detail. These 2 outcome measures allow for the full range of patients, from the very dependent up to and including the independently ambulant patients. The DEMMI may show a positive relationship to the FIM, yet the association is unknown. The association of the TUG to the 10MWT has been established²⁸; however, their relationship to the FIM is unknown.

Current practice in the participating public health inpatient rehabilitation wards is to use the DEMMI, TUG, 10MWT, and FIM to ensure physiotherapy and allow the wider multidisciplinary team to more effectively evaluate patient mobility outcomes. The 3 most frequent patient groups identified within the current patient population are expected to present clinical differences and will be analyzed for comparison. If an association is found between the outcome measures in question, clinical efficiency could be improved.

The aim of the current study is to assess the association between change scores in the FIM with evaluative measures of outcomes typically used in physiotherapy to objectively show that use of the FIM in isolation is limited in our population of patients.

Methods

Study design and setting

This retrospective descriptive observational study complied with the STROBE-RECORD guidance and checklist (available at mdedge.com/jcomjournal) and analyzed the routinely collected data from rehabilitation patients who were admitted to 5 different rehabilitation wards in 4 different public hospitals from 1 regional local health district (20-24 beds per ward) from 2015 to 2019. As this study conducted secondary analyses using existing de-identified data from a public health facility and did not involve interaction with any human subjects, ethical approval was not required.⁴⁶ Approval to conduct this study was granted by the health district’s institutional review committee, as per the National Statement on Ethical Conduct in Human Research 2015.

Participants

Patient data over a 5-year time frame were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups were identified for inclusion in this study: reconditioning, orthopedic fracture, and orthopedic replacement. (See Table 1 for the specific AN-SNAP impairment groups used in this study.)

Patient data from the less-frequent impairment groups were excluded (n = 673, 28.19%), including stroke (n = 343), brain dysfunction (n = 45), amputation of limb (n = 45), spinal cord dysfunction (n  = 36), neurological dysfunction (n = 34), cardiac (n = 24), and others (n = 25) who may have benefitted from other outcome measures due to their medical condition. Ten patient data sets were excluded for missing discharge outcome measure data, from when the patient became ill and returned to acute services or was discharged at short notice. To be included in the study, both the admission and discharge scores from the FIM and the admission and discharge scores from at least 1 of the physiotherapy outcome measures were required for each patient (n = 1704, 71.39%): Reconditioning (n = 742), Orthopedic Fracture (n = 585), and Orthopedic Replacement (n = 377). Information regarding the type of walking aid and the amount of assistance required for safe ambulation was also recorded. These items were included in the study’s descriptive analysis. Only 1.7% of these descriptors were missing.

Outcome measures

DEMMI tasks of bed mobility, sitting balance, transfers, walking, and balance were scored with an assigned value according to the patient’s performance. This was then tallied and the results scaled, to provide an overall score out of 100 available points. The total score from admission and discharge was then compared. Improvement (change) was identified by the increase in scores.

The TUG assesses a patient’s dynamic balance performance.⁴⁷ The number of seconds it took the patient to complete the procedure was recorded at admission and discharge. Improvement (change) was identified by the reduction in time taken at discharge from the admission score.

The 10MWT measures the unidirectional walking speed of a person over 10 meters and is recorded in seconds and reported in meters per second. Improvement (change) was identified by the reduction in the time taken to increase walking speed.

Concurrent to the physiotherapy measures were the FIM scores, recorded by the accredited nursing staff from each rehabilitation ward. Improvement is demonstrated by the accumulation of points on the ordinal scale of the FIM Total, including mobility, dressing, bladder and bowel care, cognition, and social interaction, and is represented as a score between 18 and 126. The FIM Motor category is reported as a score between 13 and 91.

The 2 data sets were matched by unique identifier and admission dates, then de-identified for analysis.

Statistical analysis

Patient demographic information was analyzed using descriptive statistics (mean, SD, frequencies, percentages) for each impairment group (orthopedic fracture, orthopedic replacement, reconditioning). Differences in continuous demographic variables for each impairment group were assessed using Kruskal-Wallis tests and χ² tests for categorical variables. Functional outcome scores were compared at admission, discharge, and change between the impairment groups. Association of the functional outcome change scores was determined with the Pearson correlation coefficient (r) between the FIM and the DEMMI, TUG, and 10MWT. Graphs were plotted for each of these (Figure available online at mdedge.com/jcomjournal). A strong, moderate, and weak association was described as > 0.6, > 0.4, and > 0.2, respectively.⁴⁶ Statistical significance was set at P < .05. Analyses were conducted using Stata (StataCorp LLC, USA).

Results

The patient descriptive data (site from which data were collected, admission length of stay, age at admission, discharge destination, walk aid improvement, and walk assistance improvement) from the 3 impairment groups are reported in Table 2. The functional outcomes for DEMMI, TUG, 10MWT, FIM Motor, FIM Total at admission, discharge, and the change scores are presented in Table 3.

Orthopedic fracture patients had the greatest improvement in their functional outcomes, with a DEMMI improvement of 18 points, TUG score change of 23.49 seconds (s), 10MWT change of 0.30 meters/second (m/s), FIM Motor change of 20.62, and a FIM Total change of 21.9 points. The outcome measures exceeded the minimum detectable change as reported in the literature for DEMMI (8.8 points⁴⁸), TUG (2.08 s²⁶), walking speed 0.19 m/s²⁶, and FIM Motor (14.6 points⁴⁹).

Association of functional outcomes (change scores)

There was a significant weak positive correlation between DEMMI change score and both the FIM Motor (r = 0.396) and FIM Total change scores (r = 0.373). When viewing the specific items within the FIM Motor labelled FIM Walk change, FIM MobilityBedChair change, and FIM stairs change, r values were 0.100, 0.379, and 0.126, respectively. In addition, there was a weak negative correlation between TUG change scores and both FIM Motor (r = -0.217) and FIM Total change scores (r = -0.207). There was a very weak positive correlation between 10MWT (m/s) change scores and both FIM Motor (r = 0.194) and FIM Total change scores (r = 0.187) (Table 4, Figure). There was a moderate correlation between 10MWT change (s) and TUG change (s) (r = 0.72, P < .001).

Discussion

The purpose of this study was to ascertain the association between the DEMMI, TUG, 10MWT, and FIM measures using retrospective data collected from 5 public hospital inpatient rehabilitation wards. The results of this retrospective analysis demonstrate that a variety of objective outcome measures are required for the multidisciplinary team to accurately measure a patient’s functional improvement during their inpatient rehabilitation stay. No single outcome measure in this study fully reported all mobility attributes, and we note the risk of basing decisions on a single measure evaluating rehabilitation outcomes. Although the internationally used FIM has a strong place in rehabilitation reporting and benchmarking, it does not predict change nor provide a proxy for the patient’s whole-body motor control as they extend their mobility, dynamic balance, and ambulatory ability. Multiple objective outcome measures should therefore be required to evaluate the patient’s progress and functional performance toward discharge planning.

The FIM is a measure of disability or care needs, incorporating cognitive, social, and physical components of disability. It is a valid, holistic measure of an individual’s functional ability at a given time. Rehabilitation sites internationally utilize this assessment tool to evaluate a patient’s progress and the efficacy of intervention. The strength of this measure is its widespread use and the inclusion of the personal activities of daily living to provide an overall evaluation encompassing all aspects of a person’s ability to function independently. However, as our study results suggest, patient improvement measured by the FIM Motor components were not correlated to other widely used physiotherapy measures of ambulation and balance, such as the 10MWT or TUG. This is perhaps largely because the FIM Motor components only consider the level of assistance (eg, physical assistance, assistive device, independence) and do not consider assessment of balance and gait ability as assessed in the 10MWT and TUG. The 10MWT and TUG provide assessment of velocity and dynamic balance during walking, which have been shown to predict an individual’s risk of falling.^22,23 This is a pertinent issue in the rehabilitation and geriatric population.²⁹ Furthermore, the use of the FIM as a benchmarking tool to compare facility efficiency may not provide a complete assessment of all outcomes achieved on the inpatient rehabilitation ward, such as reduced falls risk or improved ambulatory ability and balance.

It has previously been confirmed that there is a significant positive correlation between the 10MWT and the TUG.²⁷ This retrospective analysis has also supported these findings. This is possibly due to the similarity in the assessments, as they both incorporate ambulation ability and dynamic movement.

Each of the 4 outcome measures assess different yet vital aspects of an individual’s functional mobility and ambulation ability during their subacute rehabilitation journey. The diversity of patient age, functional impairment, and mobility level needs a range of outcomes to provide baselines, targets, and goal attainment for discharge home.

Consistent with the AROC AN-SNAP reporting of Length of Stay and FIM change separated into the weighted impairment groups, the data analysis of this study demonstrated significant differences between the Reconditioning, Orthopedic Fracture, and Orthopedic Replacement patient data. Tables 2 and 3 describe the differences between the groups. The fracture population in this study improved the most across each outcome measure. In contrast, the reconditioning population showed the least improvement. This may be expected due to the pathophysiological differences between the groups. Furthermore, for the elderly who sustain fractures because of a fall, rehabilitation will be required to address not only the presenting injury but also the premorbid falls risk factors which may include polypharmacy or impaired balance.

Any conclusions drawn from the findings of this study need to take into consideration that it has focused on patients from 1 local health district and therefore may not be generalizable to a wider national or international context. As this study was a retrospective study, controlling for data collection quality, measurement bias due to nonblinding and missing data is a limitation. However, clinicians regularly completed these outcome assessments and recorded this information as part of their standard care practices within this health district. There may have been slight differences in definitions of practice between the 5 rehabilitation sites. To ensure reliability, each individual site’s protocols for the FIM, DEMMI, TUG, and 10MWT were reviewed and confirmed to be consistent.

Conclusion

The FIM Motor change scores showed a weak positive association to the DEMMI change, and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. Thorough reporting of clinical outcomes is much more meaningful to assess and guide the physiotherapy component of rehabilitation. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, mobility, functional capacity, and dynamic balance.

Acknowledgements: The authors thank Anne Smith, MSHLM, BAppSc, Head of the Physiotherapy Department, and the physiotherapists and allied health assistants who have contributed to the collection of this valuable data over several years. They also thank Lina Baytieh, MS, BS, from Research Central, Illawarra Shoalhaven Local Health District, for her assistance with the analysis.

Corresponding author: Maren Jones, MPH, BS, Physiotherapy Department, Port Kembla Hospital, Illawarra Shoalhaven Local Health District, Warrawong, New South Wales, 2505 Australia; [email protected].

Financial disclosures: None.

From Illawarra Shoalhaven Local Health District, New South Wales, Australia (Maren Jones, Dr. Hewitt, Philippa King, Rhiannon Thorn, Edward Davidson, and Tiana-Lee Elphick), and Illawarra Health and Medical Research Institute, University of Wollongong, Wollongong, New South Wales, Australia (Dr. Hewitt)

Objective: To assess the association between change scores in the Functional Independence Measure (FIM) with evaluative measures used in physiotherapy to objectively show that use of the FIM in isolation is limited.

Design: Retrospective observational study.

Setting: Five rehabilitation inpatient wards from 1 public local health district in NSW Australia.

Participants: Patient data over a 5-year time frame (2015 to 2019) were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups (Australasian Rehabilitation Outcome Centre classification) were identified for inclusion in this study: Reconditioning (n = 742, mean age 76.88 years); Orthopedic Fracture (n = 585, mean age 77.46 years); and Orthopedic Replacement (n = 377, mean age 73.84 years).

Measurements: The difference between the admission and discharge scores were calculated for each measure. Kruskal-Wallis and χ² tests were used to analyze the data.

Results: Pearson correlation (r) coefficients between FIM Motor change to the de Morton’s Mobility Index (DEMMI) change was r = 0.396, FIM Motor change to the Timed Up and Go (TUG) change was r = -0.217, and the FIM Motor change to the Ten Meter Walk Test (10MWT) change was .194.

Conclusion: The FIM Motor change scores showed a weak positive association to the DEMMI change and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, functional mobility, and dynamic balance.

Keywords: physiotherapy; rehabilitation; clinical outcome measures.

Patients receive interdisciplinary inpatient rehabilitation treatment after they have sustained a lower limb fracture, a lower limb joint replacement, or have generalized deconditioning (muscle wasting and disuse atrophy) following hospitalization for surgery or illness. The degree of a patient’s impairment or loss of functional capacity, as well as their ability to manage at home safely, is assessed using standardized outcome measures during their recovery and rehabilitation.^1,2

Physiotherapists routinely use validated outcome measures to assess patient progress and to measure goal attainment through assessment of functional independence, dynamic balance performance, and ambulatory ability. These objective assessments provide clinicians with information about the effectiveness of the rehabilitation program, as well as the patient’s ability to manage in their home environment, to determine the need for assistive devices, level of caregiver support, future level of autonomy, and strategies for falls prevention.^3-7

There is a view among service providers that rehabilitation decisions can be based on a singular measure of function known as the Functional Independence Measure (FIM). This is an understandable position because not only is the FIM an internationally recognized, valid, and reliable tool, but, as a singular measure, it also means measurement consistency across rehabilitation sites is more likely. However, rehabilitation is complex, and it is risky to base decisions on a single measure, which might not capture the results of rehabilitation treatment ingredients on individual patient targets.^8,9

The patient’s progress is objectively assessed using functional outcome measures such as the FIM. Other measures used typically in our service include the de Morton’s Mobility Index (DEMMI), Timed Up and Go (TUG), and the Ten Meter Walk Test (10MWT), which measure patient mobility, balance during directional changes, and walking ability, respectively. Additional measures include patient progression to a less supportive level of assistance (ie, number of persons required to assist or level of supervision) or the selection of a walking aid (eg, forearm support frame, crutches). This progression—or lack thereof—assists in decision-making regarding the individual’s future once they are discharged from rehabilitation. Such considerations would include the need to modify the home environment, selection of assistive devices, community access (walking indoors, outdoors, and shopping), personal care needs, and age-appropriate care facility recommendations (ie, level of care). The use of outcome measures also indicates the need for further referrals to other care providers upon discharge from the rehabilitation facility.

There is widespread support in the literature for the use of the FIM, DEMMI, TUG, and 10MWT in rehabilitation population groups. For example, DEMMI has been validated in hip fracture patients during rehabilitation,¹⁰ as well as among older people hospitalized for medical illness.^11-13 It has also been shown to be a predictor of discharge destination for patients living with frailty in geriatric rehabilitation settings,¹⁴ and to have moderate predictive validity for functional independence after 4 weeks of rehabilitation.¹⁵ Similarly, TUG has been validated for use among hospitalized and community-dwelling individuals,^16-18 and for patients after joint arthroplasty^19,20 or hip fracture.²¹ It has also been shown to be an indicator of fall risk,^22-24 as well as a predictor of fracture incidence.²⁵ Furthermore, TUG has been identified as an indicator of a patient’s ability to walk in the community without the need for a walking device.²⁶ It has also been shown to be an early identifier of patients in need of rehabilitation.²⁷ Normative values for TUG have been reported, and the association with gait time established.²⁸

Gait speed has been shown to predict adverse outcomes in community-dwelling older people.²⁹ In fact, the 10MWT has been established as a powerful tool to benchmark rehabilitation recovery after a medical event.³⁰ Results of the test relate to overall quality of walking, health status, morbidity, and the rate of mortality.^31-33 Meaningful improvement, minimum detectable change (0.19-0.34 m/s), and responsiveness in common physical performance in older adults has been reported.^26,34,36

Structural and functional impairment has been used to define rehabilitation classes by the Australasian Rehabilitation Outcome Centre (AROC) in the Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP) Version 4.^37-43 Variables used for grouping are age, care type, function, and impairment for rehabilitation. FIM was developed in order to assess patients’ outcomes after inpatient multidisciplinary care, and is an internationally accepted measure of functioning.⁴⁴ It is a holistic outcome measure, which can be used to determine the patient’s level of disability and burden of care, and is widely used in both public and private inpatient rehabilitation settings. Each patient classification is reported separately within the case mix structure.⁴⁵ Inpatient rehabilitation centers are evaluated and compared by the AROC,⁴⁶ with an emphasis on length of stay and the FIM change. The most successful centers demonstrate shorter length of stay and greater FIM improvement. Although the FIM is a valuable measure, it does not provide a complete picture of the individual patient’s rehabilitation gain: ie, the specific attributes of patients’ mobility, walking ability, or balance during directional changes.

A large-scale analysis of the association between the holistic disability measure of the FIM and the more mobility- and ambulation-focused physiotherapy outcomes has not been documented.

The well-documented DEMMI accumulates points for the patient’s mobility in a similar fashion to the FIM, but with more mobility detail. These 2 outcome measures allow for the full range of patients, from the very dependent up to and including the independently ambulant patients. The DEMMI may show a positive relationship to the FIM, yet the association is unknown. The association of the TUG to the 10MWT has been established²⁸; however, their relationship to the FIM is unknown.

Current practice in the participating public health inpatient rehabilitation wards is to use the DEMMI, TUG, 10MWT, and FIM to ensure physiotherapy and allow the wider multidisciplinary team to more effectively evaluate patient mobility outcomes. The 3 most frequent patient groups identified within the current patient population are expected to present clinical differences and will be analyzed for comparison. If an association is found between the outcome measures in question, clinical efficiency could be improved.

The aim of the current study is to assess the association between change scores in the FIM with evaluative measures of outcomes typically used in physiotherapy to objectively show that use of the FIM in isolation is limited in our population of patients.

Methods

Study design and setting

This retrospective descriptive observational study complied with the STROBE-RECORD guidance and checklist (available at mdedge.com/jcomjournal) and analyzed the routinely collected data from rehabilitation patients who were admitted to 5 different rehabilitation wards in 4 different public hospitals from 1 regional local health district (20-24 beds per ward) from 2015 to 2019. As this study conducted secondary analyses using existing de-identified data from a public health facility and did not involve interaction with any human subjects, ethical approval was not required.⁴⁶ Approval to conduct this study was granted by the health district’s institutional review committee, as per the National Statement on Ethical Conduct in Human Research 2015.

Participants

Patient data over a 5-year time frame were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups were identified for inclusion in this study: reconditioning, orthopedic fracture, and orthopedic replacement. (See Table 1 for the specific AN-SNAP impairment groups used in this study.)

Patient data from the less-frequent impairment groups were excluded (n = 673, 28.19%), including stroke (n = 343), brain dysfunction (n = 45), amputation of limb (n = 45), spinal cord dysfunction (n  = 36), neurological dysfunction (n = 34), cardiac (n = 24), and others (n = 25) who may have benefitted from other outcome measures due to their medical condition. Ten patient data sets were excluded for missing discharge outcome measure data, from when the patient became ill and returned to acute services or was discharged at short notice. To be included in the study, both the admission and discharge scores from the FIM and the admission and discharge scores from at least 1 of the physiotherapy outcome measures were required for each patient (n = 1704, 71.39%): Reconditioning (n = 742), Orthopedic Fracture (n = 585), and Orthopedic Replacement (n = 377). Information regarding the type of walking aid and the amount of assistance required for safe ambulation was also recorded. These items were included in the study’s descriptive analysis. Only 1.7% of these descriptors were missing.

Outcome measures

DEMMI tasks of bed mobility, sitting balance, transfers, walking, and balance were scored with an assigned value according to the patient’s performance. This was then tallied and the results scaled, to provide an overall score out of 100 available points. The total score from admission and discharge was then compared. Improvement (change) was identified by the increase in scores.

The TUG assesses a patient’s dynamic balance performance.⁴⁷ The number of seconds it took the patient to complete the procedure was recorded at admission and discharge. Improvement (change) was identified by the reduction in time taken at discharge from the admission score.

The 10MWT measures the unidirectional walking speed of a person over 10 meters and is recorded in seconds and reported in meters per second. Improvement (change) was identified by the reduction in the time taken to increase walking speed.

Concurrent to the physiotherapy measures were the FIM scores, recorded by the accredited nursing staff from each rehabilitation ward. Improvement is demonstrated by the accumulation of points on the ordinal scale of the FIM Total, including mobility, dressing, bladder and bowel care, cognition, and social interaction, and is represented as a score between 18 and 126. The FIM Motor category is reported as a score between 13 and 91.

The 2 data sets were matched by unique identifier and admission dates, then de-identified for analysis.

Statistical analysis

Patient demographic information was analyzed using descriptive statistics (mean, SD, frequencies, percentages) for each impairment group (orthopedic fracture, orthopedic replacement, reconditioning). Differences in continuous demographic variables for each impairment group were assessed using Kruskal-Wallis tests and χ² tests for categorical variables. Functional outcome scores were compared at admission, discharge, and change between the impairment groups. Association of the functional outcome change scores was determined with the Pearson correlation coefficient (r) between the FIM and the DEMMI, TUG, and 10MWT. Graphs were plotted for each of these (Figure available online at mdedge.com/jcomjournal). A strong, moderate, and weak association was described as > 0.6, > 0.4, and > 0.2, respectively.⁴⁶ Statistical significance was set at P < .05. Analyses were conducted using Stata (StataCorp LLC, USA).

Results

The patient descriptive data (site from which data were collected, admission length of stay, age at admission, discharge destination, walk aid improvement, and walk assistance improvement) from the 3 impairment groups are reported in Table 2. The functional outcomes for DEMMI, TUG, 10MWT, FIM Motor, FIM Total at admission, discharge, and the change scores are presented in Table 3.

Orthopedic fracture patients had the greatest improvement in their functional outcomes, with a DEMMI improvement of 18 points, TUG score change of 23.49 seconds (s), 10MWT change of 0.30 meters/second (m/s), FIM Motor change of 20.62, and a FIM Total change of 21.9 points. The outcome measures exceeded the minimum detectable change as reported in the literature for DEMMI (8.8 points⁴⁸), TUG (2.08 s²⁶), walking speed 0.19 m/s²⁶, and FIM Motor (14.6 points⁴⁹).

Association of functional outcomes (change scores)

There was a significant weak positive correlation between DEMMI change score and both the FIM Motor (r = 0.396) and FIM Total change scores (r = 0.373). When viewing the specific items within the FIM Motor labelled FIM Walk change, FIM MobilityBedChair change, and FIM stairs change, r values were 0.100, 0.379, and 0.126, respectively. In addition, there was a weak negative correlation between TUG change scores and both FIM Motor (r = -0.217) and FIM Total change scores (r = -0.207). There was a very weak positive correlation between 10MWT (m/s) change scores and both FIM Motor (r = 0.194) and FIM Total change scores (r = 0.187) (Table 4, Figure). There was a moderate correlation between 10MWT change (s) and TUG change (s) (r = 0.72, P < .001).

Discussion

The purpose of this study was to ascertain the association between the DEMMI, TUG, 10MWT, and FIM measures using retrospective data collected from 5 public hospital inpatient rehabilitation wards. The results of this retrospective analysis demonstrate that a variety of objective outcome measures are required for the multidisciplinary team to accurately measure a patient’s functional improvement during their inpatient rehabilitation stay. No single outcome measure in this study fully reported all mobility attributes, and we note the risk of basing decisions on a single measure evaluating rehabilitation outcomes. Although the internationally used FIM has a strong place in rehabilitation reporting and benchmarking, it does not predict change nor provide a proxy for the patient’s whole-body motor control as they extend their mobility, dynamic balance, and ambulatory ability. Multiple objective outcome measures should therefore be required to evaluate the patient’s progress and functional performance toward discharge planning.

The FIM is a measure of disability or care needs, incorporating cognitive, social, and physical components of disability. It is a valid, holistic measure of an individual’s functional ability at a given time. Rehabilitation sites internationally utilize this assessment tool to evaluate a patient’s progress and the efficacy of intervention. The strength of this measure is its widespread use and the inclusion of the personal activities of daily living to provide an overall evaluation encompassing all aspects of a person’s ability to function independently. However, as our study results suggest, patient improvement measured by the FIM Motor components were not correlated to other widely used physiotherapy measures of ambulation and balance, such as the 10MWT or TUG. This is perhaps largely because the FIM Motor components only consider the level of assistance (eg, physical assistance, assistive device, independence) and do not consider assessment of balance and gait ability as assessed in the 10MWT and TUG. The 10MWT and TUG provide assessment of velocity and dynamic balance during walking, which have been shown to predict an individual’s risk of falling.^22,23 This is a pertinent issue in the rehabilitation and geriatric population.²⁹ Furthermore, the use of the FIM as a benchmarking tool to compare facility efficiency may not provide a complete assessment of all outcomes achieved on the inpatient rehabilitation ward, such as reduced falls risk or improved ambulatory ability and balance.

It has previously been confirmed that there is a significant positive correlation between the 10MWT and the TUG.²⁷ This retrospective analysis has also supported these findings. This is possibly due to the similarity in the assessments, as they both incorporate ambulation ability and dynamic movement.

Each of the 4 outcome measures assess different yet vital aspects of an individual’s functional mobility and ambulation ability during their subacute rehabilitation journey. The diversity of patient age, functional impairment, and mobility level needs a range of outcomes to provide baselines, targets, and goal attainment for discharge home.

Consistent with the AROC AN-SNAP reporting of Length of Stay and FIM change separated into the weighted impairment groups, the data analysis of this study demonstrated significant differences between the Reconditioning, Orthopedic Fracture, and Orthopedic Replacement patient data. Tables 2 and 3 describe the differences between the groups. The fracture population in this study improved the most across each outcome measure. In contrast, the reconditioning population showed the least improvement. This may be expected due to the pathophysiological differences between the groups. Furthermore, for the elderly who sustain fractures because of a fall, rehabilitation will be required to address not only the presenting injury but also the premorbid falls risk factors which may include polypharmacy or impaired balance.

Any conclusions drawn from the findings of this study need to take into consideration that it has focused on patients from 1 local health district and therefore may not be generalizable to a wider national or international context. As this study was a retrospective study, controlling for data collection quality, measurement bias due to nonblinding and missing data is a limitation. However, clinicians regularly completed these outcome assessments and recorded this information as part of their standard care practices within this health district. There may have been slight differences in definitions of practice between the 5 rehabilitation sites. To ensure reliability, each individual site’s protocols for the FIM, DEMMI, TUG, and 10MWT were reviewed and confirmed to be consistent.

Conclusion

The FIM Motor change scores showed a weak positive association to the DEMMI change, and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. Thorough reporting of clinical outcomes is much more meaningful to assess and guide the physiotherapy component of rehabilitation. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, mobility, functional capacity, and dynamic balance.

Acknowledgements: The authors thank Anne Smith, MSHLM, BAppSc, Head of the Physiotherapy Department, and the physiotherapists and allied health assistants who have contributed to the collection of this valuable data over several years. They also thank Lina Baytieh, MS, BS, from Research Central, Illawarra Shoalhaven Local Health District, for her assistance with the analysis.

Corresponding author: Maren Jones, MPH, BS, Physiotherapy Department, Port Kembla Hospital, Illawarra Shoalhaven Local Health District, Warrawong, New South Wales, 2505 Australia; [email protected].

Financial disclosures: None.

From Illawarra Shoalhaven Local Health District, New South Wales, Australia (Maren Jones, Dr. Hewitt, Philippa King, Rhiannon Thorn, Edward Davidson, and Tiana-Lee Elphick), and Illawarra Health and Medical Research Institute, University of Wollongong, Wollongong, New South Wales, Australia (Dr. Hewitt)

Objective: To assess the association between change scores in the Functional Independence Measure (FIM) with evaluative measures used in physiotherapy to objectively show that use of the FIM in isolation is limited.

Design: Retrospective observational study.

Setting: Five rehabilitation inpatient wards from 1 public local health district in NSW Australia.

Participants: Patient data over a 5-year time frame (2015 to 2019) were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups (Australasian Rehabilitation Outcome Centre classification) were identified for inclusion in this study: Reconditioning (n = 742, mean age 76.88 years); Orthopedic Fracture (n = 585, mean age 77.46 years); and Orthopedic Replacement (n = 377, mean age 73.84 years).

Measurements: The difference between the admission and discharge scores were calculated for each measure. Kruskal-Wallis and χ² tests were used to analyze the data.

Results: Pearson correlation (r) coefficients between FIM Motor change to the de Morton’s Mobility Index (DEMMI) change was r = 0.396, FIM Motor change to the Timed Up and Go (TUG) change was r = -0.217, and the FIM Motor change to the Ten Meter Walk Test (10MWT) change was .194.

Conclusion: The FIM Motor change scores showed a weak positive association to the DEMMI change and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, functional mobility, and dynamic balance.

Keywords: physiotherapy; rehabilitation; clinical outcome measures.

Patients receive interdisciplinary inpatient rehabilitation treatment after they have sustained a lower limb fracture, a lower limb joint replacement, or have generalized deconditioning (muscle wasting and disuse atrophy) following hospitalization for surgery or illness. The degree of a patient’s impairment or loss of functional capacity, as well as their ability to manage at home safely, is assessed using standardized outcome measures during their recovery and rehabilitation.^1,2

Physiotherapists routinely use validated outcome measures to assess patient progress and to measure goal attainment through assessment of functional independence, dynamic balance performance, and ambulatory ability. These objective assessments provide clinicians with information about the effectiveness of the rehabilitation program, as well as the patient’s ability to manage in their home environment, to determine the need for assistive devices, level of caregiver support, future level of autonomy, and strategies for falls prevention.^3-7

There is a view among service providers that rehabilitation decisions can be based on a singular measure of function known as the Functional Independence Measure (FIM). This is an understandable position because not only is the FIM an internationally recognized, valid, and reliable tool, but, as a singular measure, it also means measurement consistency across rehabilitation sites is more likely. However, rehabilitation is complex, and it is risky to base decisions on a single measure, which might not capture the results of rehabilitation treatment ingredients on individual patient targets.^8,9

The patient’s progress is objectively assessed using functional outcome measures such as the FIM. Other measures used typically in our service include the de Morton’s Mobility Index (DEMMI), Timed Up and Go (TUG), and the Ten Meter Walk Test (10MWT), which measure patient mobility, balance during directional changes, and walking ability, respectively. Additional measures include patient progression to a less supportive level of assistance (ie, number of persons required to assist or level of supervision) or the selection of a walking aid (eg, forearm support frame, crutches). This progression—or lack thereof—assists in decision-making regarding the individual’s future once they are discharged from rehabilitation. Such considerations would include the need to modify the home environment, selection of assistive devices, community access (walking indoors, outdoors, and shopping), personal care needs, and age-appropriate care facility recommendations (ie, level of care). The use of outcome measures also indicates the need for further referrals to other care providers upon discharge from the rehabilitation facility.

There is widespread support in the literature for the use of the FIM, DEMMI, TUG, and 10MWT in rehabilitation population groups. For example, DEMMI has been validated in hip fracture patients during rehabilitation,¹⁰ as well as among older people hospitalized for medical illness.^11-13 It has also been shown to be a predictor of discharge destination for patients living with frailty in geriatric rehabilitation settings,¹⁴ and to have moderate predictive validity for functional independence after 4 weeks of rehabilitation.¹⁵ Similarly, TUG has been validated for use among hospitalized and community-dwelling individuals,^16-18 and for patients after joint arthroplasty^19,20 or hip fracture.²¹ It has also been shown to be an indicator of fall risk,^22-24 as well as a predictor of fracture incidence.²⁵ Furthermore, TUG has been identified as an indicator of a patient’s ability to walk in the community without the need for a walking device.²⁶ It has also been shown to be an early identifier of patients in need of rehabilitation.²⁷ Normative values for TUG have been reported, and the association with gait time established.²⁸

Gait speed has been shown to predict adverse outcomes in community-dwelling older people.²⁹ In fact, the 10MWT has been established as a powerful tool to benchmark rehabilitation recovery after a medical event.³⁰ Results of the test relate to overall quality of walking, health status, morbidity, and the rate of mortality.^31-33 Meaningful improvement, minimum detectable change (0.19-0.34 m/s), and responsiveness in common physical performance in older adults has been reported.^26,34,36

Structural and functional impairment has been used to define rehabilitation classes by the Australasian Rehabilitation Outcome Centre (AROC) in the Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP) Version 4.^37-43 Variables used for grouping are age, care type, function, and impairment for rehabilitation. FIM was developed in order to assess patients’ outcomes after inpatient multidisciplinary care, and is an internationally accepted measure of functioning.⁴⁴ It is a holistic outcome measure, which can be used to determine the patient’s level of disability and burden of care, and is widely used in both public and private inpatient rehabilitation settings. Each patient classification is reported separately within the case mix structure.⁴⁵ Inpatient rehabilitation centers are evaluated and compared by the AROC,⁴⁶ with an emphasis on length of stay and the FIM change. The most successful centers demonstrate shorter length of stay and greater FIM improvement. Although the FIM is a valuable measure, it does not provide a complete picture of the individual patient’s rehabilitation gain: ie, the specific attributes of patients’ mobility, walking ability, or balance during directional changes.

A large-scale analysis of the association between the holistic disability measure of the FIM and the more mobility- and ambulation-focused physiotherapy outcomes has not been documented.

The well-documented DEMMI accumulates points for the patient’s mobility in a similar fashion to the FIM, but with more mobility detail. These 2 outcome measures allow for the full range of patients, from the very dependent up to and including the independently ambulant patients. The DEMMI may show a positive relationship to the FIM, yet the association is unknown. The association of the TUG to the 10MWT has been established²⁸; however, their relationship to the FIM is unknown.

Current practice in the participating public health inpatient rehabilitation wards is to use the DEMMI, TUG, 10MWT, and FIM to ensure physiotherapy and allow the wider multidisciplinary team to more effectively evaluate patient mobility outcomes. The 3 most frequent patient groups identified within the current patient population are expected to present clinical differences and will be analyzed for comparison. If an association is found between the outcome measures in question, clinical efficiency could be improved.

The aim of the current study is to assess the association between change scores in the FIM with evaluative measures of outcomes typically used in physiotherapy to objectively show that use of the FIM in isolation is limited in our population of patients.

Methods

Study design and setting

This retrospective descriptive observational study complied with the STROBE-RECORD guidance and checklist (available at mdedge.com/jcomjournal) and analyzed the routinely collected data from rehabilitation patients who were admitted to 5 different rehabilitation wards in 4 different public hospitals from 1 regional local health district (20-24 beds per ward) from 2015 to 2019. As this study conducted secondary analyses using existing de-identified data from a public health facility and did not involve interaction with any human subjects, ethical approval was not required.⁴⁶ Approval to conduct this study was granted by the health district’s institutional review committee, as per the National Statement on Ethical Conduct in Human Research 2015.

Participants

Patient data over a 5-year time frame were reviewed (N = 2378). The patient data from the 3 most prevalent impairment groups were identified for inclusion in this study: reconditioning, orthopedic fracture, and orthopedic replacement. (See Table 1 for the specific AN-SNAP impairment groups used in this study.)

Patient data from the less-frequent impairment groups were excluded (n = 673, 28.19%), including stroke (n = 343), brain dysfunction (n = 45), amputation of limb (n = 45), spinal cord dysfunction (n  = 36), neurological dysfunction (n = 34), cardiac (n = 24), and others (n = 25) who may have benefitted from other outcome measures due to their medical condition. Ten patient data sets were excluded for missing discharge outcome measure data, from when the patient became ill and returned to acute services or was discharged at short notice. To be included in the study, both the admission and discharge scores from the FIM and the admission and discharge scores from at least 1 of the physiotherapy outcome measures were required for each patient (n = 1704, 71.39%): Reconditioning (n = 742), Orthopedic Fracture (n = 585), and Orthopedic Replacement (n = 377). Information regarding the type of walking aid and the amount of assistance required for safe ambulation was also recorded. These items were included in the study’s descriptive analysis. Only 1.7% of these descriptors were missing.

Outcome measures

DEMMI tasks of bed mobility, sitting balance, transfers, walking, and balance were scored with an assigned value according to the patient’s performance. This was then tallied and the results scaled, to provide an overall score out of 100 available points. The total score from admission and discharge was then compared. Improvement (change) was identified by the increase in scores.

The TUG assesses a patient’s dynamic balance performance.⁴⁷ The number of seconds it took the patient to complete the procedure was recorded at admission and discharge. Improvement (change) was identified by the reduction in time taken at discharge from the admission score.

The 10MWT measures the unidirectional walking speed of a person over 10 meters and is recorded in seconds and reported in meters per second. Improvement (change) was identified by the reduction in the time taken to increase walking speed.

Concurrent to the physiotherapy measures were the FIM scores, recorded by the accredited nursing staff from each rehabilitation ward. Improvement is demonstrated by the accumulation of points on the ordinal scale of the FIM Total, including mobility, dressing, bladder and bowel care, cognition, and social interaction, and is represented as a score between 18 and 126. The FIM Motor category is reported as a score between 13 and 91.

The 2 data sets were matched by unique identifier and admission dates, then de-identified for analysis.

Statistical analysis

Patient demographic information was analyzed using descriptive statistics (mean, SD, frequencies, percentages) for each impairment group (orthopedic fracture, orthopedic replacement, reconditioning). Differences in continuous demographic variables for each impairment group were assessed using Kruskal-Wallis tests and χ² tests for categorical variables. Functional outcome scores were compared at admission, discharge, and change between the impairment groups. Association of the functional outcome change scores was determined with the Pearson correlation coefficient (r) between the FIM and the DEMMI, TUG, and 10MWT. Graphs were plotted for each of these (Figure available online at mdedge.com/jcomjournal). A strong, moderate, and weak association was described as > 0.6, > 0.4, and > 0.2, respectively.⁴⁶ Statistical significance was set at P < .05. Analyses were conducted using Stata (StataCorp LLC, USA).

Results

The patient descriptive data (site from which data were collected, admission length of stay, age at admission, discharge destination, walk aid improvement, and walk assistance improvement) from the 3 impairment groups are reported in Table 2. The functional outcomes for DEMMI, TUG, 10MWT, FIM Motor, FIM Total at admission, discharge, and the change scores are presented in Table 3.

Orthopedic fracture patients had the greatest improvement in their functional outcomes, with a DEMMI improvement of 18 points, TUG score change of 23.49 seconds (s), 10MWT change of 0.30 meters/second (m/s), FIM Motor change of 20.62, and a FIM Total change of 21.9 points. The outcome measures exceeded the minimum detectable change as reported in the literature for DEMMI (8.8 points⁴⁸), TUG (2.08 s²⁶), walking speed 0.19 m/s²⁶, and FIM Motor (14.6 points⁴⁹).

Association of functional outcomes (change scores)

There was a significant weak positive correlation between DEMMI change score and both the FIM Motor (r = 0.396) and FIM Total change scores (r = 0.373). When viewing the specific items within the FIM Motor labelled FIM Walk change, FIM MobilityBedChair change, and FIM stairs change, r values were 0.100, 0.379, and 0.126, respectively. In addition, there was a weak negative correlation between TUG change scores and both FIM Motor (r = -0.217) and FIM Total change scores (r = -0.207). There was a very weak positive correlation between 10MWT (m/s) change scores and both FIM Motor (r = 0.194) and FIM Total change scores (r = 0.187) (Table 4, Figure). There was a moderate correlation between 10MWT change (s) and TUG change (s) (r = 0.72, P < .001).

Discussion

The purpose of this study was to ascertain the association between the DEMMI, TUG, 10MWT, and FIM measures using retrospective data collected from 5 public hospital inpatient rehabilitation wards. The results of this retrospective analysis demonstrate that a variety of objective outcome measures are required for the multidisciplinary team to accurately measure a patient’s functional improvement during their inpatient rehabilitation stay. No single outcome measure in this study fully reported all mobility attributes, and we note the risk of basing decisions on a single measure evaluating rehabilitation outcomes. Although the internationally used FIM has a strong place in rehabilitation reporting and benchmarking, it does not predict change nor provide a proxy for the patient’s whole-body motor control as they extend their mobility, dynamic balance, and ambulatory ability. Multiple objective outcome measures should therefore be required to evaluate the patient’s progress and functional performance toward discharge planning.

The FIM is a measure of disability or care needs, incorporating cognitive, social, and physical components of disability. It is a valid, holistic measure of an individual’s functional ability at a given time. Rehabilitation sites internationally utilize this assessment tool to evaluate a patient’s progress and the efficacy of intervention. The strength of this measure is its widespread use and the inclusion of the personal activities of daily living to provide an overall evaluation encompassing all aspects of a person’s ability to function independently. However, as our study results suggest, patient improvement measured by the FIM Motor components were not correlated to other widely used physiotherapy measures of ambulation and balance, such as the 10MWT or TUG. This is perhaps largely because the FIM Motor components only consider the level of assistance (eg, physical assistance, assistive device, independence) and do not consider assessment of balance and gait ability as assessed in the 10MWT and TUG. The 10MWT and TUG provide assessment of velocity and dynamic balance during walking, which have been shown to predict an individual’s risk of falling.^22,23 This is a pertinent issue in the rehabilitation and geriatric population.²⁹ Furthermore, the use of the FIM as a benchmarking tool to compare facility efficiency may not provide a complete assessment of all outcomes achieved on the inpatient rehabilitation ward, such as reduced falls risk or improved ambulatory ability and balance.

It has previously been confirmed that there is a significant positive correlation between the 10MWT and the TUG.²⁷ This retrospective analysis has also supported these findings. This is possibly due to the similarity in the assessments, as they both incorporate ambulation ability and dynamic movement.

Each of the 4 outcome measures assess different yet vital aspects of an individual’s functional mobility and ambulation ability during their subacute rehabilitation journey. The diversity of patient age, functional impairment, and mobility level needs a range of outcomes to provide baselines, targets, and goal attainment for discharge home.

Consistent with the AROC AN-SNAP reporting of Length of Stay and FIM change separated into the weighted impairment groups, the data analysis of this study demonstrated significant differences between the Reconditioning, Orthopedic Fracture, and Orthopedic Replacement patient data. Tables 2 and 3 describe the differences between the groups. The fracture population in this study improved the most across each outcome measure. In contrast, the reconditioning population showed the least improvement. This may be expected due to the pathophysiological differences between the groups. Furthermore, for the elderly who sustain fractures because of a fall, rehabilitation will be required to address not only the presenting injury but also the premorbid falls risk factors which may include polypharmacy or impaired balance.

Any conclusions drawn from the findings of this study need to take into consideration that it has focused on patients from 1 local health district and therefore may not be generalizable to a wider national or international context. As this study was a retrospective study, controlling for data collection quality, measurement bias due to nonblinding and missing data is a limitation. However, clinicians regularly completed these outcome assessments and recorded this information as part of their standard care practices within this health district. There may have been slight differences in definitions of practice between the 5 rehabilitation sites. To ensure reliability, each individual site’s protocols for the FIM, DEMMI, TUG, and 10MWT were reviewed and confirmed to be consistent.

Conclusion

The FIM Motor change scores showed a weak positive association to the DEMMI change, and no association to the TUG and 10MWT change, demonstrating that the outcome measures do not measure the same attributes. Thorough reporting of clinical outcomes is much more meaningful to assess and guide the physiotherapy component of rehabilitation. To review rehabilitation effectiveness from a management perspective, it is recommended that all measures are reviewed to assess the burden of care, mobility, functional capacity, and dynamic balance.

Acknowledgements: The authors thank Anne Smith, MSHLM, BAppSc, Head of the Physiotherapy Department, and the physiotherapists and allied health assistants who have contributed to the collection of this valuable data over several years. They also thank Lina Baytieh, MS, BS, from Research Central, Illawarra Shoalhaven Local Health District, for her assistance with the analysis.

Corresponding author: Maren Jones, MPH, BS, Physiotherapy Department, Port Kembla Hospital, Illawarra Shoalhaven Local Health District, Warrawong, New South Wales, 2505 Australia; [email protected].

Financial disclosures: None.

From the Department of Neurology, Cedars-Sinai Medical Center, Los Angeles, CA (Drs. Ching, Darwish, Li, Wong, Simpson, and Funk), the Department of Anesthesia, Cedars-Sinai Medical Center, Los Angeles, CA (Keith Siegel), and the Department of Psychiatry, Cedars-Sinai Medical Center, Los Angeles, CA (Dr. Bamgbose).

Objectives: To reduce the incidence and duration of delirium among patients in a hospital ward through standardized delirium screening tools and nonpharmacologic interventions. To advance nursing-focused education on delirium-prevention strategies. To measure the efficacy of the interventions with the aim of reproducing best practices.

Background: Delirium is associated with poor patient outcomes but may be preventable in a significant percentage of hospitalized patients.

Methods: Following nursing-focused education to prevent delirium, we prospectively evaluated patient care outcomes in a consecutive series of patients who were admitted to a hospital medical-surgical ward within a 25-week period. All patients who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria (N = 353). Standards for Quality Improvement Reporting Excellence guidelines were adhered to.

Results: There were 187 patients in the control group, and 166 in the postintervention group. Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days), mean length of stay (8.5 days vs 5.9 days), and use of an indwelling urinary catheter (9.1% vs 2.4%).

Conclusion: A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs.

Delirium is a disorder characterized by inattention and acute changes in cognition. It is defined by the American Psychiatric Association’s fifth edition of the Diagnostic and Statistical Manual of Mental Disorders as a disturbance in attention, awareness, and cognition over hours to a few days that is not better explained by a preexisting, established, or other evolving neurocognitive disorder.¹ Delirium is common yet often under-recognized among hospitalized patients, particularly in the elderly. The incidence of delirium in elderly patients on admission is estimated to be 11% to 25%, and an additional 29% to 31% of elderly patients will develop delirium during the hospitalization.² Delirium costs the health care system an estimated $38 billion to $152 billion per year.³ It is associated with negative outcomes, such as increased new placements to nursing homes, increased mortality, increased risk of dementia, and further cognitive deterioration among patients with dementia.^4-6

Despite its prevalence, delirium may be preventable in a significant percentage of hospitalized patients. Targeted intervention strategies, such as frequent reorientation, maximizing sleep, early mobilization, restricting use of psychoactive medications, and addressing hearing or vision impairment, have been demonstrated to significantly reduce the incidence of hospital delirium.^7,8 To achieve these goals, we explored the use of a multimodal strategy centered on nursing education. We integrated consistent, standardized delirium screening and nonpharmacologic interventions as part of a preventative protocol to reduce the incidence of delirium in the hospital ward.

Methods

We evaluated a consecutive series of patients who were admitted to a designated hospital medical-surgical ward within a 25-week period between October 2019 and April 2020. All patients during this period who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria. Patients who did not have a CAM documented were excluded from the analysis. Delirium was defined according to the CAM diagnostic algorithm.⁹

Core nursing staff regularly assigned to the ward completed a multimodal training program designed to improve recognition, documentation, and prevention of hospital delirium. Prior to the training, the nurses completed a 5-point Likert scale survey assessing their level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium. Nurses completed the same survey after the study period ended.

The training curriculum for nurses began with an online module reviewing the epidemiology and risk factors for delirium. Nurses then participated in a series of in-service training sessions led by a team of physicians, during which the CAM and nonpharmacologic delirium prevention measures were reviewed then practiced first-hand. Nursing staff attended an in-person lecture reviewing the current body of literature on delirium risk factors and effective nursing interventions. After formal training was completed, nurses were instructed to document CAM screens for each patient under their care at least once every 12-hour shift for the remainder of the study. An order set, reflected in Table 1, was made available to physicians and floor nurses to assist with implementing the educational components.

Patients admitted to the hospital unit from the start of the training program (week 1) until the order set was made available (week 15) constituted our control group. The postintervention study group consisted of patients admitted for 10 weeks after the completion of the interventions (weeks 16-25). A timeline of the study events is shown in Figure 1.

Patient demographics and hospital-stay metrics determined a priori were attained via the Cedars-Sinai Enterprise Information Services core. Age, sex, medical history, and incidence of surgery with anesthesia during hospitalization were recorded. The Charlson Comorbidity Index was calculated from patients’ listed diagnoses following discharge. Primary outcomes included incidence of patients with delirium during hospitalization, percentage of tested shifts with positive CAM screens, length of hospital stay, and survival. Secondary outcomes included measures associated with delirium, including the use of chemical restraints, physical restraints, sitters, indwelling urinary catheters, and new psychiatry and neurology consults. Chemical restraints were defined as administration of a new antipsychotic medication or benzodiazepine for the specific indication of hyperactive delirium or agitation.            

Statistical analysis was conducted by a statistician, using R version 3.6.3.10P values of < .05 were considered significant. Categorical variables were analyzed using Fisher’s exact test. Continuous variables were analyzed with Welch’s t-test or, for highly skewed continuous variables, with Wilcoxon rank-sum test or Mood’s median test. All patient data were anonymized and stored securely in accordance with institutional guidelines.

Our project was deemed to represent nonhuman subject research and therefore did not require Institutional Review Board (IRB) approval upon review by our institution’s IRB committee and Office of Research Compliance and Quality Improvement. Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) guidelines were adhered to (Supplementary File can be found at mdedge.com/jcomjournal).

Results

We evaluated 353 patients who met our inclusion criteria: 187 in the control group, and 166 in the postintervention group. Ten patients were readmitted to the ward after their initial discharge; only the initial admission encounters were included in our analysis. Median age, sex, median Charlson Comorbidity Index, and incidence of surgery with anesthesia during hospitalization were comparable between the control and postintervention groups and are summarized in Table 2.

In the control group, 1572 CAMs were performed, with 74 positive CAMs recorded among 27 patients with delirium. In the postintervention group, 1298 CAMs were performed, with 12 positive CAMs recorded among 7 patients with delirium (Figure 2). Primary and secondary outcomes, as well as CAM compliance measures, are summarized in Table 3.

Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%, P = .002) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%, P = .002). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days, P = .004), mean length of stay (8.5 days vs 5.9 days, P = .003), and use of an indwelling urinary catheter (9.1% vs 2.4%, P = .012). There was a trend towards decreased incidence of chemical restraints and psychiatry consults, which did not reach statistical significance. Differences in mortality during hospitalization, physical restraint use, and sitter use could not be assessed due to low incidence.

Results of nursing surveys completed before and after the training program are listed in Table 4. After training, nurses had a greater level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium.

Discussion

Our study utilized a standardized delirium assessment tool to compare patient cohorts before and after nurse-targeted training interventions on delirium recognition and prevention. Our interventions emphasized nonpharmacologic intervention strategies, which are recommended as first-line in the management of patients with delirium.¹¹ Patients were not excluded from the analysis based on preexisting medical conditions or recent surgery with anesthesia, to allow for conditions that are representative of community hospitals. We also did not use an inclusion criterion based on age; however, the majority of our patients were greater than 70 years old, representing those at highest risk for delirium.² Significant outcomes among patients in the postintervention group include decreased incidence of delirium, lower average length of stay, decreased indwelling urinary catheter use, and increased compliance with delirium screening by nursing staff.

While the study’s focus was primarily on delirium prevention rather than treatment, these strategies may also have conferred the benefit of reversing delirium symptoms. In addition to measuring incidence of delirium, our primary outcome of percentage of tested shifts with 1 or more positive CAM was intended to assess the overall duration in which patients had delirium during their hospitalization. The reduction in shifts with positive CAMs observed in the postintervention group is notable, given that a significant percentage of patients with hospital delirium have the potential for symptom reversibility.¹²

Multiple studies have shown that admitted patients who develop delirium experience prolonged hospital stays, often up to 5 to 10 days longer.^12-14 The decreased incidence and duration of delirium in our postintervention group is a reasonable explanation for the observed decrease in average length of stay. Our study is in line with previously documented initiatives that show that nonpharmacologic interventions can effectively address downstream health and fiscal sequelae of hospital delirium. For example, a volunteer-based initiative named the Hospital Elder Life Program, from which elements in our order set were modeled after, demonstrated significant reductions in delirium incidence, length of stay, and health care costs.^14-16 Other initiatives that focused on educational training for nurses to assess and prevent delirium have also demonstrated similar positive results.^17-19 Our study provides a model for effective nursing-focused education that can be reproduced in the hospital setting.

Unlike some other studies, which identified delirium based only on physician assessments, our initiative utilized the CAM performed by floor nurses to identify delirium. While this method may have affected the sensitivity and specificity of the CAMs, it also conferred the advantage of recognizing, documenting, and intervening on delirium in real time, given that bedside nurses are often the first to encounter delirium. Furthermore, nurses were instructed to notify a physician if a patient had a new positive CAM, as reflected in Table 1.

Our study demonstrated an increase in the overall compliance with the CAM screening during the postintervention period, which is significant given the under-recognition of delirium by health care professionals.²⁰ We attribute this increase to greater realized importance and a higher level of confidence from nursing staff in recognizing and addressing delirium, as supported by survey data. While the increased screening of patients should be considered a positive outcome, it also poses the possibility that the observed decrease in delirium incidence in the postintervention group was in fact due to more CAMs performed on patients without delirium. Likewise, nurses may have become more adept at recognizing true delirium, as opposed to delirium mimics, in the latter period of the study.

Perhaps the greatest limitation of our study is the variability in performing and recording CAMs, as some patients had multiple CAMs recorded while others did not have any CAMs recorded. This may have been affected in part by the increase in COVID-19 cases in our hospital towards the latter half of the study, which resulted in changes in nursing assignments as well as patient comorbidities in ways that cannot be easily quantified. Given the limited size of our patient cohorts, certain outcomes, such as the use of sitters, physical restraints, and in-hospital mortality, were unable to be assessed for changes statistically. Causative relationships between our interventions and associated outcome measures are necessarily limited in a binary comparison between control and postintervention groups.

Within these limitations, our study demonstrates promising results in core dimensions of patient care. We anticipate further quality improvement initiatives involving greater numbers of nursing staff and patients to better quantify the impact of nonpharmacologic nursing-centered interventions for preventing hospital delirium.

Conclusion

A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs. Education and tools to equip nurses to perform standardized delirium screening and interventions should be prioritized.

Acknowledgment: The authors thanks Olena Svetlov, NP, Oscar Abarca, Jose Chavez, and Jenita Gutierrez.

Corresponding author: Jason Ching, MD, Department of Neurology, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Los Angeles, CA 90048; [email protected].

Financial disclosures: None.

Funding: This research was supported by NIH National Center for Advancing Translational Science (NCATS) UCLA CTSI Grant Number UL1TR001881.

From the Department of Neurology, Cedars-Sinai Medical Center, Los Angeles, CA (Drs. Ching, Darwish, Li, Wong, Simpson, and Funk), the Department of Anesthesia, Cedars-Sinai Medical Center, Los Angeles, CA (Keith Siegel), and the Department of Psychiatry, Cedars-Sinai Medical Center, Los Angeles, CA (Dr. Bamgbose).

Objectives: To reduce the incidence and duration of delirium among patients in a hospital ward through standardized delirium screening tools and nonpharmacologic interventions. To advance nursing-focused education on delirium-prevention strategies. To measure the efficacy of the interventions with the aim of reproducing best practices.

Background: Delirium is associated with poor patient outcomes but may be preventable in a significant percentage of hospitalized patients.

Methods: Following nursing-focused education to prevent delirium, we prospectively evaluated patient care outcomes in a consecutive series of patients who were admitted to a hospital medical-surgical ward within a 25-week period. All patients who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria (N = 353). Standards for Quality Improvement Reporting Excellence guidelines were adhered to.

Results: There were 187 patients in the control group, and 166 in the postintervention group. Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days), mean length of stay (8.5 days vs 5.9 days), and use of an indwelling urinary catheter (9.1% vs 2.4%).

Conclusion: A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs.

Delirium is a disorder characterized by inattention and acute changes in cognition. It is defined by the American Psychiatric Association’s fifth edition of the Diagnostic and Statistical Manual of Mental Disorders as a disturbance in attention, awareness, and cognition over hours to a few days that is not better explained by a preexisting, established, or other evolving neurocognitive disorder.¹ Delirium is common yet often under-recognized among hospitalized patients, particularly in the elderly. The incidence of delirium in elderly patients on admission is estimated to be 11% to 25%, and an additional 29% to 31% of elderly patients will develop delirium during the hospitalization.² Delirium costs the health care system an estimated $38 billion to $152 billion per year.³ It is associated with negative outcomes, such as increased new placements to nursing homes, increased mortality, increased risk of dementia, and further cognitive deterioration among patients with dementia.^4-6

Despite its prevalence, delirium may be preventable in a significant percentage of hospitalized patients. Targeted intervention strategies, such as frequent reorientation, maximizing sleep, early mobilization, restricting use of psychoactive medications, and addressing hearing or vision impairment, have been demonstrated to significantly reduce the incidence of hospital delirium.^7,8 To achieve these goals, we explored the use of a multimodal strategy centered on nursing education. We integrated consistent, standardized delirium screening and nonpharmacologic interventions as part of a preventative protocol to reduce the incidence of delirium in the hospital ward.

Methods

We evaluated a consecutive series of patients who were admitted to a designated hospital medical-surgical ward within a 25-week period between October 2019 and April 2020. All patients during this period who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria. Patients who did not have a CAM documented were excluded from the analysis. Delirium was defined according to the CAM diagnostic algorithm.⁹

Core nursing staff regularly assigned to the ward completed a multimodal training program designed to improve recognition, documentation, and prevention of hospital delirium. Prior to the training, the nurses completed a 5-point Likert scale survey assessing their level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium. Nurses completed the same survey after the study period ended.

The training curriculum for nurses began with an online module reviewing the epidemiology and risk factors for delirium. Nurses then participated in a series of in-service training sessions led by a team of physicians, during which the CAM and nonpharmacologic delirium prevention measures were reviewed then practiced first-hand. Nursing staff attended an in-person lecture reviewing the current body of literature on delirium risk factors and effective nursing interventions. After formal training was completed, nurses were instructed to document CAM screens for each patient under their care at least once every 12-hour shift for the remainder of the study. An order set, reflected in Table 1, was made available to physicians and floor nurses to assist with implementing the educational components.

Patients admitted to the hospital unit from the start of the training program (week 1) until the order set was made available (week 15) constituted our control group. The postintervention study group consisted of patients admitted for 10 weeks after the completion of the interventions (weeks 16-25). A timeline of the study events is shown in Figure 1.

Patient demographics and hospital-stay metrics determined a priori were attained via the Cedars-Sinai Enterprise Information Services core. Age, sex, medical history, and incidence of surgery with anesthesia during hospitalization were recorded. The Charlson Comorbidity Index was calculated from patients’ listed diagnoses following discharge. Primary outcomes included incidence of patients with delirium during hospitalization, percentage of tested shifts with positive CAM screens, length of hospital stay, and survival. Secondary outcomes included measures associated with delirium, including the use of chemical restraints, physical restraints, sitters, indwelling urinary catheters, and new psychiatry and neurology consults. Chemical restraints were defined as administration of a new antipsychotic medication or benzodiazepine for the specific indication of hyperactive delirium or agitation.            

Statistical analysis was conducted by a statistician, using R version 3.6.3.10P values of < .05 were considered significant. Categorical variables were analyzed using Fisher’s exact test. Continuous variables were analyzed with Welch’s t-test or, for highly skewed continuous variables, with Wilcoxon rank-sum test or Mood’s median test. All patient data were anonymized and stored securely in accordance with institutional guidelines.

Our project was deemed to represent nonhuman subject research and therefore did not require Institutional Review Board (IRB) approval upon review by our institution’s IRB committee and Office of Research Compliance and Quality Improvement. Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) guidelines were adhered to (Supplementary File can be found at mdedge.com/jcomjournal).

Results

We evaluated 353 patients who met our inclusion criteria: 187 in the control group, and 166 in the postintervention group. Ten patients were readmitted to the ward after their initial discharge; only the initial admission encounters were included in our analysis. Median age, sex, median Charlson Comorbidity Index, and incidence of surgery with anesthesia during hospitalization were comparable between the control and postintervention groups and are summarized in Table 2.

In the control group, 1572 CAMs were performed, with 74 positive CAMs recorded among 27 patients with delirium. In the postintervention group, 1298 CAMs were performed, with 12 positive CAMs recorded among 7 patients with delirium (Figure 2). Primary and secondary outcomes, as well as CAM compliance measures, are summarized in Table 3.

Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%, P = .002) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%, P = .002). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days, P = .004), mean length of stay (8.5 days vs 5.9 days, P = .003), and use of an indwelling urinary catheter (9.1% vs 2.4%, P = .012). There was a trend towards decreased incidence of chemical restraints and psychiatry consults, which did not reach statistical significance. Differences in mortality during hospitalization, physical restraint use, and sitter use could not be assessed due to low incidence.

Results of nursing surveys completed before and after the training program are listed in Table 4. After training, nurses had a greater level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium.

Discussion

Our study utilized a standardized delirium assessment tool to compare patient cohorts before and after nurse-targeted training interventions on delirium recognition and prevention. Our interventions emphasized nonpharmacologic intervention strategies, which are recommended as first-line in the management of patients with delirium.¹¹ Patients were not excluded from the analysis based on preexisting medical conditions or recent surgery with anesthesia, to allow for conditions that are representative of community hospitals. We also did not use an inclusion criterion based on age; however, the majority of our patients were greater than 70 years old, representing those at highest risk for delirium.² Significant outcomes among patients in the postintervention group include decreased incidence of delirium, lower average length of stay, decreased indwelling urinary catheter use, and increased compliance with delirium screening by nursing staff.

While the study’s focus was primarily on delirium prevention rather than treatment, these strategies may also have conferred the benefit of reversing delirium symptoms. In addition to measuring incidence of delirium, our primary outcome of percentage of tested shifts with 1 or more positive CAM was intended to assess the overall duration in which patients had delirium during their hospitalization. The reduction in shifts with positive CAMs observed in the postintervention group is notable, given that a significant percentage of patients with hospital delirium have the potential for symptom reversibility.¹²

Multiple studies have shown that admitted patients who develop delirium experience prolonged hospital stays, often up to 5 to 10 days longer.^12-14 The decreased incidence and duration of delirium in our postintervention group is a reasonable explanation for the observed decrease in average length of stay. Our study is in line with previously documented initiatives that show that nonpharmacologic interventions can effectively address downstream health and fiscal sequelae of hospital delirium. For example, a volunteer-based initiative named the Hospital Elder Life Program, from which elements in our order set were modeled after, demonstrated significant reductions in delirium incidence, length of stay, and health care costs.^14-16 Other initiatives that focused on educational training for nurses to assess and prevent delirium have also demonstrated similar positive results.^17-19 Our study provides a model for effective nursing-focused education that can be reproduced in the hospital setting.

Unlike some other studies, which identified delirium based only on physician assessments, our initiative utilized the CAM performed by floor nurses to identify delirium. While this method may have affected the sensitivity and specificity of the CAMs, it also conferred the advantage of recognizing, documenting, and intervening on delirium in real time, given that bedside nurses are often the first to encounter delirium. Furthermore, nurses were instructed to notify a physician if a patient had a new positive CAM, as reflected in Table 1.

Our study demonstrated an increase in the overall compliance with the CAM screening during the postintervention period, which is significant given the under-recognition of delirium by health care professionals.²⁰ We attribute this increase to greater realized importance and a higher level of confidence from nursing staff in recognizing and addressing delirium, as supported by survey data. While the increased screening of patients should be considered a positive outcome, it also poses the possibility that the observed decrease in delirium incidence in the postintervention group was in fact due to more CAMs performed on patients without delirium. Likewise, nurses may have become more adept at recognizing true delirium, as opposed to delirium mimics, in the latter period of the study.

Perhaps the greatest limitation of our study is the variability in performing and recording CAMs, as some patients had multiple CAMs recorded while others did not have any CAMs recorded. This may have been affected in part by the increase in COVID-19 cases in our hospital towards the latter half of the study, which resulted in changes in nursing assignments as well as patient comorbidities in ways that cannot be easily quantified. Given the limited size of our patient cohorts, certain outcomes, such as the use of sitters, physical restraints, and in-hospital mortality, were unable to be assessed for changes statistically. Causative relationships between our interventions and associated outcome measures are necessarily limited in a binary comparison between control and postintervention groups.

Within these limitations, our study demonstrates promising results in core dimensions of patient care. We anticipate further quality improvement initiatives involving greater numbers of nursing staff and patients to better quantify the impact of nonpharmacologic nursing-centered interventions for preventing hospital delirium.

Conclusion

A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs. Education and tools to equip nurses to perform standardized delirium screening and interventions should be prioritized.

Acknowledgment: The authors thanks Olena Svetlov, NP, Oscar Abarca, Jose Chavez, and Jenita Gutierrez.

Corresponding author: Jason Ching, MD, Department of Neurology, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Los Angeles, CA 90048; [email protected].

Financial disclosures: None.

Funding: This research was supported by NIH National Center for Advancing Translational Science (NCATS) UCLA CTSI Grant Number UL1TR001881.

From the Department of Neurology, Cedars-Sinai Medical Center, Los Angeles, CA (Drs. Ching, Darwish, Li, Wong, Simpson, and Funk), the Department of Anesthesia, Cedars-Sinai Medical Center, Los Angeles, CA (Keith Siegel), and the Department of Psychiatry, Cedars-Sinai Medical Center, Los Angeles, CA (Dr. Bamgbose).

Objectives: To reduce the incidence and duration of delirium among patients in a hospital ward through standardized delirium screening tools and nonpharmacologic interventions. To advance nursing-focused education on delirium-prevention strategies. To measure the efficacy of the interventions with the aim of reproducing best practices.

Background: Delirium is associated with poor patient outcomes but may be preventable in a significant percentage of hospitalized patients.

Methods: Following nursing-focused education to prevent delirium, we prospectively evaluated patient care outcomes in a consecutive series of patients who were admitted to a hospital medical-surgical ward within a 25-week period. All patients who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria (N = 353). Standards for Quality Improvement Reporting Excellence guidelines were adhered to.

Results: There were 187 patients in the control group, and 166 in the postintervention group. Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days), mean length of stay (8.5 days vs 5.9 days), and use of an indwelling urinary catheter (9.1% vs 2.4%).

Conclusion: A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs.

Delirium is a disorder characterized by inattention and acute changes in cognition. It is defined by the American Psychiatric Association’s fifth edition of the Diagnostic and Statistical Manual of Mental Disorders as a disturbance in attention, awareness, and cognition over hours to a few days that is not better explained by a preexisting, established, or other evolving neurocognitive disorder.¹ Delirium is common yet often under-recognized among hospitalized patients, particularly in the elderly. The incidence of delirium in elderly patients on admission is estimated to be 11% to 25%, and an additional 29% to 31% of elderly patients will develop delirium during the hospitalization.² Delirium costs the health care system an estimated $38 billion to $152 billion per year.³ It is associated with negative outcomes, such as increased new placements to nursing homes, increased mortality, increased risk of dementia, and further cognitive deterioration among patients with dementia.^4-6

Despite its prevalence, delirium may be preventable in a significant percentage of hospitalized patients. Targeted intervention strategies, such as frequent reorientation, maximizing sleep, early mobilization, restricting use of psychoactive medications, and addressing hearing or vision impairment, have been demonstrated to significantly reduce the incidence of hospital delirium.^7,8 To achieve these goals, we explored the use of a multimodal strategy centered on nursing education. We integrated consistent, standardized delirium screening and nonpharmacologic interventions as part of a preventative protocol to reduce the incidence of delirium in the hospital ward.

Methods

We evaluated a consecutive series of patients who were admitted to a designated hospital medical-surgical ward within a 25-week period between October 2019 and April 2020. All patients during this period who had at least 1 Confusion Assessment Method (CAM) documented by a nurse during hospitalization met our inclusion criteria. Patients who did not have a CAM documented were excluded from the analysis. Delirium was defined according to the CAM diagnostic algorithm.⁹

Core nursing staff regularly assigned to the ward completed a multimodal training program designed to improve recognition, documentation, and prevention of hospital delirium. Prior to the training, the nurses completed a 5-point Likert scale survey assessing their level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium. Nurses completed the same survey after the study period ended.

The training curriculum for nurses began with an online module reviewing the epidemiology and risk factors for delirium. Nurses then participated in a series of in-service training sessions led by a team of physicians, during which the CAM and nonpharmacologic delirium prevention measures were reviewed then practiced first-hand. Nursing staff attended an in-person lecture reviewing the current body of literature on delirium risk factors and effective nursing interventions. After formal training was completed, nurses were instructed to document CAM screens for each patient under their care at least once every 12-hour shift for the remainder of the study. An order set, reflected in Table 1, was made available to physicians and floor nurses to assist with implementing the educational components.

Patients admitted to the hospital unit from the start of the training program (week 1) until the order set was made available (week 15) constituted our control group. The postintervention study group consisted of patients admitted for 10 weeks after the completion of the interventions (weeks 16-25). A timeline of the study events is shown in Figure 1.

Patient demographics and hospital-stay metrics determined a priori were attained via the Cedars-Sinai Enterprise Information Services core. Age, sex, medical history, and incidence of surgery with anesthesia during hospitalization were recorded. The Charlson Comorbidity Index was calculated from patients’ listed diagnoses following discharge. Primary outcomes included incidence of patients with delirium during hospitalization, percentage of tested shifts with positive CAM screens, length of hospital stay, and survival. Secondary outcomes included measures associated with delirium, including the use of chemical restraints, physical restraints, sitters, indwelling urinary catheters, and new psychiatry and neurology consults. Chemical restraints were defined as administration of a new antipsychotic medication or benzodiazepine for the specific indication of hyperactive delirium or agitation.            

Statistical analysis was conducted by a statistician, using R version 3.6.3.10P values of < .05 were considered significant. Categorical variables were analyzed using Fisher’s exact test. Continuous variables were analyzed with Welch’s t-test or, for highly skewed continuous variables, with Wilcoxon rank-sum test or Mood’s median test. All patient data were anonymized and stored securely in accordance with institutional guidelines.

Our project was deemed to represent nonhuman subject research and therefore did not require Institutional Review Board (IRB) approval upon review by our institution’s IRB committee and Office of Research Compliance and Quality Improvement. Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) guidelines were adhered to (Supplementary File can be found at mdedge.com/jcomjournal).

Results

We evaluated 353 patients who met our inclusion criteria: 187 in the control group, and 166 in the postintervention group. Ten patients were readmitted to the ward after their initial discharge; only the initial admission encounters were included in our analysis. Median age, sex, median Charlson Comorbidity Index, and incidence of surgery with anesthesia during hospitalization were comparable between the control and postintervention groups and are summarized in Table 2.

In the control group, 1572 CAMs were performed, with 74 positive CAMs recorded among 27 patients with delirium. In the postintervention group, 1298 CAMs were performed, with 12 positive CAMs recorded among 7 patients with delirium (Figure 2). Primary and secondary outcomes, as well as CAM compliance measures, are summarized in Table 3.

Compared to the control group, the postintervention group had a significant decrease in the incidence of delirium during hospitalization (14.4% vs 4.2%, P = .002) and a significant decrease in the mean percentage of tested nursing shifts with 1 or more positive CAM (4.9% vs 1.1%, P = .002). Significant differences in secondary outcomes between the control and postintervention groups included median length of stay (6 days vs 4 days, P = .004), mean length of stay (8.5 days vs 5.9 days, P = .003), and use of an indwelling urinary catheter (9.1% vs 2.4%, P = .012). There was a trend towards decreased incidence of chemical restraints and psychiatry consults, which did not reach statistical significance. Differences in mortality during hospitalization, physical restraint use, and sitter use could not be assessed due to low incidence.

Results of nursing surveys completed before and after the training program are listed in Table 4. After training, nurses had a greater level of confidence with recognizing delirium risk factors, preventing delirium, addressing delirium, utilizing the CAM tool, and educating others about delirium.

Discussion

Our study utilized a standardized delirium assessment tool to compare patient cohorts before and after nurse-targeted training interventions on delirium recognition and prevention. Our interventions emphasized nonpharmacologic intervention strategies, which are recommended as first-line in the management of patients with delirium.¹¹ Patients were not excluded from the analysis based on preexisting medical conditions or recent surgery with anesthesia, to allow for conditions that are representative of community hospitals. We also did not use an inclusion criterion based on age; however, the majority of our patients were greater than 70 years old, representing those at highest risk for delirium.² Significant outcomes among patients in the postintervention group include decreased incidence of delirium, lower average length of stay, decreased indwelling urinary catheter use, and increased compliance with delirium screening by nursing staff.

While the study’s focus was primarily on delirium prevention rather than treatment, these strategies may also have conferred the benefit of reversing delirium symptoms. In addition to measuring incidence of delirium, our primary outcome of percentage of tested shifts with 1 or more positive CAM was intended to assess the overall duration in which patients had delirium during their hospitalization. The reduction in shifts with positive CAMs observed in the postintervention group is notable, given that a significant percentage of patients with hospital delirium have the potential for symptom reversibility.¹²

Multiple studies have shown that admitted patients who develop delirium experience prolonged hospital stays, often up to 5 to 10 days longer.^12-14 The decreased incidence and duration of delirium in our postintervention group is a reasonable explanation for the observed decrease in average length of stay. Our study is in line with previously documented initiatives that show that nonpharmacologic interventions can effectively address downstream health and fiscal sequelae of hospital delirium. For example, a volunteer-based initiative named the Hospital Elder Life Program, from which elements in our order set were modeled after, demonstrated significant reductions in delirium incidence, length of stay, and health care costs.^14-16 Other initiatives that focused on educational training for nurses to assess and prevent delirium have also demonstrated similar positive results.^17-19 Our study provides a model for effective nursing-focused education that can be reproduced in the hospital setting.

Unlike some other studies, which identified delirium based only on physician assessments, our initiative utilized the CAM performed by floor nurses to identify delirium. While this method may have affected the sensitivity and specificity of the CAMs, it also conferred the advantage of recognizing, documenting, and intervening on delirium in real time, given that bedside nurses are often the first to encounter delirium. Furthermore, nurses were instructed to notify a physician if a patient had a new positive CAM, as reflected in Table 1.

Our study demonstrated an increase in the overall compliance with the CAM screening during the postintervention period, which is significant given the under-recognition of delirium by health care professionals.²⁰ We attribute this increase to greater realized importance and a higher level of confidence from nursing staff in recognizing and addressing delirium, as supported by survey data. While the increased screening of patients should be considered a positive outcome, it also poses the possibility that the observed decrease in delirium incidence in the postintervention group was in fact due to more CAMs performed on patients without delirium. Likewise, nurses may have become more adept at recognizing true delirium, as opposed to delirium mimics, in the latter period of the study.

Perhaps the greatest limitation of our study is the variability in performing and recording CAMs, as some patients had multiple CAMs recorded while others did not have any CAMs recorded. This may have been affected in part by the increase in COVID-19 cases in our hospital towards the latter half of the study, which resulted in changes in nursing assignments as well as patient comorbidities in ways that cannot be easily quantified. Given the limited size of our patient cohorts, certain outcomes, such as the use of sitters, physical restraints, and in-hospital mortality, were unable to be assessed for changes statistically. Causative relationships between our interventions and associated outcome measures are necessarily limited in a binary comparison between control and postintervention groups.

Within these limitations, our study demonstrates promising results in core dimensions of patient care. We anticipate further quality improvement initiatives involving greater numbers of nursing staff and patients to better quantify the impact of nonpharmacologic nursing-centered interventions for preventing hospital delirium.

Conclusion

A multimodal strategy involving nursing-focused training and nonpharmacologic interventions to address hospital delirium is associated with improved patient care outcomes and nursing confidence. Nurses play an integral role in the early recognition and prevention of hospital delirium, which directly translates to reducing burdens in both patient functionality and health care costs. Education and tools to equip nurses to perform standardized delirium screening and interventions should be prioritized.

Acknowledgment: The authors thanks Olena Svetlov, NP, Oscar Abarca, Jose Chavez, and Jenita Gutierrez.

Corresponding author: Jason Ching, MD, Department of Neurology, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Los Angeles, CA 90048; [email protected].

Financial disclosures: None.

Funding: This research was supported by NIH National Center for Advancing Translational Science (NCATS) UCLA CTSI Grant Number UL1TR001881.

Addition of echocardiogram measurement of biventricular dysfunction improved the accuracy of prognosis among patients with cardiac shock (CS) in the cardiac intensive care unit.

In patients in the cardiac ICU with CS, biventricular dysfunction (BVD), as assessed using transthoracic echocardiography, improves clinical risk stratification when combined with the Society for Cardiovascular Angiography and Interventions shock stage.

No improvements in risk stratification was seen with patients with left or right ventricular systolic dysfunction (LVSD or RVSD) alone, according to an article published in the journal Chest.

Ventricular systolic dysfunction is commonly seen in patients who have suffered cardiac shock, most often on the left side. Although echocardiography is often performed on these patients during diagnosis, previous studies looking at ventricular dysfunction used invasive hemodynamic parameters, which made it challenging to incorporate their findings into general cardiac ICU practice.
 

Pinning down cardiac shock

Although treatment of acute MI and heart failure has improved greatly, particularly with the implementation of percutaneous coronary intervention (primary PCI) for ST-segment elevation MI. This has reduced the rate of future heart failure, but cardiac shock can occur before or after the procedure, with a 30-day mortality of 30%-40%. This outcome hasn’t improved in the last 20 years.

Efforts to improve cardiac shock outcomes through percutaneous mechanical circulatory support devices have been hindered by the fact that CS patients are heterogeneous, and prognosis may depend on a range of factors.

SCAI was developed as a five-stage classification system for CS to improve communication of patient status, as well as to improve differentiation among patients participation in clinical trials. It does not include measures of ventricular dysfunction.
 

Simple measure boosts prognosis accuracy

The new work adds an additional layer to the SCAI shock stage. “Adding echocardiography allows discrimination between levels of risk for each SCAI stage,” said David Baran, MD, who was asked for comment. Dr. Baran was the lead author on the original SCAI study and is system director of advanced heart failure at Sentara Heart Hospital, as well as a professor of medicine at Eastern Virginia Medical School, both in Norfolk.

The work also underscores the value of repeated measures of prognosis during a patient’s stay in the ICU. “If a patient is not improving, it may prompt a consideration of whether transfer or consultation with a tertiary center may be of value. Conversely, if a patient doesn’t have high-risk features and is responding to therapy, it is reassuring to have data supporting low mortality with that care plan,” said Dr. Baran.

The study may be biased, since not every patient undergoes an echocardiogram. Still, “the authors make a convincing case that biventricular dysfunction is a powerful negative marker across the spectrum of SCAI stages,” said Dr. Baran.

Echocardiography is simple and generally available, and some are even portable and used with a smartphone. But patient body size interferes with echocardiography, as can the presence of a ventilator or multiple surgical dressings. “The key advantage of echo is that it is completely noninvasive and can be brought to the patient in the ICU, unlike other testing which involves moving the patient to the testing environment,” said Dr. Baran.

The researchers analyzed data from 3,158 patients admitted to the cardiac ICU at the Mayo Clinic Hospital St. Mary’s Campus in Rochester, Minn., 51.8% of whom had acute coronary syndromes. They defined LVSD as a left ventricular ejection fraction less than 40%, and RVSD as at least moderate systolic dysfunction determined by semiquantitative measurement. BVD constituted the presence of both LVSD and RVSD. They examined the association of in-hospital mortality with these parameters combined with SCAI stage.
 

BVD a risk factor

Overall in-hospital mortality was 10%. A total of 22.3% of patients had LVSD and 11.8% had RVSD; 16.4% had moderate or greater BVD. There was no association between LVSD or RVSD and in-hospital mortality after adjustment for SCAI stage, but there was a significant association for BVD (adjusted hazard ratio, 1.815; P = .0023). When combined with SCAI, BVC led to an improved ability to predict hospital mortality (area under the curve, 0.784 vs. 0.766; P < .001). Adding semiquantitative RVSD and LVSD led to more improvement (AUC, 0.794; P < .01 vs. both).

RVSD was associated with higher in-hospital mortality (adjusted odds ratio, 1.421; P = .02), and there was a trend toward greater mortality with LVSD (aOR, 1.336; P = .06). There was little change when SCAI shock stage A patients were excluded (aOR, 1.840; P < .001).

Patients with BVD had greater in-hospital mortality than those without ventricular dysfunction (aOR, 1.815; P = .0023), but other between-group comparisons were not significant.

The researchers performed a classification and regression tree analysis using left ventricular ejection fraction (LVEF) and semiquantitative RVSD. It found that RVSD was a better predictor of in-hospital mortality than LVSD, and the best cutoff for LVSD was different among patients with RVSD and patients without RVSD.

Patients with mild or greater RVD and LVEF greater than 24% were considered high risk; those with borderline or low RVSD and LVEF less than 33%, or mild or greater RVSD with LVEF of at least 24%, were considered intermediate risk. Patients with borderline or no RVSD and LVEF of at least 33% were considered low risk. Hospital mortality was 22% in the high-risk group, 12.2% in the intermediate group, and 3.3% in the low-risk group (aOR vs. intermediate, 0.493; P = .0006; aOR vs. high risk, 0.357; P < .0001).

The study authors disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.

Addition of echocardiogram measurement of biventricular dysfunction improved the accuracy of prognosis among patients with cardiac shock (CS) in the cardiac intensive care unit.

In patients in the cardiac ICU with CS, biventricular dysfunction (BVD), as assessed using transthoracic echocardiography, improves clinical risk stratification when combined with the Society for Cardiovascular Angiography and Interventions shock stage.

No improvements in risk stratification was seen with patients with left or right ventricular systolic dysfunction (LVSD or RVSD) alone, according to an article published in the journal Chest.

Ventricular systolic dysfunction is commonly seen in patients who have suffered cardiac shock, most often on the left side. Although echocardiography is often performed on these patients during diagnosis, previous studies looking at ventricular dysfunction used invasive hemodynamic parameters, which made it challenging to incorporate their findings into general cardiac ICU practice.
 

Pinning down cardiac shock

Although treatment of acute MI and heart failure has improved greatly, particularly with the implementation of percutaneous coronary intervention (primary PCI) for ST-segment elevation MI. This has reduced the rate of future heart failure, but cardiac shock can occur before or after the procedure, with a 30-day mortality of 30%-40%. This outcome hasn’t improved in the last 20 years.

Efforts to improve cardiac shock outcomes through percutaneous mechanical circulatory support devices have been hindered by the fact that CS patients are heterogeneous, and prognosis may depend on a range of factors.

SCAI was developed as a five-stage classification system for CS to improve communication of patient status, as well as to improve differentiation among patients participation in clinical trials. It does not include measures of ventricular dysfunction.
 

Simple measure boosts prognosis accuracy

The new work adds an additional layer to the SCAI shock stage. “Adding echocardiography allows discrimination between levels of risk for each SCAI stage,” said David Baran, MD, who was asked for comment. Dr. Baran was the lead author on the original SCAI study and is system director of advanced heart failure at Sentara Heart Hospital, as well as a professor of medicine at Eastern Virginia Medical School, both in Norfolk.

The work also underscores the value of repeated measures of prognosis during a patient’s stay in the ICU. “If a patient is not improving, it may prompt a consideration of whether transfer or consultation with a tertiary center may be of value. Conversely, if a patient doesn’t have high-risk features and is responding to therapy, it is reassuring to have data supporting low mortality with that care plan,” said Dr. Baran.

The study may be biased, since not every patient undergoes an echocardiogram. Still, “the authors make a convincing case that biventricular dysfunction is a powerful negative marker across the spectrum of SCAI stages,” said Dr. Baran.

Echocardiography is simple and generally available, and some are even portable and used with a smartphone. But patient body size interferes with echocardiography, as can the presence of a ventilator or multiple surgical dressings. “The key advantage of echo is that it is completely noninvasive and can be brought to the patient in the ICU, unlike other testing which involves moving the patient to the testing environment,” said Dr. Baran.

The researchers analyzed data from 3,158 patients admitted to the cardiac ICU at the Mayo Clinic Hospital St. Mary’s Campus in Rochester, Minn., 51.8% of whom had acute coronary syndromes. They defined LVSD as a left ventricular ejection fraction less than 40%, and RVSD as at least moderate systolic dysfunction determined by semiquantitative measurement. BVD constituted the presence of both LVSD and RVSD. They examined the association of in-hospital mortality with these parameters combined with SCAI stage.
 

BVD a risk factor

Overall in-hospital mortality was 10%. A total of 22.3% of patients had LVSD and 11.8% had RVSD; 16.4% had moderate or greater BVD. There was no association between LVSD or RVSD and in-hospital mortality after adjustment for SCAI stage, but there was a significant association for BVD (adjusted hazard ratio, 1.815; P = .0023). When combined with SCAI, BVC led to an improved ability to predict hospital mortality (area under the curve, 0.784 vs. 0.766; P < .001). Adding semiquantitative RVSD and LVSD led to more improvement (AUC, 0.794; P < .01 vs. both).

RVSD was associated with higher in-hospital mortality (adjusted odds ratio, 1.421; P = .02), and there was a trend toward greater mortality with LVSD (aOR, 1.336; P = .06). There was little change when SCAI shock stage A patients were excluded (aOR, 1.840; P < .001).

Patients with BVD had greater in-hospital mortality than those without ventricular dysfunction (aOR, 1.815; P = .0023), but other between-group comparisons were not significant.

The researchers performed a classification and regression tree analysis using left ventricular ejection fraction (LVEF) and semiquantitative RVSD. It found that RVSD was a better predictor of in-hospital mortality than LVSD, and the best cutoff for LVSD was different among patients with RVSD and patients without RVSD.

Patients with mild or greater RVD and LVEF greater than 24% were considered high risk; those with borderline or low RVSD and LVEF less than 33%, or mild or greater RVSD with LVEF of at least 24%, were considered intermediate risk. Patients with borderline or no RVSD and LVEF of at least 33% were considered low risk. Hospital mortality was 22% in the high-risk group, 12.2% in the intermediate group, and 3.3% in the low-risk group (aOR vs. intermediate, 0.493; P = .0006; aOR vs. high risk, 0.357; P < .0001).

The study authors disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.

Addition of echocardiogram measurement of biventricular dysfunction improved the accuracy of prognosis among patients with cardiac shock (CS) in the cardiac intensive care unit.

In patients in the cardiac ICU with CS, biventricular dysfunction (BVD), as assessed using transthoracic echocardiography, improves clinical risk stratification when combined with the Society for Cardiovascular Angiography and Interventions shock stage.

No improvements in risk stratification was seen with patients with left or right ventricular systolic dysfunction (LVSD or RVSD) alone, according to an article published in the journal Chest.

Ventricular systolic dysfunction is commonly seen in patients who have suffered cardiac shock, most often on the left side. Although echocardiography is often performed on these patients during diagnosis, previous studies looking at ventricular dysfunction used invasive hemodynamic parameters, which made it challenging to incorporate their findings into general cardiac ICU practice.
 

Pinning down cardiac shock

Although treatment of acute MI and heart failure has improved greatly, particularly with the implementation of percutaneous coronary intervention (primary PCI) for ST-segment elevation MI. This has reduced the rate of future heart failure, but cardiac shock can occur before or after the procedure, with a 30-day mortality of 30%-40%. This outcome hasn’t improved in the last 20 years.

Efforts to improve cardiac shock outcomes through percutaneous mechanical circulatory support devices have been hindered by the fact that CS patients are heterogeneous, and prognosis may depend on a range of factors.

SCAI was developed as a five-stage classification system for CS to improve communication of patient status, as well as to improve differentiation among patients participation in clinical trials. It does not include measures of ventricular dysfunction.
 

Simple measure boosts prognosis accuracy

The new work adds an additional layer to the SCAI shock stage. “Adding echocardiography allows discrimination between levels of risk for each SCAI stage,” said David Baran, MD, who was asked for comment. Dr. Baran was the lead author on the original SCAI study and is system director of advanced heart failure at Sentara Heart Hospital, as well as a professor of medicine at Eastern Virginia Medical School, both in Norfolk.

The work also underscores the value of repeated measures of prognosis during a patient’s stay in the ICU. “If a patient is not improving, it may prompt a consideration of whether transfer or consultation with a tertiary center may be of value. Conversely, if a patient doesn’t have high-risk features and is responding to therapy, it is reassuring to have data supporting low mortality with that care plan,” said Dr. Baran.

The study may be biased, since not every patient undergoes an echocardiogram. Still, “the authors make a convincing case that biventricular dysfunction is a powerful negative marker across the spectrum of SCAI stages,” said Dr. Baran.

Echocardiography is simple and generally available, and some are even portable and used with a smartphone. But patient body size interferes with echocardiography, as can the presence of a ventilator or multiple surgical dressings. “The key advantage of echo is that it is completely noninvasive and can be brought to the patient in the ICU, unlike other testing which involves moving the patient to the testing environment,” said Dr. Baran.

The researchers analyzed data from 3,158 patients admitted to the cardiac ICU at the Mayo Clinic Hospital St. Mary’s Campus in Rochester, Minn., 51.8% of whom had acute coronary syndromes. They defined LVSD as a left ventricular ejection fraction less than 40%, and RVSD as at least moderate systolic dysfunction determined by semiquantitative measurement. BVD constituted the presence of both LVSD and RVSD. They examined the association of in-hospital mortality with these parameters combined with SCAI stage.
 

BVD a risk factor

Overall in-hospital mortality was 10%. A total of 22.3% of patients had LVSD and 11.8% had RVSD; 16.4% had moderate or greater BVD. There was no association between LVSD or RVSD and in-hospital mortality after adjustment for SCAI stage, but there was a significant association for BVD (adjusted hazard ratio, 1.815; P = .0023). When combined with SCAI, BVC led to an improved ability to predict hospital mortality (area under the curve, 0.784 vs. 0.766; P < .001). Adding semiquantitative RVSD and LVSD led to more improvement (AUC, 0.794; P < .01 vs. both).

RVSD was associated with higher in-hospital mortality (adjusted odds ratio, 1.421; P = .02), and there was a trend toward greater mortality with LVSD (aOR, 1.336; P = .06). There was little change when SCAI shock stage A patients were excluded (aOR, 1.840; P < .001).

Patients with BVD had greater in-hospital mortality than those without ventricular dysfunction (aOR, 1.815; P = .0023), but other between-group comparisons were not significant.

The researchers performed a classification and regression tree analysis using left ventricular ejection fraction (LVEF) and semiquantitative RVSD. It found that RVSD was a better predictor of in-hospital mortality than LVSD, and the best cutoff for LVSD was different among patients with RVSD and patients without RVSD.

Patients with mild or greater RVD and LVEF greater than 24% were considered high risk; those with borderline or low RVSD and LVEF less than 33%, or mild or greater RVSD with LVEF of at least 24%, were considered intermediate risk. Patients with borderline or no RVSD and LVEF of at least 33% were considered low risk. Hospital mortality was 22% in the high-risk group, 12.2% in the intermediate group, and 3.3% in the low-risk group (aOR vs. intermediate, 0.493; P = .0006; aOR vs. high risk, 0.357; P < .0001).

The study authors disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.