Hospital Medicine

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized patients often receive intravenous fluids (IVF) when they cannot meet physiologic needs through oral intake in the setting of medical or surgical illness. Prescribing the optimal IVF solution to the appropriate patient is a complex decision and often occurs without the same degree of institutionalized restrictions or guidance developed for other inpatient pharmacologic agents. There is wide variation in clinical utilization of IVF due to the lack of data to guide decision making.¹ When data do exist, they typically focus on a limited number of clinical situations.² Thus, even though IVF are often considered low-risk, the frequency and lack of consistency with which they are used can result in errors, complications, and over-use of medical resources.³

(Evidence quality: not described in the guideline, recommendation strength: not described in the guideline)

Recommendation 1

To aid in fluid management and avoid complications, the guidelines recommend that patients on IVF require careful assessment of volume status, including a detailed history, physical exam, clinical monitoring, and daily labs.²

Clinical history should focus on understanding fluid losses and intake; physical exam should include vital signs, evidence of orthostatic hypotension, capillary refill, jugular venous pulsation, and assessment for pulmonary edema. Subsequent clinical monitoring should include fluid balance (Ins and Outs) and daily weights. All patients starting or continuing IVF should have a basic metabolic panel at least daily according to the guidelines, though the authors note this frequency may be too high for some patients and needs further study.²

Recommendation 2

The guidelines describe four types of IV fluids that can be administered: crystalloids, balanced crystalloids, glucose solutions, and non blood-product colloids.²

Crystalloids include isotonic saline with 154 millimoles (mmol) of sodium and chloride. Balanced crystalloids, such as lactated Ringer’s solution, are more physiologic, with less sodium and chloride, and the addition of magnesium, potassium, and calcium. Glucose solutions are quickly metabolized and, thus, are an effective way to deliver free water. Non blood-product colloids include particles that are retained within the circulation, including proteins such as human albumin.

Recommendation 3

For each indication to administer IVF, the guidelines recommend the following formulations and considerations:²

For general resuscitation, use crystalloids with sodium content of 130-154 mmol, delivered in a bolus of at least 500 milliliters (mL) over 15 minutes or less. For sepsis, infuse at least 30 mL/kg.⁴ For routine maintenance, restrict the volume to 25-30 mL/kg/day of water, and include 1 mmol/kg/day of potassium, sodium, and chloride along with 50-100 g/day of glucose to prevent starvation ketosis, though glucose should be avoided in most diabetic patients. With obesity, adjust the IVF to ideal body weight, and for patients who are older, frail, or admitted with renal or cardiac impairment, consider prescribing a lower range of fluid (20-25 mL/kg/day). For redistribution or replacement, use sodium chloride or balanced crystalloids or consider colloids, which have a theoretical advantage in expanding intravascular volume while limiting interstitial edema. Note that colloids are more expensive, and definitive evidence supporting increased efficacy is lacking. Clinicians should monitor closely for hypovolemia, hypervolemia, and electrolyte abnormalities, particularly hypo- and hypernatremia that carry associated mental status implications and risk of central pontine myelinolysis. The inadvertent overuse of IVF is common in hospital settings, particularly when maintenance fluids are not discontinued upon patient improvement or when patients move between care areas. Thus, regular clinical reassessment of volume status is important.

Recommendation 4

In both noncritically ill and critically ill hospitalized patients, there is a benefit to using balanced crystalloids compared to isotonic saline in preventing major adverse kidney events and death.^5,6

Two important studies in 2018 added new information to the existing NICE guidelines, addressing the previously unanswered question of the benefits of balanced crystalloids versus isotonic saline, one among non-critically ill patients and the other among critically ill patients.^5,6 Prior data suggested that the use of isotonic saline is associated with multiple complications, including hyperchloremic metabolic acidosis, acute kidney injury, and death. In the non-critically ill population, the use of balanced crystalloids resulted in lower incidence of major adverse kidney events (absolute difference of 0.9%), but did not change the number of hospital days (the primary outcome).⁵ In the critically ill population the use of balanced crystalloids resulted in lower rates of death, new renal replacement therapy, or persistent renal dysfunction,⁶ and the authors found preferential use of balanced crystalloids could prevent one out of every 94 patients admitted to the ICU from experiencing these adverse outcomes. Given the similar cost associated with isotonic saline and balanced crystalloids, these new findings suggest hospitalists should select balanced crystalloids if there is no compelling clinical reason to use isotonic saline.

While conflicts of interest are often a concern in clinical guidelines due to influence by pharmaceutical, device, and specialty interests, the United Kingdom’s National Clinical Guideline Centre (NGC), which developed the NICE guidelines, is hosted by the Royal College of Physicians and has governance partnerships with the Royal College of Surgeons of England, Royal College of General Practitioners, and Royal College of Nursing. Each guideline produced by the NGC is overseen by an independent guideline committee comprised of healthcare professionals and patient representatives, and as a result, concern for conflicts of interest is low.

The NICE guidelines were created by a multidisciplinary team from multiple clinical specialties, and reviewed evidence addressing both clinical and health economic outcomes. Importantly, data from randomized controlled studies was relatively limited. The data excluded patients under 16 years of age, pregnant women, and those with severe liver or renal disease, diabetes or burns, as well as those in intensive care settings. Unfortunately, many medical patients cared for by hospitalists fall into one or more of these categories, limiting applicability of the guidelines.

Two important studies in 2018 added new information to the existing NICE guidelines, as outlined in Recommendation 4.^5,6 Both of these studies occurred at a single institution, limiting their generalizability, though each study included a diverse patient population. In the ICU study, treating clinicians were aware of the composition of the assigned crystalloid so the decision to initiate renal-replacement therapy may have been susceptible to treatment bias. In addition, censoring of data collection at hospital discharge may have underestimated the true incidence of death at 30 days and overestimated persistent renal dysfunction at 30 days. Importantly, the trial design did not allow comparison of lactated Ringer’s solution versus Plasma-Lyte. The non-ICU study evaluated patients who began treatment in the emergency department and were subsequently admitted to non-ICU inpatient units—a population that mirrors much of hospitalist practice, however the un-blinded design makes bias a concern. Finally, lactated Ringer’s solution represented more than 95% of the balanced crystalloids used in the trial, so additional study is required to compare Plasma-Lyte with both saline and lactated Ringer’s solution.

More evidence is needed to better understand the appropriate use of IVF in specific clinical scenarios, including to determine if balanced solutions, as compared with isotonic saline, are superior across a spectrum of clinical conditions. For patients with an indication for maintenance fluid administration, determining if a higher sodium content reduces the risk of hyponatremia without increasing the risk of volume overload will help guide practice. Finally, more comprehensive study of the incidence of overuse and complications as a consequence of IVF, as well as the optimal frequency of lab monitoring, is needed to guide understanding of how practicing hospitalists and health systems can help reduce harm and waste

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.

Hospitalized children with inadequate fluid intake are often administered maintenance intravenous fluids (IVFs) to support metabolic needs and sensible losses. Historically, hypotonic IVFs have been the standard, based on theoretical water and electrolyte requirements for estimated energy expenditure.¹ However, when combined with increased levels of arginine vasopressin (AVP) seen in acutely ill children which impairs free-water excretion,² hypotonic IVF can result in hyponatremia. The recently published guideline by the American Academy of Pediatrics (AAP)³ is the first to provide an evidence-based recommendation on the use of maintenance IVF therapy in children.

Patients between the ages of 28 days and 18 years should receive isotonic solutions with appropriate potassium chloride and dextrose for maintenance IVFs (evidence quality: high; recommendation strength: strong)

Isotonic fluids, such as 0.9% NaCl (normal saline), Hartmann solution and PlasmaLyte, contain a sodium concentration similar to that of plasma (135-144 mEq/L). Lactated Ringer solution (LR) is near-isotonic (sodium 130 mEq/L), but was not used in any of the reviewed studies and therefore not included in the recommendation. Excluded are patients with neurosurgical disorders, congenital or acquired cardiac disease, hepatic disease, cancer, renal dysfunction, diabetes insipidus, voluminous watery diarrhea, severe burns, or patients in the neonatal intensive care unit.

The primary benefit of the AAP recommendation is the reduced risk of iatrogenic hyponatremia and its associated sequelae, including complications or impact on cost of care. The number needed to treat with isotonic fluids was 7.5 to prevent any hyponatremia and 27.8 to prevent moderate hyponatremia (<130 mEq/L). Increases in readmission rates, length of stay, and cost of hospitalization have been reported in a recent meta-analysis reviewing the economic burden of hyponatremia in both adults and children.⁴

Potential harms from the use of isotonic fluids include hypernatremia, hyperchloremic metabolic acidosis, and fluid overload, although available data have not demonstrated an increased risk of these complications. In light of a recent normal saline (NS) shortage in the United States, limited availability is also a consideration. Plasmalyte is more costly than NS and is currently incompatible with the addition of dextrose.

Methods in Preparing Guideline

The guideline development committee included broad representation by pediatric experts in primary care, hospital medicine, emergency medicine, critical care medicine, nephrology, anesthesiology, surgery and quality improvement, as well as a guideline methodologist/informatician and epidemiologist.

Search strategies from recently published systematic reviews of clinical trials comparing isotonic with hypotonic maintenance IVFs were used to identify studies eligible for inclusion. A total of 17 studies with 2,455 total patients were initially identified and included. One additional study meeting inclusion criteria was found after the committee convened and excluded from the guideline.⁵ Three reviewers from the subcommittee performed a structured critical appraisal of each article. The methods of each trial were assessed for risk-of-bias in multiple domains, including randomization, allocation concealment, performance, detection, attrition and reporting. Forest plots were generated using random-effects models and Mantel-Haenzel statistics with the outcome of hyponatremia. The guideline underwent review by various stakeholders including AAP councils, committees, and sections, and individuals considered experts in the field.

A strength of the guideline is the high quality of the evidence and the consistent findings. All of the included studies were randomized clinical trials and the number of included patients was large. Of the 17 included studies, 16 reported a risk ratio favoring isotonic fluids over hypotonic fluids in the prevention of developing hyponatremia; the results of the study that favored hypotonic fluids were not statistically significant on their own. A sensitivity analysis was performed to exclude one study with a 20% weight, determined by multiple factors such as sample size, confidence interval, and an unusually high rate of hyponatremia in the isotonic and hypotonic fluids groups (33.3 % and 70%, respectively).⁶ After exclusion, there was no change in the overall estimated risk in hypotonic fluids leading to hyponatremia. Only one trial had two sources of high risk of bias (allocation concealment, attrition) and the remaining had only low or unclear risk of biases in the various domains.

The study that was excluded due to its late identification similarly shows increased risk of hyponatremia in groups administered hypotonic fluids (risk ratio 6.5-8.5), and would likely not affect the estimated risk.⁵

Despite differences in types of patients enrolled, rate of administered fluids, type of IVF, frequency of lab testing, and study duration, the I² (degree of heterogeneity) of the forest plot of all included studies remained low at 14% and the increased risk of hyponatremia from hypotonic fluids remained consistent.

Due to study design differences, a limitation of the guideline is that no recommendation is made regarding the type of isotonic fluids and the rate of IVF administration. Additionally, due to the low frequency of clinically significant sequelae of hyponatremia, such as hyponatremic encephalopathy, it remains uncertain how many patients would need to be treated with isotonic fluids to prevent a rare but potentially devastating event.

Sources of Potential Conflict of Interest or Bias

The guideline was developed and funded by the AAP. A formal conflict of interest management policy was followed, and subcommittee members had no conflicts of interests or financial relationships relevant to the guideline to disclose.

Generalizability

Given the large number of patients included in the studies and heterogeneity of the population included, the recommendation applies to most patients cared for by pediatric hospitalists. Several patient exclusions relevant to the pediatric hospitalist deserve mention: neonates, kidney disease, and voluminous diarrhea. Neonates under the age of 28 days, including febrile neonates, are excluded from the guideline because of the immature concentrating abilities of neonatal kidneys. Patients with renal impairment were excluded from the guideline recommendation because several studies excluded patients with kidney disease. Hospitalists often care for children who sustain prerenal acute kidney injury from severe dehydration. In this condition, the kidney conserves water through the release of AVP. While an excluded population, these patients would be even more susceptible to develop hyponatremia if administered hypotonic fluids. Patients with “voluminous diarrhea” are excluded from the guideline because those with gastroenteritis with ongoing losses may require IVFs at rates higher than maintenance, and are particularly vulnerable to electrolyte derangements. The guideline, however, does not define voluminous diarrhea, leaving it to the discretion of the treating clinician.

Finally, it is critical to mention that IVF should be considered a therapy to be judiciously used, and discontinued when possible. While the guideline addresses the choice of fluid composition, alternatives to orally or enterally hydrate a patient are always preferred.

While the guideline strongly recommends isotonic fluids for maintenance therapy, the choice of isotonic fluid remains with the clinician. Most included studies used NS for their isotonic groups, but Hartmann’s solution and Plasmalyte were represented in a few studies. LR, one of the more widely used balanced solutions, though slightly hypotonic (130 mEq/L), was not studied. The exclusion of LR from the included studies is unfortunate, as the benefit of balanced solutions compared to NS after significant fluid resuscitation has been shown in the setting of severe sepsis and shock.⁷ Hyperchloremic metabolic acidosis after fluid resuscitation with NS has raised concern about continuing NS as maintenance fluid and possibly worsening acidosis or hyperchloremia and its adverse effects.⁸ Further studies on the potential benefit of LR as maintenance fluid, or the potential harms of unbalanced solutions as maintenance fluids in the setting of significant resuscitation are needed.

Disclosures

The authors have nothing to disclose.

“Better to remain silent and be thought a fool than to speak and remove all doubt..”
—Abraham Lincoln

Have you ever been in a meeting with a supervisor wondering when you will get a chance to speak? Or have you walked away from an interview not knowing much about the candidate because you were talking all the time? If so, it might be time to consider your TLR: Talking to Listening Ratio. The TLR is a leadership pearl of great value. By keeping track of how much you talk versus how much you listen, you learn how and when to keep quiet.

As Mark Goulston wrote, “There are three stages of speaking to other people. In the first stage, you are on task, relevant and concise . . . the second stage (is) when it feels so good to talk, you don’t even notice the other person is not listening. The third stage occurs after you have lost track of what you were saying and begin to realize you might need to reel the other person back in.” Rather than finding a way to re-engage the other person by giving them a chance to talk while you listen, “. . . the usual impulse is to talk even more in an effort to regain their interest.”¹

When you are talking, you are not listening—and when you are not listening, you are not learning. Executives who do all the talking at meetings do not have the opportunity to hear the ideas of others. Poor listening can make it appear as if you don’t care what others think. Worse, being a hypocompetent listener can turn you into an ineffective leader—one who does not have the trust or respect of others.

The TLR is highly relevant for hospitalists: physicians and nurses who do all the talking are not noticing what patients or families want to say or what potentially mistaken conclusions they are drawing. Similarly, quality improvement and patient safety champions who do all the talking are not discovering what frontline clinicians think about an initiative or what barriers need to be overcome for success. They are also not hearing novel approaches to the problem or different priorities that should be addressed instead.

Your goal: ensure that your TLR is less than 1. How? Make it a habit to reflect on your TLR after an encounter with a patient, colleague, or supervisor and ask yourself, “Did I listen well?” In addition to its value in monitoring your own talkativeness, use the TLR to measure others. For example, when interviewing a new hire, apply TLR to discover how much patience would be required to work with a candidate. We once interviewed a physician whose TLR was north of 20 . . . we passed on hiring them. The TLR is also helpful for managing meetings. If you find yourself in one with an over-talker (TLR >5), point to the agenda and redirect the discussion. If it’s a direct report or colleague that’s doing all the talking, remind them that you have another meeting in 30 minutes, so they will need to move things along. Better yet: share the TLR pearl with them so that they can reflect on their performance. If you’re dealing with an under-talker (eg, TLR<0.5), encourage them to voice their opinion. Who knows—you might learn a thing or two.

The most surprising aspect to us about TLR is how oblivious people tend to be about it. High TLR’ers have little idea about the effect they have on people while those with an extremely low TLR (less than 0.2) wonder why they didn’t get picked for a project or promotion. Aim for a TLR between 0.5 and 0.7. Doing so will make you a better leader and follower.

Disclosures

Drs. Saint and Chopra are co-authors of the upcoming book, “Thirty Rules for Healthcare Leaders,” from which this article is adapted. Both authors have no other relevant conflicts of interest.

“Better to remain silent and be thought a fool than to speak and remove all doubt..”
—Abraham Lincoln

Have you ever been in a meeting with a supervisor wondering when you will get a chance to speak? Or have you walked away from an interview not knowing much about the candidate because you were talking all the time? If so, it might be time to consider your TLR: Talking to Listening Ratio. The TLR is a leadership pearl of great value. By keeping track of how much you talk versus how much you listen, you learn how and when to keep quiet.

As Mark Goulston wrote, “There are three stages of speaking to other people. In the first stage, you are on task, relevant and concise . . . the second stage (is) when it feels so good to talk, you don’t even notice the other person is not listening. The third stage occurs after you have lost track of what you were saying and begin to realize you might need to reel the other person back in.” Rather than finding a way to re-engage the other person by giving them a chance to talk while you listen, “. . . the usual impulse is to talk even more in an effort to regain their interest.”¹

When you are talking, you are not listening—and when you are not listening, you are not learning. Executives who do all the talking at meetings do not have the opportunity to hear the ideas of others. Poor listening can make it appear as if you don’t care what others think. Worse, being a hypocompetent listener can turn you into an ineffective leader—one who does not have the trust or respect of others.

The TLR is highly relevant for hospitalists: physicians and nurses who do all the talking are not noticing what patients or families want to say or what potentially mistaken conclusions they are drawing. Similarly, quality improvement and patient safety champions who do all the talking are not discovering what frontline clinicians think about an initiative or what barriers need to be overcome for success. They are also not hearing novel approaches to the problem or different priorities that should be addressed instead.

Your goal: ensure that your TLR is less than 1. How? Make it a habit to reflect on your TLR after an encounter with a patient, colleague, or supervisor and ask yourself, “Did I listen well?” In addition to its value in monitoring your own talkativeness, use the TLR to measure others. For example, when interviewing a new hire, apply TLR to discover how much patience would be required to work with a candidate. We once interviewed a physician whose TLR was north of 20 . . . we passed on hiring them. The TLR is also helpful for managing meetings. If you find yourself in one with an over-talker (TLR >5), point to the agenda and redirect the discussion. If it’s a direct report or colleague that’s doing all the talking, remind them that you have another meeting in 30 minutes, so they will need to move things along. Better yet: share the TLR pearl with them so that they can reflect on their performance. If you’re dealing with an under-talker (eg, TLR<0.5), encourage them to voice their opinion. Who knows—you might learn a thing or two.

The most surprising aspect to us about TLR is how oblivious people tend to be about it. High TLR’ers have little idea about the effect they have on people while those with an extremely low TLR (less than 0.2) wonder why they didn’t get picked for a project or promotion. Aim for a TLR between 0.5 and 0.7. Doing so will make you a better leader and follower.

Disclosures

Drs. Saint and Chopra are co-authors of the upcoming book, “Thirty Rules for Healthcare Leaders,” from which this article is adapted. Both authors have no other relevant conflicts of interest.

“Better to remain silent and be thought a fool than to speak and remove all doubt..”
—Abraham Lincoln

Have you ever been in a meeting with a supervisor wondering when you will get a chance to speak? Or have you walked away from an interview not knowing much about the candidate because you were talking all the time? If so, it might be time to consider your TLR: Talking to Listening Ratio. The TLR is a leadership pearl of great value. By keeping track of how much you talk versus how much you listen, you learn how and when to keep quiet.

As Mark Goulston wrote, “There are three stages of speaking to other people. In the first stage, you are on task, relevant and concise . . . the second stage (is) when it feels so good to talk, you don’t even notice the other person is not listening. The third stage occurs after you have lost track of what you were saying and begin to realize you might need to reel the other person back in.” Rather than finding a way to re-engage the other person by giving them a chance to talk while you listen, “. . . the usual impulse is to talk even more in an effort to regain their interest.”¹

When you are talking, you are not listening—and when you are not listening, you are not learning. Executives who do all the talking at meetings do not have the opportunity to hear the ideas of others. Poor listening can make it appear as if you don’t care what others think. Worse, being a hypocompetent listener can turn you into an ineffective leader—one who does not have the trust or respect of others.

The TLR is highly relevant for hospitalists: physicians and nurses who do all the talking are not noticing what patients or families want to say or what potentially mistaken conclusions they are drawing. Similarly, quality improvement and patient safety champions who do all the talking are not discovering what frontline clinicians think about an initiative or what barriers need to be overcome for success. They are also not hearing novel approaches to the problem or different priorities that should be addressed instead.

Your goal: ensure that your TLR is less than 1. How? Make it a habit to reflect on your TLR after an encounter with a patient, colleague, or supervisor and ask yourself, “Did I listen well?” In addition to its value in monitoring your own talkativeness, use the TLR to measure others. For example, when interviewing a new hire, apply TLR to discover how much patience would be required to work with a candidate. We once interviewed a physician whose TLR was north of 20 . . . we passed on hiring them. The TLR is also helpful for managing meetings. If you find yourself in one with an over-talker (TLR >5), point to the agenda and redirect the discussion. If it’s a direct report or colleague that’s doing all the talking, remind them that you have another meeting in 30 minutes, so they will need to move things along. Better yet: share the TLR pearl with them so that they can reflect on their performance. If you’re dealing with an under-talker (eg, TLR<0.5), encourage them to voice their opinion. Who knows—you might learn a thing or two.

The most surprising aspect to us about TLR is how oblivious people tend to be about it. High TLR’ers have little idea about the effect they have on people while those with an extremely low TLR (less than 0.2) wonder why they didn’t get picked for a project or promotion. Aim for a TLR between 0.5 and 0.7. Doing so will make you a better leader and follower.

Disclosures

Drs. Saint and Chopra are co-authors of the upcoming book, “Thirty Rules for Healthcare Leaders,” from which this article is adapted. Both authors have no other relevant conflicts of interest.

Inspired by the ABIM Foundation’s Choosing Wisely® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 67-year-old man is admitted to a telemetry ward for an acute myocardial infarction and treated with percutaneous coronary intervention. He is currently on day three of antibiotics for a methicillin-resistant Staphylococcus aureus (MRSA) lower extremity soft tissue infection that is healing without a draining wound. He is placed on contact precautions based on institutional infection control guidelines. The hospitalist overhears members of the team commenting on having to don gowns to see this patient each day and wonders aloud whether care is impacted by the use of contact precautions.

Contact precautions (CP) for patients with methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus (VRE) infections are common in several hospitals. CP pose a significant burden to health systems, with an estimated 20%-25% of hospitalized patients on CP for MRSA or VRE alone.¹ CP are becoming increasingly more prevalent with state laws and the Veterans Affairs (VA) hospital system requiring active surveillance cultures (ASC) and subsequent CP when ASC are positive.²

Supporters highlight the utility of CP in preventing the spread of infection, controlling outbreaks, and protecting healthcare workers from certain transmissible diseases. The Centers for Disease Control and Prevention (CDC) recommended CP after prior studies demonstrated their effectiveness during outbreaks of transmissible infections.³ CP were included in bundles alongside interventions such as improving hand hygiene, chlorhexidine gluconate (CHG) bathing, and ASC with targeted or universal decolonization.² The VA MRSA bundle, for example, demonstrated a reduction of healthcare-associated MRSA in the ICU by 62% after implementation. The Society for Healthcare Epidemiology of America Research Network (SHEA) and the Infectious Diseases Society of America (IDSA) recommend CP for MRSA-infected and colonized patients in acute care settings to control outbreaks.^4,5 The CDC also has broad recommendations supporting CP for all patients infected and previously identified as being colonized with target multidrug-resistant organisms (MDROs) without identifying which are considered to be “targets.”⁶

Despite current guidelines, cluster-randomized trials have not shown a benefit of initiating CP over usual care for the prevention of acquiring MRSA or VRE in the hospital. One study demonstrated no change in MRSA and VRE acquisition with broad screening and subsequent CP.⁷ Another study evaluated a universal gown and glove policy in an ICU setting and found a reduction in MRSA acquisition, but no reduction in VRE acquisition.⁸ A third study investigated hand hygiene and daily CHG bathing and noted a reduction in MRSA transmission rates, where CP for screened colonized patients had no effect on transmission of MRSA or VRE.⁹

In addition, a prospective trial at a large academic center over two six-month intervals utilized universal gloving with emollient-impregnated gloves compared with CP and found no difference in MDRO acquisition. Universal gloving was associated with higher hand hygiene rates than CP.¹⁰ Another more recent retrospective observational study compared universal contact precautions (UCP) in ICUs to a historical nine-year baseline and concurrently to other nonuniversal CP ICUs. There was no significant decrease in MDROs during the UCP period compared with baseline or with non-UCP units.¹¹Further interest in and scrutiny of CP prompted a recently published meta-analysis of 14 studies in which CP were eliminated. The rates of transmission of MRSA, VRE, or other MDROs studied were not impacted by discontinuation.¹² One of the studies included two large academic medical centers and assessed the impact of discontinuing CP for endemic MRSA and VRE. The bundled intervention included the discontinuation of CP for all carriers of MRSA and VRE, except patients with draining wounds, maintaining high hand hygiene rates, and CHG baths for nearly all patients. There was no significant increase in transmission rates, and the intervention saved the health system an estimated $643,776 and 45,277 hours per year in healthcare worker time previously spent on donning and doffing personal protective equipment.¹³ Another large academic hospital published a time series approach of seven interventions to reduce healthcare-associated infections and noted no increase in MRSA or VRE transmission when CP were discontinued when combined with other horizontal preventions.¹⁴ Results were found to be similar in a high-risk population of patients with hematologic malignancies and hematopoietic stem cell transplantation, where both surveillance and CP for VRE were discontinued and did not impact the rates of VRE bacteremia.¹⁵

Multiple studies have examined the deleterious effects of CP, including a comprehensive systematic literature review of various adverse outcomes linked with CP.¹⁶ CP decrease the amount of time that healthcare workers (HCW) spend with patients,¹⁷ create delays at admission and discharge,¹⁸ increase symptoms of anxiety and depression in patients,^19,20 and decrease patient satisfaction with care.^21,22 In a study conducted at the Cleveland Clinic Hospital, physician communication, staff responsiveness, patients’ perception of cleanliness, and their willingness to recommend the hospital on the Hospital Consumer Assessment of Healthcare Providers and Systems survey were lower in each category for patients on CP when compared with patients not on CP.²² Patients who are on CP are six times more likely to experience an adverse event in the hospital, including falls and pressure ulcers.²³ A recent study from a large academic medical center demonstrated that noninfectious adverse events were reduced by 72% after discontinuing CP for MRSA and VRE. These events included postoperative respiratory failure, hemorrhage or hematoma, thrombosis, wound dehiscence, pressure ulcers, and falls or trauma.²⁴

The financial costs of unnecessary CP have also been studied. A recent retrospective study examining a large cohort of patients on CP for MRSA demonstrated that when compared with nonisolated patients, those on MRSA CP had a 30% increase in length of stay and a 43% increase in costs of care. Patients isolated for MRSA were 4.4% more likely than nonisolated individuals to be readmitted within 30 days after discharge, unrelated to MRSA.²⁵ These data contribute to the growing evidence that a conscientious, patient-centered approach to CP is preferred to overly broad policies that compromise patient safety.

Contact precautions for MRSA and VRE should be used to interrupt transmission during uncontrolled outbreaks, and in patients with open wounds, uncontained secretions, or incontinent diarrhea.

In addition, there are other commonly encountered organisms for which CP should be continued. CP should be used for active Clostridium difficile infection to prevent transmission. Due to the paucity of data regarding prevention of novel and highly resistant organisms and the complexity in treating these MDROs, it is reasonable to initiate CP in these cases.²⁶ Examples include active infection with multidrug resistance, including carbapenem-resistant Enterobacteriaceae, highly drug-resistant Pseudomonas aeruginosa, and other emerging MDROs such as vancomycin-resistant or -indeterminate S. aureus (VRSA or VISA) and Candida auris.²⁷ Limiting CP to instances where there is clear evidence to support will ensure patient safety and limit the harms associated with CP.

Horizontal prevention aims to reduce the burden of all microorganisms. This includes techniques such as hand hygiene, antimicrobial stewardship, CHG bathing, and environmental cleaning methods to decrease colonization of all MDROs in hospital rooms. Compared with vertical prevention strategies that use active surveillance testing for colonization and CP, horizontal interventions are the most effective means to reduce transmission of MDROs.²⁸ The simplest and the most well-studied method for reducing transmission of all organisms in the hospital remains hand hygiene.²⁹ High institutional hand hygiene rates of at least 90% are critical to the success of any initiative that seeks to eliminate CP.

CHG bathing has also been studied across multiple patient settings for reducing MRSA and VRE acquisition, catheter-associated urinary tract infections, and central line-associated bacterial infections.³⁰ In addition, hospital-wide daily CHG bathing has been associated with decreased C. difficile infection, and the baths were well tolerated by patients.³¹

SHEA recently released recommendations for timing of discontinuation of CP for patients with MDROs and emphasized that hospital systems must take an individual approach to discontinuing CP that takes into account local prevalence, risk, and resources.³² The decision to not place a patient on CP is one side of this high-value coin. The other side is knowing when it is appropriate to discontinue CP.

• Discontinue the use of CP for MRSA and VRE in hospitals with low endemic rates and high hand hygiene compliance.
• Improve horizontal preventions by promoting hand hygiene, antimicrobial stewardship, and considering CHG bathing for all patients.
• Create a systematic approach to discontinuing CP and compare transmission of MRSA and VRE rates through microbiology surveillance before and after discontinuation.

Contact precautions for MRSA and VRE are another example of a “Thing We Do for No Reason”. For most patients with MRSA and VRE, CP have not been shown to effectively reduce transmission. In addition, CP are expensive and associated with increased rates of patient adverse events. Hospitalists can lead the effort to ensure optimal hand hygiene and work with local infection control teams to reevaluate the utility of CP for patients with MRSA and VRE.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

There are no conflicts of interest for any authors, financial or other.

Inspired by the ABIM Foundation’s Choosing Wisely® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 67-year-old man is admitted to a telemetry ward for an acute myocardial infarction and treated with percutaneous coronary intervention. He is currently on day three of antibiotics for a methicillin-resistant Staphylococcus aureus (MRSA) lower extremity soft tissue infection that is healing without a draining wound. He is placed on contact precautions based on institutional infection control guidelines. The hospitalist overhears members of the team commenting on having to don gowns to see this patient each day and wonders aloud whether care is impacted by the use of contact precautions.

Contact precautions (CP) for patients with methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus (VRE) infections are common in several hospitals. CP pose a significant burden to health systems, with an estimated 20%-25% of hospitalized patients on CP for MRSA or VRE alone.¹ CP are becoming increasingly more prevalent with state laws and the Veterans Affairs (VA) hospital system requiring active surveillance cultures (ASC) and subsequent CP when ASC are positive.²

Supporters highlight the utility of CP in preventing the spread of infection, controlling outbreaks, and protecting healthcare workers from certain transmissible diseases. The Centers for Disease Control and Prevention (CDC) recommended CP after prior studies demonstrated their effectiveness during outbreaks of transmissible infections.³ CP were included in bundles alongside interventions such as improving hand hygiene, chlorhexidine gluconate (CHG) bathing, and ASC with targeted or universal decolonization.² The VA MRSA bundle, for example, demonstrated a reduction of healthcare-associated MRSA in the ICU by 62% after implementation. The Society for Healthcare Epidemiology of America Research Network (SHEA) and the Infectious Diseases Society of America (IDSA) recommend CP for MRSA-infected and colonized patients in acute care settings to control outbreaks.^4,5 The CDC also has broad recommendations supporting CP for all patients infected and previously identified as being colonized with target multidrug-resistant organisms (MDROs) without identifying which are considered to be “targets.”⁶

Despite current guidelines, cluster-randomized trials have not shown a benefit of initiating CP over usual care for the prevention of acquiring MRSA or VRE in the hospital. One study demonstrated no change in MRSA and VRE acquisition with broad screening and subsequent CP.⁷ Another study evaluated a universal gown and glove policy in an ICU setting and found a reduction in MRSA acquisition, but no reduction in VRE acquisition.⁸ A third study investigated hand hygiene and daily CHG bathing and noted a reduction in MRSA transmission rates, where CP for screened colonized patients had no effect on transmission of MRSA or VRE.⁹

In addition, a prospective trial at a large academic center over two six-month intervals utilized universal gloving with emollient-impregnated gloves compared with CP and found no difference in MDRO acquisition. Universal gloving was associated with higher hand hygiene rates than CP.¹⁰ Another more recent retrospective observational study compared universal contact precautions (UCP) in ICUs to a historical nine-year baseline and concurrently to other nonuniversal CP ICUs. There was no significant decrease in MDROs during the UCP period compared with baseline or with non-UCP units.¹¹Further interest in and scrutiny of CP prompted a recently published meta-analysis of 14 studies in which CP were eliminated. The rates of transmission of MRSA, VRE, or other MDROs studied were not impacted by discontinuation.¹² One of the studies included two large academic medical centers and assessed the impact of discontinuing CP for endemic MRSA and VRE. The bundled intervention included the discontinuation of CP for all carriers of MRSA and VRE, except patients with draining wounds, maintaining high hand hygiene rates, and CHG baths for nearly all patients. There was no significant increase in transmission rates, and the intervention saved the health system an estimated $643,776 and 45,277 hours per year in healthcare worker time previously spent on donning and doffing personal protective equipment.¹³ Another large academic hospital published a time series approach of seven interventions to reduce healthcare-associated infections and noted no increase in MRSA or VRE transmission when CP were discontinued when combined with other horizontal preventions.¹⁴ Results were found to be similar in a high-risk population of patients with hematologic malignancies and hematopoietic stem cell transplantation, where both surveillance and CP for VRE were discontinued and did not impact the rates of VRE bacteremia.¹⁵

Multiple studies have examined the deleterious effects of CP, including a comprehensive systematic literature review of various adverse outcomes linked with CP.¹⁶ CP decrease the amount of time that healthcare workers (HCW) spend with patients,¹⁷ create delays at admission and discharge,¹⁸ increase symptoms of anxiety and depression in patients,^19,20 and decrease patient satisfaction with care.^21,22 In a study conducted at the Cleveland Clinic Hospital, physician communication, staff responsiveness, patients’ perception of cleanliness, and their willingness to recommend the hospital on the Hospital Consumer Assessment of Healthcare Providers and Systems survey were lower in each category for patients on CP when compared with patients not on CP.²² Patients who are on CP are six times more likely to experience an adverse event in the hospital, including falls and pressure ulcers.²³ A recent study from a large academic medical center demonstrated that noninfectious adverse events were reduced by 72% after discontinuing CP for MRSA and VRE. These events included postoperative respiratory failure, hemorrhage or hematoma, thrombosis, wound dehiscence, pressure ulcers, and falls or trauma.²⁴

The financial costs of unnecessary CP have also been studied. A recent retrospective study examining a large cohort of patients on CP for MRSA demonstrated that when compared with nonisolated patients, those on MRSA CP had a 30% increase in length of stay and a 43% increase in costs of care. Patients isolated for MRSA were 4.4% more likely than nonisolated individuals to be readmitted within 30 days after discharge, unrelated to MRSA.²⁵ These data contribute to the growing evidence that a conscientious, patient-centered approach to CP is preferred to overly broad policies that compromise patient safety.

Contact precautions for MRSA and VRE should be used to interrupt transmission during uncontrolled outbreaks, and in patients with open wounds, uncontained secretions, or incontinent diarrhea.

In addition, there are other commonly encountered organisms for which CP should be continued. CP should be used for active Clostridium difficile infection to prevent transmission. Due to the paucity of data regarding prevention of novel and highly resistant organisms and the complexity in treating these MDROs, it is reasonable to initiate CP in these cases.²⁶ Examples include active infection with multidrug resistance, including carbapenem-resistant Enterobacteriaceae, highly drug-resistant Pseudomonas aeruginosa, and other emerging MDROs such as vancomycin-resistant or -indeterminate S. aureus (VRSA or VISA) and Candida auris.²⁷ Limiting CP to instances where there is clear evidence to support will ensure patient safety and limit the harms associated with CP.

Horizontal prevention aims to reduce the burden of all microorganisms. This includes techniques such as hand hygiene, antimicrobial stewardship, CHG bathing, and environmental cleaning methods to decrease colonization of all MDROs in hospital rooms. Compared with vertical prevention strategies that use active surveillance testing for colonization and CP, horizontal interventions are the most effective means to reduce transmission of MDROs.²⁸ The simplest and the most well-studied method for reducing transmission of all organisms in the hospital remains hand hygiene.²⁹ High institutional hand hygiene rates of at least 90% are critical to the success of any initiative that seeks to eliminate CP.

CHG bathing has also been studied across multiple patient settings for reducing MRSA and VRE acquisition, catheter-associated urinary tract infections, and central line-associated bacterial infections.³⁰ In addition, hospital-wide daily CHG bathing has been associated with decreased C. difficile infection, and the baths were well tolerated by patients.³¹

SHEA recently released recommendations for timing of discontinuation of CP for patients with MDROs and emphasized that hospital systems must take an individual approach to discontinuing CP that takes into account local prevalence, risk, and resources.³² The decision to not place a patient on CP is one side of this high-value coin. The other side is knowing when it is appropriate to discontinue CP.

• Discontinue the use of CP for MRSA and VRE in hospitals with low endemic rates and high hand hygiene compliance.
• Improve horizontal preventions by promoting hand hygiene, antimicrobial stewardship, and considering CHG bathing for all patients.
• Create a systematic approach to discontinuing CP and compare transmission of MRSA and VRE rates through microbiology surveillance before and after discontinuation.

Contact precautions for MRSA and VRE are another example of a “Thing We Do for No Reason”. For most patients with MRSA and VRE, CP have not been shown to effectively reduce transmission. In addition, CP are expensive and associated with increased rates of patient adverse events. Hospitalists can lead the effort to ensure optimal hand hygiene and work with local infection control teams to reevaluate the utility of CP for patients with MRSA and VRE.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

There are no conflicts of interest for any authors, financial or other.

Inspired by the ABIM Foundation’s Choosing Wisely® campaign, the “Things We Do for No Reason” (TWDFNR) series reviews practices that have become common parts of hospital care but may provide little value to our patients. Practices reviewed in the TWDFNR series do not represent “black and white” conclusions or clinical practice standards but are meant as a starting place for research and active discussions among hospitalists and patients. We invite you to be part of that discussion. https://www.choosingwisely.org/

A 67-year-old man is admitted to a telemetry ward for an acute myocardial infarction and treated with percutaneous coronary intervention. He is currently on day three of antibiotics for a methicillin-resistant Staphylococcus aureus (MRSA) lower extremity soft tissue infection that is healing without a draining wound. He is placed on contact precautions based on institutional infection control guidelines. The hospitalist overhears members of the team commenting on having to don gowns to see this patient each day and wonders aloud whether care is impacted by the use of contact precautions.

Contact precautions (CP) for patients with methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus (VRE) infections are common in several hospitals. CP pose a significant burden to health systems, with an estimated 20%-25% of hospitalized patients on CP for MRSA or VRE alone.¹ CP are becoming increasingly more prevalent with state laws and the Veterans Affairs (VA) hospital system requiring active surveillance cultures (ASC) and subsequent CP when ASC are positive.²

Supporters highlight the utility of CP in preventing the spread of infection, controlling outbreaks, and protecting healthcare workers from certain transmissible diseases. The Centers for Disease Control and Prevention (CDC) recommended CP after prior studies demonstrated their effectiveness during outbreaks of transmissible infections.³ CP were included in bundles alongside interventions such as improving hand hygiene, chlorhexidine gluconate (CHG) bathing, and ASC with targeted or universal decolonization.² The VA MRSA bundle, for example, demonstrated a reduction of healthcare-associated MRSA in the ICU by 62% after implementation. The Society for Healthcare Epidemiology of America Research Network (SHEA) and the Infectious Diseases Society of America (IDSA) recommend CP for MRSA-infected and colonized patients in acute care settings to control outbreaks.^4,5 The CDC also has broad recommendations supporting CP for all patients infected and previously identified as being colonized with target multidrug-resistant organisms (MDROs) without identifying which are considered to be “targets.”⁶

Despite current guidelines, cluster-randomized trials have not shown a benefit of initiating CP over usual care for the prevention of acquiring MRSA or VRE in the hospital. One study demonstrated no change in MRSA and VRE acquisition with broad screening and subsequent CP.⁷ Another study evaluated a universal gown and glove policy in an ICU setting and found a reduction in MRSA acquisition, but no reduction in VRE acquisition.⁸ A third study investigated hand hygiene and daily CHG bathing and noted a reduction in MRSA transmission rates, where CP for screened colonized patients had no effect on transmission of MRSA or VRE.⁹

In addition, a prospective trial at a large academic center over two six-month intervals utilized universal gloving with emollient-impregnated gloves compared with CP and found no difference in MDRO acquisition. Universal gloving was associated with higher hand hygiene rates than CP.¹⁰ Another more recent retrospective observational study compared universal contact precautions (UCP) in ICUs to a historical nine-year baseline and concurrently to other nonuniversal CP ICUs. There was no significant decrease in MDROs during the UCP period compared with baseline or with non-UCP units.¹¹Further interest in and scrutiny of CP prompted a recently published meta-analysis of 14 studies in which CP were eliminated. The rates of transmission of MRSA, VRE, or other MDROs studied were not impacted by discontinuation.¹² One of the studies included two large academic medical centers and assessed the impact of discontinuing CP for endemic MRSA and VRE. The bundled intervention included the discontinuation of CP for all carriers of MRSA and VRE, except patients with draining wounds, maintaining high hand hygiene rates, and CHG baths for nearly all patients. There was no significant increase in transmission rates, and the intervention saved the health system an estimated $643,776 and 45,277 hours per year in healthcare worker time previously spent on donning and doffing personal protective equipment.¹³ Another large academic hospital published a time series approach of seven interventions to reduce healthcare-associated infections and noted no increase in MRSA or VRE transmission when CP were discontinued when combined with other horizontal preventions.¹⁴ Results were found to be similar in a high-risk population of patients with hematologic malignancies and hematopoietic stem cell transplantation, where both surveillance and CP for VRE were discontinued and did not impact the rates of VRE bacteremia.¹⁵

Multiple studies have examined the deleterious effects of CP, including a comprehensive systematic literature review of various adverse outcomes linked with CP.¹⁶ CP decrease the amount of time that healthcare workers (HCW) spend with patients,¹⁷ create delays at admission and discharge,¹⁸ increase symptoms of anxiety and depression in patients,^19,20 and decrease patient satisfaction with care.^21,22 In a study conducted at the Cleveland Clinic Hospital, physician communication, staff responsiveness, patients’ perception of cleanliness, and their willingness to recommend the hospital on the Hospital Consumer Assessment of Healthcare Providers and Systems survey were lower in each category for patients on CP when compared with patients not on CP.²² Patients who are on CP are six times more likely to experience an adverse event in the hospital, including falls and pressure ulcers.²³ A recent study from a large academic medical center demonstrated that noninfectious adverse events were reduced by 72% after discontinuing CP for MRSA and VRE. These events included postoperative respiratory failure, hemorrhage or hematoma, thrombosis, wound dehiscence, pressure ulcers, and falls or trauma.²⁴

The financial costs of unnecessary CP have also been studied. A recent retrospective study examining a large cohort of patients on CP for MRSA demonstrated that when compared with nonisolated patients, those on MRSA CP had a 30% increase in length of stay and a 43% increase in costs of care. Patients isolated for MRSA were 4.4% more likely than nonisolated individuals to be readmitted within 30 days after discharge, unrelated to MRSA.²⁵ These data contribute to the growing evidence that a conscientious, patient-centered approach to CP is preferred to overly broad policies that compromise patient safety.

Contact precautions for MRSA and VRE should be used to interrupt transmission during uncontrolled outbreaks, and in patients with open wounds, uncontained secretions, or incontinent diarrhea.

In addition, there are other commonly encountered organisms for which CP should be continued. CP should be used for active Clostridium difficile infection to prevent transmission. Due to the paucity of data regarding prevention of novel and highly resistant organisms and the complexity in treating these MDROs, it is reasonable to initiate CP in these cases.²⁶ Examples include active infection with multidrug resistance, including carbapenem-resistant Enterobacteriaceae, highly drug-resistant Pseudomonas aeruginosa, and other emerging MDROs such as vancomycin-resistant or -indeterminate S. aureus (VRSA or VISA) and Candida auris.²⁷ Limiting CP to instances where there is clear evidence to support will ensure patient safety and limit the harms associated with CP.

Horizontal prevention aims to reduce the burden of all microorganisms. This includes techniques such as hand hygiene, antimicrobial stewardship, CHG bathing, and environmental cleaning methods to decrease colonization of all MDROs in hospital rooms. Compared with vertical prevention strategies that use active surveillance testing for colonization and CP, horizontal interventions are the most effective means to reduce transmission of MDROs.²⁸ The simplest and the most well-studied method for reducing transmission of all organisms in the hospital remains hand hygiene.²⁹ High institutional hand hygiene rates of at least 90% are critical to the success of any initiative that seeks to eliminate CP.

CHG bathing has also been studied across multiple patient settings for reducing MRSA and VRE acquisition, catheter-associated urinary tract infections, and central line-associated bacterial infections.³⁰ In addition, hospital-wide daily CHG bathing has been associated with decreased C. difficile infection, and the baths were well tolerated by patients.³¹

SHEA recently released recommendations for timing of discontinuation of CP for patients with MDROs and emphasized that hospital systems must take an individual approach to discontinuing CP that takes into account local prevalence, risk, and resources.³² The decision to not place a patient on CP is one side of this high-value coin. The other side is knowing when it is appropriate to discontinue CP.

• Discontinue the use of CP for MRSA and VRE in hospitals with low endemic rates and high hand hygiene compliance.
• Improve horizontal preventions by promoting hand hygiene, antimicrobial stewardship, and considering CHG bathing for all patients.
• Create a systematic approach to discontinuing CP and compare transmission of MRSA and VRE rates through microbiology surveillance before and after discontinuation.

Contact precautions for MRSA and VRE are another example of a “Thing We Do for No Reason”. For most patients with MRSA and VRE, CP have not been shown to effectively reduce transmission. In addition, CP are expensive and associated with increased rates of patient adverse events. Hospitalists can lead the effort to ensure optimal hand hygiene and work with local infection control teams to reevaluate the utility of CP for patients with MRSA and VRE.

Do you think this is a low-value practice? Is this truly a “Thing We Do for No Reason?” Share what you do in your practice and join in the conversation online by retweeting it on Twitter (#TWDFNR) and liking it on Facebook. We invite you to propose ideas for other “Things We Do for No Reason” topics by emailing [email protected].

Disclosures

There are no conflicts of interest for any authors, financial or other.

Despite media attention to gender inequality in multiple professions, medicine has only recently begun to identify disparities facing women in academic medicine, focusing primarily on women faculty rather than trainees. Because of the unique and poorly understood juxtaposition of forces affecting their experience, focusing on women medical trainees may provide a representative framework to understand the larger, complex problem of gender equity in medicine. Rather than being a complicated problem with component parts that can be separately addressed, gender equity in medicine is a complex problem—one composed of a myriad of interrelated human and systemic factors. Such a complex problem demands innovative, open-minded, user-centered interventions. Here, we outline some of the factors unique to women trainees, including lack of female role models in leadership, gender bias, sexual harassment, work-life imbalance, and few formal leadership training programs. We propose one potential strategy, leadership programs specifically targeted to women residents and fellows. We recently implemented this strategy at our institution in the form of a day-long symposium of skill-building sessions for women residents and fellows.

Although women have achieved equal representation in several medical training programs, there is still a dearth of women in high-profile leadership positions within academic medicine. Although women comprised 46% of United States medical school applicants and residents in 2015-2016, underrepresentation persists at the level of associate professor (35% women), full professor (22%), department chair (14%), and dean (16%).¹ Many potential women leaders may not self-identify as such due to the limited exposure to women role models in positions of power and may in fact be ready for leadership roles earlier but not apply until later in their careers as compared with men.^2,3 The lack of role models with a shared background is an even more severe problem for women of color and all of these factors contribute to the “leaky pipeline” phenomenon.⁴ We aimed to address this mindset and help women see themselves as leaders by overcoming “second-generation gender bias” through our work.²

Due to the intense and inflexible nature of residency and fellowship training programs, many women choose to defer milestones such as childbearing.⁵ Women trainees are more likely than their male colleagues to avoid having a child during residency due to perceived threat to their career and negative perceptions of colleagues.^5,6 Women who are in a domestic partnership often bear the brunt of the household work regardless of the careers of the two partners, a phenomenon termed the “second shift.”⁷ This work-life imbalance has been shown to correlate with depressive symptoms in women internal medicine trainees.⁸

Recently, a trainee published on the experience of medical residents being asked whether they were ever called “nurse.” All the women in the room put up their hands; none of their male colleagues did.⁹ At issue is not the relative importance of the professions of medicine and nursing, but rather the gender stereotypes in medicine that lead to automatic categorization of women into one group. Although the majority of women residents likely have had personal experiences with bias and microaggressions, few are explicitly taught the tools to address these. Beyond microaggressions, women trainees are also subject to more sexual harassment than their male colleagues.¹⁰ In addition, women living at the intersections of race, ethnicity, and gender are faced with even higher rates of harassment.¹¹ Reporting sexual assault and harassment can be particularly difficult as a trainee because of the risk of retaliation, fear of poor evaluations from superiors, and lack of confidence in the reporting process.¹⁰

Finally, women trainees often receive little training about the skills required for career advancement to achieve parity with their male colleagues. Women are less likely to negotiate due to concerns about backlash or due to general lack of awareness about the importance of negotiation.¹² Women are asked to volunteer for “nonpromotable” tasks more often than men by colleagues of both sexes, a barrier to women reaching their full career potential and a difficult workplace scenario to navigate.¹³ Unlike the fields of business, law, and technology, for example, women in medicine do not routinely have training courses that incorporate skills such as navigating difficult conversations, conflict resolution, curriculum vitae writing, and negotiation. Various solutions have been offered to address some of the barriers facing women in medicine (such as the Association of American Medical Colleges and Executive Leadership in Academic Medicine leadership courses), but typically these focus on faculty rather than trainees. Given that women physicians practicing in the inpatient setting have been shown to have better patient outcomes¹⁴ and organizations with female leadership outperform those led by men,¹⁵ equipping our women trainees to thrive in the clinical and leadership environments is an essential step in fulfilling our mission as high-quality training programs.

At our institution, we recognized the need for training in leadership skills for women medical trainees and designed a day-long symposium for internal medicine women residents and fellows. Before developing the curriculum, we conducted a needs assessment to ascertain which skills women wanted to develop; women overwhelmingly wanted to learn about public speaking skills, work-life integration, and mentoring. Based on these responses, a group spanning multiple levels of training (residency, fellowship, and faculty) designed a combination of large-group lectures and small-group workshops termed “Women in Leadership Development” (WILD). The day-long curriculum included sessions on public speaking skills, women as change agents, mentorship, conflict resolution, and addressing microaggressions and concluded with a networking event for women faculty and trainees (Table).

Acknowledgments

The authors are indebted to Catherine Lucey, MD, for her framing of the issues for women in medicine as a complex problem and to Sarah Schaeffer, MD, for her unpublished data on microaggressions at our institution. The authors are also grateful to the UCSF Department of Medicine and the UCSF Chancellor’s Advisory Committee on the Status of Women for their financial support of the WILD (Women In Leadership Development) program.

Disclosures

The authors declare no conflict of interest.

Funding

The authors report no external funding source for this study.

Despite media attention to gender inequality in multiple professions, medicine has only recently begun to identify disparities facing women in academic medicine, focusing primarily on women faculty rather than trainees. Because of the unique and poorly understood juxtaposition of forces affecting their experience, focusing on women medical trainees may provide a representative framework to understand the larger, complex problem of gender equity in medicine. Rather than being a complicated problem with component parts that can be separately addressed, gender equity in medicine is a complex problem—one composed of a myriad of interrelated human and systemic factors. Such a complex problem demands innovative, open-minded, user-centered interventions. Here, we outline some of the factors unique to women trainees, including lack of female role models in leadership, gender bias, sexual harassment, work-life imbalance, and few formal leadership training programs. We propose one potential strategy, leadership programs specifically targeted to women residents and fellows. We recently implemented this strategy at our institution in the form of a day-long symposium of skill-building sessions for women residents and fellows.

Although women have achieved equal representation in several medical training programs, there is still a dearth of women in high-profile leadership positions within academic medicine. Although women comprised 46% of United States medical school applicants and residents in 2015-2016, underrepresentation persists at the level of associate professor (35% women), full professor (22%), department chair (14%), and dean (16%).¹ Many potential women leaders may not self-identify as such due to the limited exposure to women role models in positions of power and may in fact be ready for leadership roles earlier but not apply until later in their careers as compared with men.^2,3 The lack of role models with a shared background is an even more severe problem for women of color and all of these factors contribute to the “leaky pipeline” phenomenon.⁴ We aimed to address this mindset and help women see themselves as leaders by overcoming “second-generation gender bias” through our work.²

Due to the intense and inflexible nature of residency and fellowship training programs, many women choose to defer milestones such as childbearing.⁵ Women trainees are more likely than their male colleagues to avoid having a child during residency due to perceived threat to their career and negative perceptions of colleagues.^5,6 Women who are in a domestic partnership often bear the brunt of the household work regardless of the careers of the two partners, a phenomenon termed the “second shift.”⁷ This work-life imbalance has been shown to correlate with depressive symptoms in women internal medicine trainees.⁸

Recently, a trainee published on the experience of medical residents being asked whether they were ever called “nurse.” All the women in the room put up their hands; none of their male colleagues did.⁹ At issue is not the relative importance of the professions of medicine and nursing, but rather the gender stereotypes in medicine that lead to automatic categorization of women into one group. Although the majority of women residents likely have had personal experiences with bias and microaggressions, few are explicitly taught the tools to address these. Beyond microaggressions, women trainees are also subject to more sexual harassment than their male colleagues.¹⁰ In addition, women living at the intersections of race, ethnicity, and gender are faced with even higher rates of harassment.¹¹ Reporting sexual assault and harassment can be particularly difficult as a trainee because of the risk of retaliation, fear of poor evaluations from superiors, and lack of confidence in the reporting process.¹⁰

Finally, women trainees often receive little training about the skills required for career advancement to achieve parity with their male colleagues. Women are less likely to negotiate due to concerns about backlash or due to general lack of awareness about the importance of negotiation.¹² Women are asked to volunteer for “nonpromotable” tasks more often than men by colleagues of both sexes, a barrier to women reaching their full career potential and a difficult workplace scenario to navigate.¹³ Unlike the fields of business, law, and technology, for example, women in medicine do not routinely have training courses that incorporate skills such as navigating difficult conversations, conflict resolution, curriculum vitae writing, and negotiation. Various solutions have been offered to address some of the barriers facing women in medicine (such as the Association of American Medical Colleges and Executive Leadership in Academic Medicine leadership courses), but typically these focus on faculty rather than trainees. Given that women physicians practicing in the inpatient setting have been shown to have better patient outcomes¹⁴ and organizations with female leadership outperform those led by men,¹⁵ equipping our women trainees to thrive in the clinical and leadership environments is an essential step in fulfilling our mission as high-quality training programs.

At our institution, we recognized the need for training in leadership skills for women medical trainees and designed a day-long symposium for internal medicine women residents and fellows. Before developing the curriculum, we conducted a needs assessment to ascertain which skills women wanted to develop; women overwhelmingly wanted to learn about public speaking skills, work-life integration, and mentoring. Based on these responses, a group spanning multiple levels of training (residency, fellowship, and faculty) designed a combination of large-group lectures and small-group workshops termed “Women in Leadership Development” (WILD). The day-long curriculum included sessions on public speaking skills, women as change agents, mentorship, conflict resolution, and addressing microaggressions and concluded with a networking event for women faculty and trainees (Table).

Acknowledgments

The authors are indebted to Catherine Lucey, MD, for her framing of the issues for women in medicine as a complex problem and to Sarah Schaeffer, MD, for her unpublished data on microaggressions at our institution. The authors are also grateful to the UCSF Department of Medicine and the UCSF Chancellor’s Advisory Committee on the Status of Women for their financial support of the WILD (Women In Leadership Development) program.

Disclosures

The authors declare no conflict of interest.

Funding

The authors report no external funding source for this study.

Despite media attention to gender inequality in multiple professions, medicine has only recently begun to identify disparities facing women in academic medicine, focusing primarily on women faculty rather than trainees. Because of the unique and poorly understood juxtaposition of forces affecting their experience, focusing on women medical trainees may provide a representative framework to understand the larger, complex problem of gender equity in medicine. Rather than being a complicated problem with component parts that can be separately addressed, gender equity in medicine is a complex problem—one composed of a myriad of interrelated human and systemic factors. Such a complex problem demands innovative, open-minded, user-centered interventions. Here, we outline some of the factors unique to women trainees, including lack of female role models in leadership, gender bias, sexual harassment, work-life imbalance, and few formal leadership training programs. We propose one potential strategy, leadership programs specifically targeted to women residents and fellows. We recently implemented this strategy at our institution in the form of a day-long symposium of skill-building sessions for women residents and fellows.

Although women have achieved equal representation in several medical training programs, there is still a dearth of women in high-profile leadership positions within academic medicine. Although women comprised 46% of United States medical school applicants and residents in 2015-2016, underrepresentation persists at the level of associate professor (35% women), full professor (22%), department chair (14%), and dean (16%).¹ Many potential women leaders may not self-identify as such due to the limited exposure to women role models in positions of power and may in fact be ready for leadership roles earlier but not apply until later in their careers as compared with men.^2,3 The lack of role models with a shared background is an even more severe problem for women of color and all of these factors contribute to the “leaky pipeline” phenomenon.⁴ We aimed to address this mindset and help women see themselves as leaders by overcoming “second-generation gender bias” through our work.²

Due to the intense and inflexible nature of residency and fellowship training programs, many women choose to defer milestones such as childbearing.⁵ Women trainees are more likely than their male colleagues to avoid having a child during residency due to perceived threat to their career and negative perceptions of colleagues.^5,6 Women who are in a domestic partnership often bear the brunt of the household work regardless of the careers of the two partners, a phenomenon termed the “second shift.”⁷ This work-life imbalance has been shown to correlate with depressive symptoms in women internal medicine trainees.⁸

Recently, a trainee published on the experience of medical residents being asked whether they were ever called “nurse.” All the women in the room put up their hands; none of their male colleagues did.⁹ At issue is not the relative importance of the professions of medicine and nursing, but rather the gender stereotypes in medicine that lead to automatic categorization of women into one group. Although the majority of women residents likely have had personal experiences with bias and microaggressions, few are explicitly taught the tools to address these. Beyond microaggressions, women trainees are also subject to more sexual harassment than their male colleagues.¹⁰ In addition, women living at the intersections of race, ethnicity, and gender are faced with even higher rates of harassment.¹¹ Reporting sexual assault and harassment can be particularly difficult as a trainee because of the risk of retaliation, fear of poor evaluations from superiors, and lack of confidence in the reporting process.¹⁰

Finally, women trainees often receive little training about the skills required for career advancement to achieve parity with their male colleagues. Women are less likely to negotiate due to concerns about backlash or due to general lack of awareness about the importance of negotiation.¹² Women are asked to volunteer for “nonpromotable” tasks more often than men by colleagues of both sexes, a barrier to women reaching their full career potential and a difficult workplace scenario to navigate.¹³ Unlike the fields of business, law, and technology, for example, women in medicine do not routinely have training courses that incorporate skills such as navigating difficult conversations, conflict resolution, curriculum vitae writing, and negotiation. Various solutions have been offered to address some of the barriers facing women in medicine (such as the Association of American Medical Colleges and Executive Leadership in Academic Medicine leadership courses), but typically these focus on faculty rather than trainees. Given that women physicians practicing in the inpatient setting have been shown to have better patient outcomes¹⁴ and organizations with female leadership outperform those led by men,¹⁵ equipping our women trainees to thrive in the clinical and leadership environments is an essential step in fulfilling our mission as high-quality training programs.

At our institution, we recognized the need for training in leadership skills for women medical trainees and designed a day-long symposium for internal medicine women residents and fellows. Before developing the curriculum, we conducted a needs assessment to ascertain which skills women wanted to develop; women overwhelmingly wanted to learn about public speaking skills, work-life integration, and mentoring. Based on these responses, a group spanning multiple levels of training (residency, fellowship, and faculty) designed a combination of large-group lectures and small-group workshops termed “Women in Leadership Development” (WILD). The day-long curriculum included sessions on public speaking skills, women as change agents, mentorship, conflict resolution, and addressing microaggressions and concluded with a networking event for women faculty and trainees (Table).

Acknowledgments

The authors are indebted to Catherine Lucey, MD, for her framing of the issues for women in medicine as a complex problem and to Sarah Schaeffer, MD, for her unpublished data on microaggressions at our institution. The authors are also grateful to the UCSF Department of Medicine and the UCSF Chancellor’s Advisory Committee on the Status of Women for their financial support of the WILD (Women In Leadership Development) program.

Disclosures

The authors declare no conflict of interest.

Funding

The authors report no external funding source for this study.

A 69-year-old woman presented to the clinic with pain in the right great toe lasting several days. She was prescribed colchicine and indomethacin empirically for gout. She took one tablet of colchicine (0.6 mg) every hour until her stools became loose after the eighth tablet. Her toe pain resolved, but two days later she developed bilateral lower extremity pruritus and paresthesia and presented to the emergency department (ED). On physical examination, no rash, weakness, or sensory deficits were observed, and she was able to ambulate without assistance. Her patellar reflexes were normal. The complete blood count was notable for an absolute lymphocyte count of 6,120/µL (normal: 1,100-4,800), and the comprehensive metabolic panel was normal. Serum creatine kinase (CK) was 341 U/L (normal: 24-170) and uric acid 7.7 mg/dL (normal: 2.4-6.4). Her lower extremity symptoms were attributed to colchicine, which was discontinued. She was prescribed diphenhydramine and discharged home.

Monoarthritis of the hallux is the classic manifestation of gout, although other considerations include pseudogout, sesamoiditis, and trauma. The typical side effects of colchicine include diarrhea and myositis. Colchicine-induced muscle injury often results in a modest elevation of CK levels and is associated with myalgia.

Paresthesia is defined as abnormal sensory symptoms that most commonly localize to the peripheral nerves or spinal cord. Acute neuropathies or myelopathies might result from vasculitis, heavy metal toxicity, vitamin deficiencies, and paraneoplastic neurologic syndromes. The normal motor, sensory, and reflex examination, however, make these unlikely.

The neuro-anatomic localization of pruritus is poorly understood but is proposed to include peripheral nerves, spinothalamic tracts, and thalami. Acute pruritus (lasting <6 weeks) typically results from a primary dermatologic process such as a drug reaction, eczema, or xerosis. Less common causes include uremia, cholestasis, and thyroid disease. Pruritus can also be seen with malignancy, most commonly hematologic or paraneoplastic syndromes, or with connective tissue diseases. At this stage, it is unclear whether her pruritus and paresthesia are part of a unifying disease process.

Five days later she re-presented to the ED with nausea and emesis after eating at a restaurant. Her symptoms improved with intravenous fluids, and she was discharged. Four days later she returned with difficulty ambulating, bilateral leg cramping, and continued pruritus and paresthesia. The chemistry panel was normal except for a potassium level of 2.6 mmol/L and a bicarbonate level of 32 mmol/L. She was admitted to the hospital because of severe hypokalemia and impaired ability to ambulate. Her potassium was replenished. Her CK was elevated (3,551 U/L on hospital day 7). She was given cyclobenzaprine, gabapentin, oxycodone, acetaminophen, and prednisone (40 mg); her cramping only mildly improved, and she remained unable to walk. On hospital day five she had visual hallucinations and confusion, which did not resolve with administration of haloperidol; a head CT was unremarkable. On hospital day eight the patient, with her family’s support, left the hospital and presented to a different ED for a second opinion.

Difficulty ambulating often results from weakness, sensory impairment, cerebellar ataxia, extrapyramidal dysfunction (eg, parkinsonism), and pain. In this patient, leg cramping suggests pain or true weakness due to a myopathic process as a contributing factor. Symptoms of muscle disease include cramps, myalgia, and difficulty walking. Causes of elevated CK and myalgia include inflammatory myopathies, endocrinopathies, drugs, infections, and electrolyte abnormalities (eg, hypokalemia). Her age and acuity of presentation decrease the likelihood of a metabolic myopathy due to a disorder of glycogen storage, lipid metabolism, or mitochondrial function. Her hypokalemic metabolic alkalosis likely resulted from vomiting. Hypokalemic periodic paralysis is unlikely as exacerbations typically only last hours to days. As such, her difficulty ambulating, muscle cramps, and elevated CK strongly support a primary myopathic disorder, although additional information regarding the neurologic examination is still required.

Acute changes in mental status without corresponding changes in cranial nerve, motor, or sensory function are common in the hospital setting and frequently relate to delirium, which is the most likely explanation for her confusion. Her age and exposure to muscle relaxants, opiates, and corticosteroids increase her risk considerably. Other possible explanations for isolated changes in mental status include nonconvulsive seizures, central nervous system (CNS) infection, and strokes that involve the thalamus, nondominant parietal lobe, and reticular activating system. A shower of emboli resulting in small multifocal strokes can have the same effect.

She was re-evaluated by her new providers. Her only prior medical history was hypertension, which was treated at home with atenolol and amlodipine. She had emigrated from Nigeria to the US many years prior. She occasionally consumed alcohol and never smoked tobacco or used illicit drugs. She was unsure if she had received a tetanus booster in the past 10 years.

On physical examination, her temperature was 36°C, blood pressure 149/70 mm Hg, pulse 56 beats per minute, respiratory rate 18 breaths per minute, and oxygen saturation 98% on ambient air. She was diaphoretic and appeared anxious, grabbing both bedrails out of fear of falling. Cardiovascular, pulmonary, abdominal, and skin examinations were normal. She was alert and oriented to her identity, her location, and the time. Cranial nerves II to XII were normal. Tone was normal in her upper extremities but markedly increased in her lower extremities and back. There were spontaneous and stimulus-induced painful spasms, predominantly involving her axial muscles and distal lower extremities. Muscle bulk was normal. Strength was normal in the upper extremities and could not be assessed in the lower extremities due to rigidity. Reflexes were 2+ and symmetric throughout with downgoing toes on Babinski testing. A sensory examination was normal. Gait could not be tested because of the severe muscle spasms. The patient was admitted to the hospital.

Localized muscle spasms may be caused by muscle overuse, but more generalized spasms are associated with systemic diseases such as electrolyte disturbances, toxidromes, tetanus, peripheral nerve hyperexcitability syndromes (including Isaacs syndrome and Morvan syndrome), or stiff person syndrome (SPS). Hypokalemia is unlikely the cause as its correction did not improve her symptoms. Although tetanus is rare in the United States, it remains endemic in the developing world and can cause focal as well as generalized stimulus-induced spasms. The patient should be asked about potential exposure to Clostridium tetani infection, such as incurring a puncture wound. It is also important to consider neuroleptic malignant syndrome and serotonin syndrome, which can cause confusion, elevated CK, and increased muscle tone. Her confusion, however, was transient and the elevated CK preceded the administration of haloperidol.

SPS and progressive encephalomyelitis with rigidity and myoclonus (PERM) provide better explanations for her presentation. Both diseases cause severe spasms, impaired ambulation, and stiffness. They differ in their acuity of onset, accompanying symptoms, antibody associations, and responses to treatment. The rapid onset, paresthesia, and confusion seen in this patient are atypical of SPS. SPS usually presents with subacute-to-chronic stiffness or soreness of muscles in the back and lower extremities, followed by the upper extremities. Rigidity, stimulation-provoked spasms, hyperlordosis, and difficulty ambulating are typically later-stage findings. Her rapid escalation of symptoms is more consistent with PERM, which is often more acute and progressive than typical SPS; however, unlike this patient, PERM commonly causes widespread CNS dysfunction, including persistent encephalopathy, cranial neuropathies, hyperreflexia, and autonomic instability. Both are rare diagnoses that can manifest as a paraneoplastic neurologic syndrome.

Blood tests showed a leukocyte count of 17,350/µL, neutrophils 8,720/µL (normal: 1,500–7,800), lymphocytes 6,130/µL, hemoglobin 11.3 g/dL, and platelets 231,000/µL. The basic metabolic panel was normal. Serum total protein was 6.7 g/dL with albumin 3.5 g/dL. Aspartate aminotransferase (AST) was 94 U/L (normal: 0-31), alanine aminotransferase (ALT) 56 U/L (normal: 0-31), alkaline phosphatase 45 U/L, and total bilirubin 1.1 mg/dL. Vitamin B12 was 868 pg/mL. Hemoglobin A1c and thyrotropin levels were normal. Creatine kinase was 3,757 U/L and lactate dehydrogenase (LDH) 435 U/L (normal: 122-220). The syphilis treponemal test and hepatitis B surface antigen were negative. HIV and hepatitis C antibodies were nonreactive. The anti-nuclear antibody screen was negative and complement C3 and C4 were normal.

Neutrophilia likely reflects glucocorticoid-induced demargination, as opposed to an infectious process, given the temporal association with steroid administration. Persistent mild lymphocytosis is nonspecific but more likely to reflect a reactive rather than a clonal process. Elevated LDH and CK, as well as a greater increase of AST relative to ALT, suggest muscle injury, although mild concomitant hepatic injury cannot be excluded. Normal or negative serum studies for TSH, HIV, ANA, peripheral blood smear, and creatinine eliminate many of the systemic causes of her pruritus, but malignancy and associated paraneoplastic etiologies remain considerations.

The initial work-up for SPS includes electromyography (EMG) which would show spontaneous muscle activity. Her poorly localized sensory abnormalities, transient vestibular symptoms, and confusion warrant an MRI of the brain and spine to evaluate for inflammation (eg, encephalomyelitis), which could be consistent with PERM.

An MRI of the brain and cervicothoracic spine without contrast was significantly limited by motion artifact but without obvious intracranial or cord signal abnormalities. Electromyography demonstrated spontaneous muscle activity in both lower extremities with co-contraction of agonist and antagonist muscles (hamstrings and quadriceps as well as medial gastrocnemius and tibialis anterior). Sensory and motor nerve conductions were normal. Cerebral spinal fluid (CSF) contained six leukocytes (96% lymphocytes) and three red blood cells per microliter; glucose was 67 mg/dL and protein 24 mg/dL. There were two oligoclonal bands unique to the CSF. Cytology was negative for malignant cells.

The EMG narrows the differential diagnosis considerably. Co-contraction of opposing flexor and extensor groups (with predominance of extensors) on EMG is a diagnostic criterion for SPS and explains the myalgia and elevated CK. Her normal MRI studies effectively ruled out any focal lesion and did not show signs of encephalitis. Oligoclonal bands in the CSF are a sensitive marker of intrathecal inflammation, although not specific to one diagnosis. The mildly elevated cell count also supports CNS inflammation. In the setting of a lymphocytic pleocytosis and unique oligoclonal bands, it is important to consider infectious, neoplastic, autoimmune, and paraneoplastic causes of neuroinflammatory disorders.

Serum analyses, including antiglutamic acid decarboxylase 65 (GAD65) antibody and anti-amphiphysin antibody, should be ordered. The anti-GAD65 antibody is most commonly elevated in the setting of autoimmune diabetes mellitus; the titer, however, is usually dramatically higher in SPS. The CSF titer of anti-GAD65 antibodies is more specific than the serum titer for SPS. Antibodies against amphiphysin are typically elevated in paraneoplastic SPS, and anti-glycine receptor antibodies are associated with PERM, which commonly does not have elevated anti-GAD65 antibodies.

The serum GAD65 antibody level was greater than 265,000 × 10³ IU/µL (normal <5,000), and the CSF level was 11.2 nmol/L (normal: ≤0.02). Serum amphiphysin antibody testing was negative.

Significantly elevated serum and CSF anti-GAD65 antibody levels are highly suggestive of SPS. Stiff person syndrome with rapidly progressive clinical symptoms raises the concern of a paraneoplastic neurologic syndrome. Although anti-amphiphysin antibody – the antibody classically associated with breast cancer and SPS – was negative, anti-GAD65 antibody has been implicated in paraneoplastic SPS with thymoma, lymphoma, and thyroid carcinoma. Paraneoplastic neurologic syndrome can predate a detectable malignancy by several years. As SPS and lymphoma are associated with pruritus and lymphocytosis, imaging is indicated to search for malignancy. Antiglycine receptor antibody, associated with PERM, is not routinely available commercially.

Computed tomography of the chest, abdomen, and pelvis with intravenous contrast revealed a 3.9 × 8.0 × 7.0 cm anterior mediastinal mass (Figure 1, Panel A). Biopsy of the mass demonstrated a thymoma. Given that the patient exhibited no further signs of CNS involvement, her initial transiently altered mental status was attributed to opioids and steroids. As she did not meet the clinical criteria for PERM, testing of antiglycine antibodies was not pursued.

She received scheduled baclofen and diazepam with as needed cyclobenzaprine for continued muscle spasms. Over the next several days, her stiffness, spasms, and myoclonic jerks slowly improved, and she was able to attempt physical therapy (Appendix Video 1; https://youtu.be/d0gLpTgqaCs). She subsequently received intravenous immunoglobulin (IVIG) with further improvement. After five months of scheduled diazepam and baclofen, she was able to ambulate with minimal assistance (Appendix Video 2; https://youtu.be/I00i638u00o). Given the absence of safe tissue planes for resection, the patient received neoadjuvant chemotherapy with four cycles of cyclophosphamide, doxorubicin, and cisplatin. Tumor size decreased to 1.7 × 6.5 × 5.2 cm (Figure 1, Panel B), and she subsequently underwent resection (Figure 2). Pathological analysis demonstrated a type B1 thymoma.

SPS is a condition of muscle stiffness and spasticity. Diagnosis is difficult and often delayed due to its rarity, with an approximate prevalence of one to two cases per million people.¹ SPS typically occurs in middle age, and women are diagnosed twice as often as men. Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, high-titer antibody to GAD65 or amphiphysin, and the absence of an alternate diagnosis.² Variant syndromes have been described, including a milder variant limited to the limbs, a severe variant with brainstem and spinal cord involvement, and a paraneoplastic variant.³ This patient’s clinical presentation, EMG findings, and extraordinarily high anti-GAD titers in the serum and CSF were diagnostic of SPS.

The pathophysiology of SPS is associated with autoantibodies targeting proteins such as GAD65, amphiphysin, gephyrin, and GABA_A receptor-associated protein (GABARAP). These proteins are critical to gamma-aminobutyric acid (GABA) signaling, the primary inhibitory neurotransmitter pathway in the CNS (Figure 3).⁴ The formation of GABA from glutamate is catalyzed by GAD65. Gamma-aminobutyric acid is loaded into secretory vesicles, and amphiphysin facilitates vesicle recycling from the synaptic space.⁵ In the postsynaptic neuron, GABA binds the GABA_A receptor, leading to neuronal hyperpolarization and resistance to excitation. The GABA_A receptor is clustered on the plasma membrane through a scaffold formed by gephyrin. GABARAP facilitates this clustering, in part by linking GABA_A receptors and gephyrin.⁶ Autoantibodies to these proteins may be pathogenic; however, the direct effects on their targets are unclear. The end result is decreased GABAergic activity, leading to continuous activation of opposing muscle groups. The resulting stiffness is characteristic of this disorder. Colchicine is known to antagonize GABA_A receptor signaling, and this may have brought the underlying diagnosis of SPS to clinical attention.^7,8

• Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, and high-titer antibody to GAD65 or amphiphysin.
• SPS typically occurs in middle age, and women are diagnosed twice as often as men.
• Symptomatic treatment of SPS targets the GABAergic system. Typically, high doses of scheduled benzodiazepines and baclofenare necessary.
• SPS occasionally occurs as a paraneoplastic neurologic syndrome, most commonly in association with breast cancer.

Acknowledgments

The authors wish to thank Jason Kern, MD for his preparation and interpretation of the pathologic image; and the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education, for supporting Reza Manesh, MD.

Disclosures

The authors have nothing to disclose.

Appendix Video 1: This video was taken during a physical therapy session after 1 week of scheduled benzodiazepine and 2 days of intravenous immunoglobulin. It was difficult for the patient to stand without assistance due to severe stiffness. (https://youtu.be/d0gLpTgqaCs)

Appendix Video 2: This video was taken 5 months after scheduled diazepam and baclofen, and 1 week prior to thymectomy. (https://youtu.be/I00i638u00o)

A 69-year-old woman presented to the clinic with pain in the right great toe lasting several days. She was prescribed colchicine and indomethacin empirically for gout. She took one tablet of colchicine (0.6 mg) every hour until her stools became loose after the eighth tablet. Her toe pain resolved, but two days later she developed bilateral lower extremity pruritus and paresthesia and presented to the emergency department (ED). On physical examination, no rash, weakness, or sensory deficits were observed, and she was able to ambulate without assistance. Her patellar reflexes were normal. The complete blood count was notable for an absolute lymphocyte count of 6,120/µL (normal: 1,100-4,800), and the comprehensive metabolic panel was normal. Serum creatine kinase (CK) was 341 U/L (normal: 24-170) and uric acid 7.7 mg/dL (normal: 2.4-6.4). Her lower extremity symptoms were attributed to colchicine, which was discontinued. She was prescribed diphenhydramine and discharged home.

Monoarthritis of the hallux is the classic manifestation of gout, although other considerations include pseudogout, sesamoiditis, and trauma. The typical side effects of colchicine include diarrhea and myositis. Colchicine-induced muscle injury often results in a modest elevation of CK levels and is associated with myalgia.

Paresthesia is defined as abnormal sensory symptoms that most commonly localize to the peripheral nerves or spinal cord. Acute neuropathies or myelopathies might result from vasculitis, heavy metal toxicity, vitamin deficiencies, and paraneoplastic neurologic syndromes. The normal motor, sensory, and reflex examination, however, make these unlikely.

The neuro-anatomic localization of pruritus is poorly understood but is proposed to include peripheral nerves, spinothalamic tracts, and thalami. Acute pruritus (lasting <6 weeks) typically results from a primary dermatologic process such as a drug reaction, eczema, or xerosis. Less common causes include uremia, cholestasis, and thyroid disease. Pruritus can also be seen with malignancy, most commonly hematologic or paraneoplastic syndromes, or with connective tissue diseases. At this stage, it is unclear whether her pruritus and paresthesia are part of a unifying disease process.

Five days later she re-presented to the ED with nausea and emesis after eating at a restaurant. Her symptoms improved with intravenous fluids, and she was discharged. Four days later she returned with difficulty ambulating, bilateral leg cramping, and continued pruritus and paresthesia. The chemistry panel was normal except for a potassium level of 2.6 mmol/L and a bicarbonate level of 32 mmol/L. She was admitted to the hospital because of severe hypokalemia and impaired ability to ambulate. Her potassium was replenished. Her CK was elevated (3,551 U/L on hospital day 7). She was given cyclobenzaprine, gabapentin, oxycodone, acetaminophen, and prednisone (40 mg); her cramping only mildly improved, and she remained unable to walk. On hospital day five she had visual hallucinations and confusion, which did not resolve with administration of haloperidol; a head CT was unremarkable. On hospital day eight the patient, with her family’s support, left the hospital and presented to a different ED for a second opinion.

Difficulty ambulating often results from weakness, sensory impairment, cerebellar ataxia, extrapyramidal dysfunction (eg, parkinsonism), and pain. In this patient, leg cramping suggests pain or true weakness due to a myopathic process as a contributing factor. Symptoms of muscle disease include cramps, myalgia, and difficulty walking. Causes of elevated CK and myalgia include inflammatory myopathies, endocrinopathies, drugs, infections, and electrolyte abnormalities (eg, hypokalemia). Her age and acuity of presentation decrease the likelihood of a metabolic myopathy due to a disorder of glycogen storage, lipid metabolism, or mitochondrial function. Her hypokalemic metabolic alkalosis likely resulted from vomiting. Hypokalemic periodic paralysis is unlikely as exacerbations typically only last hours to days. As such, her difficulty ambulating, muscle cramps, and elevated CK strongly support a primary myopathic disorder, although additional information regarding the neurologic examination is still required.

Acute changes in mental status without corresponding changes in cranial nerve, motor, or sensory function are common in the hospital setting and frequently relate to delirium, which is the most likely explanation for her confusion. Her age and exposure to muscle relaxants, opiates, and corticosteroids increase her risk considerably. Other possible explanations for isolated changes in mental status include nonconvulsive seizures, central nervous system (CNS) infection, and strokes that involve the thalamus, nondominant parietal lobe, and reticular activating system. A shower of emboli resulting in small multifocal strokes can have the same effect.

She was re-evaluated by her new providers. Her only prior medical history was hypertension, which was treated at home with atenolol and amlodipine. She had emigrated from Nigeria to the US many years prior. She occasionally consumed alcohol and never smoked tobacco or used illicit drugs. She was unsure if she had received a tetanus booster in the past 10 years.

On physical examination, her temperature was 36°C, blood pressure 149/70 mm Hg, pulse 56 beats per minute, respiratory rate 18 breaths per minute, and oxygen saturation 98% on ambient air. She was diaphoretic and appeared anxious, grabbing both bedrails out of fear of falling. Cardiovascular, pulmonary, abdominal, and skin examinations were normal. She was alert and oriented to her identity, her location, and the time. Cranial nerves II to XII were normal. Tone was normal in her upper extremities but markedly increased in her lower extremities and back. There were spontaneous and stimulus-induced painful spasms, predominantly involving her axial muscles and distal lower extremities. Muscle bulk was normal. Strength was normal in the upper extremities and could not be assessed in the lower extremities due to rigidity. Reflexes were 2+ and symmetric throughout with downgoing toes on Babinski testing. A sensory examination was normal. Gait could not be tested because of the severe muscle spasms. The patient was admitted to the hospital.

Localized muscle spasms may be caused by muscle overuse, but more generalized spasms are associated with systemic diseases such as electrolyte disturbances, toxidromes, tetanus, peripheral nerve hyperexcitability syndromes (including Isaacs syndrome and Morvan syndrome), or stiff person syndrome (SPS). Hypokalemia is unlikely the cause as its correction did not improve her symptoms. Although tetanus is rare in the United States, it remains endemic in the developing world and can cause focal as well as generalized stimulus-induced spasms. The patient should be asked about potential exposure to Clostridium tetani infection, such as incurring a puncture wound. It is also important to consider neuroleptic malignant syndrome and serotonin syndrome, which can cause confusion, elevated CK, and increased muscle tone. Her confusion, however, was transient and the elevated CK preceded the administration of haloperidol.

SPS and progressive encephalomyelitis with rigidity and myoclonus (PERM) provide better explanations for her presentation. Both diseases cause severe spasms, impaired ambulation, and stiffness. They differ in their acuity of onset, accompanying symptoms, antibody associations, and responses to treatment. The rapid onset, paresthesia, and confusion seen in this patient are atypical of SPS. SPS usually presents with subacute-to-chronic stiffness or soreness of muscles in the back and lower extremities, followed by the upper extremities. Rigidity, stimulation-provoked spasms, hyperlordosis, and difficulty ambulating are typically later-stage findings. Her rapid escalation of symptoms is more consistent with PERM, which is often more acute and progressive than typical SPS; however, unlike this patient, PERM commonly causes widespread CNS dysfunction, including persistent encephalopathy, cranial neuropathies, hyperreflexia, and autonomic instability. Both are rare diagnoses that can manifest as a paraneoplastic neurologic syndrome.

Blood tests showed a leukocyte count of 17,350/µL, neutrophils 8,720/µL (normal: 1,500–7,800), lymphocytes 6,130/µL, hemoglobin 11.3 g/dL, and platelets 231,000/µL. The basic metabolic panel was normal. Serum total protein was 6.7 g/dL with albumin 3.5 g/dL. Aspartate aminotransferase (AST) was 94 U/L (normal: 0-31), alanine aminotransferase (ALT) 56 U/L (normal: 0-31), alkaline phosphatase 45 U/L, and total bilirubin 1.1 mg/dL. Vitamin B12 was 868 pg/mL. Hemoglobin A1c and thyrotropin levels were normal. Creatine kinase was 3,757 U/L and lactate dehydrogenase (LDH) 435 U/L (normal: 122-220). The syphilis treponemal test and hepatitis B surface antigen were negative. HIV and hepatitis C antibodies were nonreactive. The anti-nuclear antibody screen was negative and complement C3 and C4 were normal.

Neutrophilia likely reflects glucocorticoid-induced demargination, as opposed to an infectious process, given the temporal association with steroid administration. Persistent mild lymphocytosis is nonspecific but more likely to reflect a reactive rather than a clonal process. Elevated LDH and CK, as well as a greater increase of AST relative to ALT, suggest muscle injury, although mild concomitant hepatic injury cannot be excluded. Normal or negative serum studies for TSH, HIV, ANA, peripheral blood smear, and creatinine eliminate many of the systemic causes of her pruritus, but malignancy and associated paraneoplastic etiologies remain considerations.

The initial work-up for SPS includes electromyography (EMG) which would show spontaneous muscle activity. Her poorly localized sensory abnormalities, transient vestibular symptoms, and confusion warrant an MRI of the brain and spine to evaluate for inflammation (eg, encephalomyelitis), which could be consistent with PERM.

An MRI of the brain and cervicothoracic spine without contrast was significantly limited by motion artifact but without obvious intracranial or cord signal abnormalities. Electromyography demonstrated spontaneous muscle activity in both lower extremities with co-contraction of agonist and antagonist muscles (hamstrings and quadriceps as well as medial gastrocnemius and tibialis anterior). Sensory and motor nerve conductions were normal. Cerebral spinal fluid (CSF) contained six leukocytes (96% lymphocytes) and three red blood cells per microliter; glucose was 67 mg/dL and protein 24 mg/dL. There were two oligoclonal bands unique to the CSF. Cytology was negative for malignant cells.

The EMG narrows the differential diagnosis considerably. Co-contraction of opposing flexor and extensor groups (with predominance of extensors) on EMG is a diagnostic criterion for SPS and explains the myalgia and elevated CK. Her normal MRI studies effectively ruled out any focal lesion and did not show signs of encephalitis. Oligoclonal bands in the CSF are a sensitive marker of intrathecal inflammation, although not specific to one diagnosis. The mildly elevated cell count also supports CNS inflammation. In the setting of a lymphocytic pleocytosis and unique oligoclonal bands, it is important to consider infectious, neoplastic, autoimmune, and paraneoplastic causes of neuroinflammatory disorders.

Serum analyses, including antiglutamic acid decarboxylase 65 (GAD65) antibody and anti-amphiphysin antibody, should be ordered. The anti-GAD65 antibody is most commonly elevated in the setting of autoimmune diabetes mellitus; the titer, however, is usually dramatically higher in SPS. The CSF titer of anti-GAD65 antibodies is more specific than the serum titer for SPS. Antibodies against amphiphysin are typically elevated in paraneoplastic SPS, and anti-glycine receptor antibodies are associated with PERM, which commonly does not have elevated anti-GAD65 antibodies.

The serum GAD65 antibody level was greater than 265,000 × 10³ IU/µL (normal <5,000), and the CSF level was 11.2 nmol/L (normal: ≤0.02). Serum amphiphysin antibody testing was negative.

Significantly elevated serum and CSF anti-GAD65 antibody levels are highly suggestive of SPS. Stiff person syndrome with rapidly progressive clinical symptoms raises the concern of a paraneoplastic neurologic syndrome. Although anti-amphiphysin antibody – the antibody classically associated with breast cancer and SPS – was negative, anti-GAD65 antibody has been implicated in paraneoplastic SPS with thymoma, lymphoma, and thyroid carcinoma. Paraneoplastic neurologic syndrome can predate a detectable malignancy by several years. As SPS and lymphoma are associated with pruritus and lymphocytosis, imaging is indicated to search for malignancy. Antiglycine receptor antibody, associated with PERM, is not routinely available commercially.

Computed tomography of the chest, abdomen, and pelvis with intravenous contrast revealed a 3.9 × 8.0 × 7.0 cm anterior mediastinal mass (Figure 1, Panel A). Biopsy of the mass demonstrated a thymoma. Given that the patient exhibited no further signs of CNS involvement, her initial transiently altered mental status was attributed to opioids and steroids. As she did not meet the clinical criteria for PERM, testing of antiglycine antibodies was not pursued.

She received scheduled baclofen and diazepam with as needed cyclobenzaprine for continued muscle spasms. Over the next several days, her stiffness, spasms, and myoclonic jerks slowly improved, and she was able to attempt physical therapy (Appendix Video 1; https://youtu.be/d0gLpTgqaCs). She subsequently received intravenous immunoglobulin (IVIG) with further improvement. After five months of scheduled diazepam and baclofen, she was able to ambulate with minimal assistance (Appendix Video 2; https://youtu.be/I00i638u00o). Given the absence of safe tissue planes for resection, the patient received neoadjuvant chemotherapy with four cycles of cyclophosphamide, doxorubicin, and cisplatin. Tumor size decreased to 1.7 × 6.5 × 5.2 cm (Figure 1, Panel B), and she subsequently underwent resection (Figure 2). Pathological analysis demonstrated a type B1 thymoma.

SPS is a condition of muscle stiffness and spasticity. Diagnosis is difficult and often delayed due to its rarity, with an approximate prevalence of one to two cases per million people.¹ SPS typically occurs in middle age, and women are diagnosed twice as often as men. Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, high-titer antibody to GAD65 or amphiphysin, and the absence of an alternate diagnosis.² Variant syndromes have been described, including a milder variant limited to the limbs, a severe variant with brainstem and spinal cord involvement, and a paraneoplastic variant.³ This patient’s clinical presentation, EMG findings, and extraordinarily high anti-GAD titers in the serum and CSF were diagnostic of SPS.

The pathophysiology of SPS is associated with autoantibodies targeting proteins such as GAD65, amphiphysin, gephyrin, and GABA_A receptor-associated protein (GABARAP). These proteins are critical to gamma-aminobutyric acid (GABA) signaling, the primary inhibitory neurotransmitter pathway in the CNS (Figure 3).⁴ The formation of GABA from glutamate is catalyzed by GAD65. Gamma-aminobutyric acid is loaded into secretory vesicles, and amphiphysin facilitates vesicle recycling from the synaptic space.⁵ In the postsynaptic neuron, GABA binds the GABA_A receptor, leading to neuronal hyperpolarization and resistance to excitation. The GABA_A receptor is clustered on the plasma membrane through a scaffold formed by gephyrin. GABARAP facilitates this clustering, in part by linking GABA_A receptors and gephyrin.⁶ Autoantibodies to these proteins may be pathogenic; however, the direct effects on their targets are unclear. The end result is decreased GABAergic activity, leading to continuous activation of opposing muscle groups. The resulting stiffness is characteristic of this disorder. Colchicine is known to antagonize GABA_A receptor signaling, and this may have brought the underlying diagnosis of SPS to clinical attention.^7,8

• Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, and high-titer antibody to GAD65 or amphiphysin.
• SPS typically occurs in middle age, and women are diagnosed twice as often as men.
• Symptomatic treatment of SPS targets the GABAergic system. Typically, high doses of scheduled benzodiazepines and baclofenare necessary.
• SPS occasionally occurs as a paraneoplastic neurologic syndrome, most commonly in association with breast cancer.

Acknowledgments

The authors wish to thank Jason Kern, MD for his preparation and interpretation of the pathologic image; and the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education, for supporting Reza Manesh, MD.

Disclosures

The authors have nothing to disclose.

Appendix Video 1: This video was taken during a physical therapy session after 1 week of scheduled benzodiazepine and 2 days of intravenous immunoglobulin. It was difficult for the patient to stand without assistance due to severe stiffness. (https://youtu.be/d0gLpTgqaCs)

Appendix Video 2: This video was taken 5 months after scheduled diazepam and baclofen, and 1 week prior to thymectomy. (https://youtu.be/I00i638u00o)

A 69-year-old woman presented to the clinic with pain in the right great toe lasting several days. She was prescribed colchicine and indomethacin empirically for gout. She took one tablet of colchicine (0.6 mg) every hour until her stools became loose after the eighth tablet. Her toe pain resolved, but two days later she developed bilateral lower extremity pruritus and paresthesia and presented to the emergency department (ED). On physical examination, no rash, weakness, or sensory deficits were observed, and she was able to ambulate without assistance. Her patellar reflexes were normal. The complete blood count was notable for an absolute lymphocyte count of 6,120/µL (normal: 1,100-4,800), and the comprehensive metabolic panel was normal. Serum creatine kinase (CK) was 341 U/L (normal: 24-170) and uric acid 7.7 mg/dL (normal: 2.4-6.4). Her lower extremity symptoms were attributed to colchicine, which was discontinued. She was prescribed diphenhydramine and discharged home.

Monoarthritis of the hallux is the classic manifestation of gout, although other considerations include pseudogout, sesamoiditis, and trauma. The typical side effects of colchicine include diarrhea and myositis. Colchicine-induced muscle injury often results in a modest elevation of CK levels and is associated with myalgia.

Paresthesia is defined as abnormal sensory symptoms that most commonly localize to the peripheral nerves or spinal cord. Acute neuropathies or myelopathies might result from vasculitis, heavy metal toxicity, vitamin deficiencies, and paraneoplastic neurologic syndromes. The normal motor, sensory, and reflex examination, however, make these unlikely.

The neuro-anatomic localization of pruritus is poorly understood but is proposed to include peripheral nerves, spinothalamic tracts, and thalami. Acute pruritus (lasting <6 weeks) typically results from a primary dermatologic process such as a drug reaction, eczema, or xerosis. Less common causes include uremia, cholestasis, and thyroid disease. Pruritus can also be seen with malignancy, most commonly hematologic or paraneoplastic syndromes, or with connective tissue diseases. At this stage, it is unclear whether her pruritus and paresthesia are part of a unifying disease process.

Five days later she re-presented to the ED with nausea and emesis after eating at a restaurant. Her symptoms improved with intravenous fluids, and she was discharged. Four days later she returned with difficulty ambulating, bilateral leg cramping, and continued pruritus and paresthesia. The chemistry panel was normal except for a potassium level of 2.6 mmol/L and a bicarbonate level of 32 mmol/L. She was admitted to the hospital because of severe hypokalemia and impaired ability to ambulate. Her potassium was replenished. Her CK was elevated (3,551 U/L on hospital day 7). She was given cyclobenzaprine, gabapentin, oxycodone, acetaminophen, and prednisone (40 mg); her cramping only mildly improved, and she remained unable to walk. On hospital day five she had visual hallucinations and confusion, which did not resolve with administration of haloperidol; a head CT was unremarkable. On hospital day eight the patient, with her family’s support, left the hospital and presented to a different ED for a second opinion.

Difficulty ambulating often results from weakness, sensory impairment, cerebellar ataxia, extrapyramidal dysfunction (eg, parkinsonism), and pain. In this patient, leg cramping suggests pain or true weakness due to a myopathic process as a contributing factor. Symptoms of muscle disease include cramps, myalgia, and difficulty walking. Causes of elevated CK and myalgia include inflammatory myopathies, endocrinopathies, drugs, infections, and electrolyte abnormalities (eg, hypokalemia). Her age and acuity of presentation decrease the likelihood of a metabolic myopathy due to a disorder of glycogen storage, lipid metabolism, or mitochondrial function. Her hypokalemic metabolic alkalosis likely resulted from vomiting. Hypokalemic periodic paralysis is unlikely as exacerbations typically only last hours to days. As such, her difficulty ambulating, muscle cramps, and elevated CK strongly support a primary myopathic disorder, although additional information regarding the neurologic examination is still required.

Acute changes in mental status without corresponding changes in cranial nerve, motor, or sensory function are common in the hospital setting and frequently relate to delirium, which is the most likely explanation for her confusion. Her age and exposure to muscle relaxants, opiates, and corticosteroids increase her risk considerably. Other possible explanations for isolated changes in mental status include nonconvulsive seizures, central nervous system (CNS) infection, and strokes that involve the thalamus, nondominant parietal lobe, and reticular activating system. A shower of emboli resulting in small multifocal strokes can have the same effect.

She was re-evaluated by her new providers. Her only prior medical history was hypertension, which was treated at home with atenolol and amlodipine. She had emigrated from Nigeria to the US many years prior. She occasionally consumed alcohol and never smoked tobacco or used illicit drugs. She was unsure if she had received a tetanus booster in the past 10 years.

On physical examination, her temperature was 36°C, blood pressure 149/70 mm Hg, pulse 56 beats per minute, respiratory rate 18 breaths per minute, and oxygen saturation 98% on ambient air. She was diaphoretic and appeared anxious, grabbing both bedrails out of fear of falling. Cardiovascular, pulmonary, abdominal, and skin examinations were normal. She was alert and oriented to her identity, her location, and the time. Cranial nerves II to XII were normal. Tone was normal in her upper extremities but markedly increased in her lower extremities and back. There were spontaneous and stimulus-induced painful spasms, predominantly involving her axial muscles and distal lower extremities. Muscle bulk was normal. Strength was normal in the upper extremities and could not be assessed in the lower extremities due to rigidity. Reflexes were 2+ and symmetric throughout with downgoing toes on Babinski testing. A sensory examination was normal. Gait could not be tested because of the severe muscle spasms. The patient was admitted to the hospital.

Localized muscle spasms may be caused by muscle overuse, but more generalized spasms are associated with systemic diseases such as electrolyte disturbances, toxidromes, tetanus, peripheral nerve hyperexcitability syndromes (including Isaacs syndrome and Morvan syndrome), or stiff person syndrome (SPS). Hypokalemia is unlikely the cause as its correction did not improve her symptoms. Although tetanus is rare in the United States, it remains endemic in the developing world and can cause focal as well as generalized stimulus-induced spasms. The patient should be asked about potential exposure to Clostridium tetani infection, such as incurring a puncture wound. It is also important to consider neuroleptic malignant syndrome and serotonin syndrome, which can cause confusion, elevated CK, and increased muscle tone. Her confusion, however, was transient and the elevated CK preceded the administration of haloperidol.

SPS and progressive encephalomyelitis with rigidity and myoclonus (PERM) provide better explanations for her presentation. Both diseases cause severe spasms, impaired ambulation, and stiffness. They differ in their acuity of onset, accompanying symptoms, antibody associations, and responses to treatment. The rapid onset, paresthesia, and confusion seen in this patient are atypical of SPS. SPS usually presents with subacute-to-chronic stiffness or soreness of muscles in the back and lower extremities, followed by the upper extremities. Rigidity, stimulation-provoked spasms, hyperlordosis, and difficulty ambulating are typically later-stage findings. Her rapid escalation of symptoms is more consistent with PERM, which is often more acute and progressive than typical SPS; however, unlike this patient, PERM commonly causes widespread CNS dysfunction, including persistent encephalopathy, cranial neuropathies, hyperreflexia, and autonomic instability. Both are rare diagnoses that can manifest as a paraneoplastic neurologic syndrome.

Blood tests showed a leukocyte count of 17,350/µL, neutrophils 8,720/µL (normal: 1,500–7,800), lymphocytes 6,130/µL, hemoglobin 11.3 g/dL, and platelets 231,000/µL. The basic metabolic panel was normal. Serum total protein was 6.7 g/dL with albumin 3.5 g/dL. Aspartate aminotransferase (AST) was 94 U/L (normal: 0-31), alanine aminotransferase (ALT) 56 U/L (normal: 0-31), alkaline phosphatase 45 U/L, and total bilirubin 1.1 mg/dL. Vitamin B12 was 868 pg/mL. Hemoglobin A1c and thyrotropin levels were normal. Creatine kinase was 3,757 U/L and lactate dehydrogenase (LDH) 435 U/L (normal: 122-220). The syphilis treponemal test and hepatitis B surface antigen were negative. HIV and hepatitis C antibodies were nonreactive. The anti-nuclear antibody screen was negative and complement C3 and C4 were normal.

Neutrophilia likely reflects glucocorticoid-induced demargination, as opposed to an infectious process, given the temporal association with steroid administration. Persistent mild lymphocytosis is nonspecific but more likely to reflect a reactive rather than a clonal process. Elevated LDH and CK, as well as a greater increase of AST relative to ALT, suggest muscle injury, although mild concomitant hepatic injury cannot be excluded. Normal or negative serum studies for TSH, HIV, ANA, peripheral blood smear, and creatinine eliminate many of the systemic causes of her pruritus, but malignancy and associated paraneoplastic etiologies remain considerations.

The initial work-up for SPS includes electromyography (EMG) which would show spontaneous muscle activity. Her poorly localized sensory abnormalities, transient vestibular symptoms, and confusion warrant an MRI of the brain and spine to evaluate for inflammation (eg, encephalomyelitis), which could be consistent with PERM.

An MRI of the brain and cervicothoracic spine without contrast was significantly limited by motion artifact but without obvious intracranial or cord signal abnormalities. Electromyography demonstrated spontaneous muscle activity in both lower extremities with co-contraction of agonist and antagonist muscles (hamstrings and quadriceps as well as medial gastrocnemius and tibialis anterior). Sensory and motor nerve conductions were normal. Cerebral spinal fluid (CSF) contained six leukocytes (96% lymphocytes) and three red blood cells per microliter; glucose was 67 mg/dL and protein 24 mg/dL. There were two oligoclonal bands unique to the CSF. Cytology was negative for malignant cells.

The EMG narrows the differential diagnosis considerably. Co-contraction of opposing flexor and extensor groups (with predominance of extensors) on EMG is a diagnostic criterion for SPS and explains the myalgia and elevated CK. Her normal MRI studies effectively ruled out any focal lesion and did not show signs of encephalitis. Oligoclonal bands in the CSF are a sensitive marker of intrathecal inflammation, although not specific to one diagnosis. The mildly elevated cell count also supports CNS inflammation. In the setting of a lymphocytic pleocytosis and unique oligoclonal bands, it is important to consider infectious, neoplastic, autoimmune, and paraneoplastic causes of neuroinflammatory disorders.

Serum analyses, including antiglutamic acid decarboxylase 65 (GAD65) antibody and anti-amphiphysin antibody, should be ordered. The anti-GAD65 antibody is most commonly elevated in the setting of autoimmune diabetes mellitus; the titer, however, is usually dramatically higher in SPS. The CSF titer of anti-GAD65 antibodies is more specific than the serum titer for SPS. Antibodies against amphiphysin are typically elevated in paraneoplastic SPS, and anti-glycine receptor antibodies are associated with PERM, which commonly does not have elevated anti-GAD65 antibodies.

The serum GAD65 antibody level was greater than 265,000 × 10³ IU/µL (normal <5,000), and the CSF level was 11.2 nmol/L (normal: ≤0.02). Serum amphiphysin antibody testing was negative.

Significantly elevated serum and CSF anti-GAD65 antibody levels are highly suggestive of SPS. Stiff person syndrome with rapidly progressive clinical symptoms raises the concern of a paraneoplastic neurologic syndrome. Although anti-amphiphysin antibody – the antibody classically associated with breast cancer and SPS – was negative, anti-GAD65 antibody has been implicated in paraneoplastic SPS with thymoma, lymphoma, and thyroid carcinoma. Paraneoplastic neurologic syndrome can predate a detectable malignancy by several years. As SPS and lymphoma are associated with pruritus and lymphocytosis, imaging is indicated to search for malignancy. Antiglycine receptor antibody, associated with PERM, is not routinely available commercially.

Computed tomography of the chest, abdomen, and pelvis with intravenous contrast revealed a 3.9 × 8.0 × 7.0 cm anterior mediastinal mass (Figure 1, Panel A). Biopsy of the mass demonstrated a thymoma. Given that the patient exhibited no further signs of CNS involvement, her initial transiently altered mental status was attributed to opioids and steroids. As she did not meet the clinical criteria for PERM, testing of antiglycine antibodies was not pursued.

She received scheduled baclofen and diazepam with as needed cyclobenzaprine for continued muscle spasms. Over the next several days, her stiffness, spasms, and myoclonic jerks slowly improved, and she was able to attempt physical therapy (Appendix Video 1; https://youtu.be/d0gLpTgqaCs). She subsequently received intravenous immunoglobulin (IVIG) with further improvement. After five months of scheduled diazepam and baclofen, she was able to ambulate with minimal assistance (Appendix Video 2; https://youtu.be/I00i638u00o). Given the absence of safe tissue planes for resection, the patient received neoadjuvant chemotherapy with four cycles of cyclophosphamide, doxorubicin, and cisplatin. Tumor size decreased to 1.7 × 6.5 × 5.2 cm (Figure 1, Panel B), and she subsequently underwent resection (Figure 2). Pathological analysis demonstrated a type B1 thymoma.

SPS is a condition of muscle stiffness and spasticity. Diagnosis is difficult and often delayed due to its rarity, with an approximate prevalence of one to two cases per million people.¹ SPS typically occurs in middle age, and women are diagnosed twice as often as men. Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, high-titer antibody to GAD65 or amphiphysin, and the absence of an alternate diagnosis.² Variant syndromes have been described, including a milder variant limited to the limbs, a severe variant with brainstem and spinal cord involvement, and a paraneoplastic variant.³ This patient’s clinical presentation, EMG findings, and extraordinarily high anti-GAD titers in the serum and CSF were diagnostic of SPS.

The pathophysiology of SPS is associated with autoantibodies targeting proteins such as GAD65, amphiphysin, gephyrin, and GABA_A receptor-associated protein (GABARAP). These proteins are critical to gamma-aminobutyric acid (GABA) signaling, the primary inhibitory neurotransmitter pathway in the CNS (Figure 3).⁴ The formation of GABA from glutamate is catalyzed by GAD65. Gamma-aminobutyric acid is loaded into secretory vesicles, and amphiphysin facilitates vesicle recycling from the synaptic space.⁵ In the postsynaptic neuron, GABA binds the GABA_A receptor, leading to neuronal hyperpolarization and resistance to excitation. The GABA_A receptor is clustered on the plasma membrane through a scaffold formed by gephyrin. GABARAP facilitates this clustering, in part by linking GABA_A receptors and gephyrin.⁶ Autoantibodies to these proteins may be pathogenic; however, the direct effects on their targets are unclear. The end result is decreased GABAergic activity, leading to continuous activation of opposing muscle groups. The resulting stiffness is characteristic of this disorder. Colchicine is known to antagonize GABA_A receptor signaling, and this may have brought the underlying diagnosis of SPS to clinical attention.^7,8

• Classic SPS is characterized by axial and limb muscle stiffness, episodic spasms precipitated by tactile or auditory stimuli, continuous motor unit activity in agonist and antagonist muscles on EMG, and high-titer antibody to GAD65 or amphiphysin.
• SPS typically occurs in middle age, and women are diagnosed twice as often as men.
• Symptomatic treatment of SPS targets the GABAergic system. Typically, high doses of scheduled benzodiazepines and baclofenare necessary.
• SPS occasionally occurs as a paraneoplastic neurologic syndrome, most commonly in association with breast cancer.

Acknowledgments

The authors wish to thank Jason Kern, MD for his preparation and interpretation of the pathologic image; and the Jeremiah A. Barondess Fellowship in the Clinical Transaction of the New York Academy of Medicine, in collaboration with the Accreditation Council for Graduate Medical Education, for supporting Reza Manesh, MD.

Disclosures

The authors have nothing to disclose.

Appendix Video 1: This video was taken during a physical therapy session after 1 week of scheduled benzodiazepine and 2 days of intravenous immunoglobulin. It was difficult for the patient to stand without assistance due to severe stiffness. (https://youtu.be/d0gLpTgqaCs)

Appendix Video 2: This video was taken 5 months after scheduled diazepam and baclofen, and 1 week prior to thymectomy. (https://youtu.be/I00i638u00o)

Contact precautions (CP), the use of gowns and gloves as personal protective equipment when caring for patients who are colonized or infected with one or more multidrug-resistant organisms (MDROs), is an important infection prevention intervention utilized to prevent pathogens from being transmitted among patients in healthcare settings. Recently, certain healthcare facilities have taken steps to limit the use of CP for patients colonized or infected with MDROs that are considered to be endemic, namely methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococci (VRE). In this issue of the Journal of Hospital Medicine, authors Young et al. argue that CP for MRSA and VRE is an intervention that should be eliminated as part of the Choosing Wisely^® campaign because it is a “thing we do for no reason.”¹ We respectfully disagree with this characterization of CP for MRSA and VRE, and we assert instead that CP are a necessary practice that should be continued.

Young et al. refer to published studies and a recent meta-analysis that did not conclusively show a benefit of CP for MRSA and VRE.² The quasi-experimental studies cited have major methodological flaws that limit their ability to demonstrate the effect of CP. Most importantly, these studies fail to account for the fact that among patients who develop an infection following hospital-acquired MRSA colonization, approximately 70% of the infections are identified after discharge.³ When such studies do not restrict their outcome measure to include only those infections occurring among patients with hospital-acquired colonization, and do not take steps to accurately identify postdischarge infections that occur in such patients, their results are biased toward the null and difficult to interpret. Due to several serious challenges to study feasibility, including the need for an extremely large sample size, a very long period of follow-up, and the need to control for a variety of other concurrent infection prevention measures, there may never be a study that conclusively proves that CP, apart from other infection prevention interventions, has a significant impact. However, despite these limitations, one of the recent multicenter randomized controlled trials, cited by the authors as evidence against the use of CP, was able to demonstrate a significant reduction in MRSA transmission using universal gowns and gloves for all intensive care unit patients, even in sites that utilized other effective strategies, including chlorhexidine bathing.^4,5

In this issue of the Journal of Hospital Medicine^®, Young et al. acknowledge that CP are generally utilized as part of a comprehensive package of infection prevention approaches that also includes hand hygiene, environmental cleaning, antimicrobial stewardship, and evidence-based interventions to prevent device- and procedure-related infections. This multifaceted approach makes it more difficult to determine the attributable effect of CP alone. However, there is a strong rationale for using CP to prevent transmission, and there are numerous examples where the use of bundled approaches that include CP was associated with success. In the Netherlands, CP were part of an aggressive “search and destroy” approach to MRSA associated with almost total elimination of MRSA from hospitals in that country. The United Kingdom achieved an 80% decrease in MRSA bacteremia following a series of aggressive intervention policies designed to prevent MRSA transmission, including use of screening and CP.⁶ In the United States, the Veterans Affairs system utilizes this type of approach and reported a 62% decrease in MRSA rates. Subsequent analysis showed that the downward trend of hospital-onset MRSA infections was observed only among patients who were not carrying MRSA at the time of admission, suggesting that preventing transmission was an important contributor to the overall trends.^7,8 More broadly, healthcare-associated MRSA rates in the United States have decreased dramatically over the past decade,^9,10 a period during which more than 81% of hospitals reported using CP for patients colonized or infected with MRSA as part of the bundle of infection prevention approaches.¹¹ Given these decreases, and the potential role that CP played in achieving these results, we, along with others,¹² urge caution about the dangers of abandoning CP prematurely and without data to indicate that it is safe to stop.

Although some studies report adverse events associated with CP, including a reduced number of visits from healthcare personnel and increased anxiety and depression, these studies rarely control for important confounding variables such as the severity of illness or the presence of anxiety and depression at the time of hospital admission.^13-15 The highest-quality evidence in studies that control for severity of illness and the presence of depression at the time of admission suggests that CP are not associated with an increased incidence of adverse events.^16,17

Interestingly, Young et al. acknowledge that CP are important and should be continued for patients infected or colonized with certain MDROs, including carbapenem-resistant Enterobacteriaceae, multidrug-resistant Pseudomonas aeruginosa, and Candida auris. They even suggest continuing CP for patients with certain types of antimicrobial-resistant Staphylococcus aureus isolates that are resistant or intermediate to vancomycin (Vancomycin-resistant Staphylococcus aureus [VRSA] or Vancomycin-intermediate Staphylococcus aureus [VISA]) and for which transmission has rarely been documented in the United States. It is unclear why they believe that CP are indicated and useful to prevent transmission of these multidrug-resistant pathogens while advocating that CP are not useful or indicated to prevent transmission of MRSA and VRE. One must consider whether it makes sense to use such a selective approach to using CP for patients with some, but not all, MDROs.

The authors state that CP should be employed to help interrupt outbreaks and for patients with high-risk situations such as open wounds, uncontained secretions, or incontinent diarrhea. We agree that there is appeal to a risk-based approach in which CP are applied based on the likelihood that an individual patient may be carrying and shedding an MDRO. However, to our knowledge, there are no validated algorithms available for this purpose, and it appears likely that using such algorithms would result in an increase in the proportion of patients cared for using CP, rather than a decrease.

The use of CP when caring for patients colonized or infected with an MDRO is considered to be a standard of care. Based on experimental, clinical, and epidemiologic studies and a strong theoretical rationale, the use of CP is currently recommended by the United States Centers for Disease Control and Prevention (CDC), the Healthcare Infection Control Practices Advisory Committee (HICPAC),¹⁸ the Society for Healthcare Epidemiology of America (SHEA),¹⁹ and the Infectious Diseases Society of America.²⁰ Many healthcare facilities continue to employ CP for patients with a wide array of MDROs, including MRSA and VRE, and many infection prevention experts continue to support and utilize this approach. In response to the growing movement to discontinue CP, the CDC recently reaffirmed its support and recommendation for the use of CP when caring for patients colonized or infected with MRSA.²¹

In summary, a bundled, multifaceted approach to infection prevention and transmission of MDROs is extremely important, and we caution against stopping CP for MRSA and VRE before data are available on the potential harm of that approach. Study limitations make it difficult to demonstrate the individual contribution of CP, but CP are an important component of a comprehensive infection prevention MDRO bundle that has successfully reduced healthcare-associated MRSA. Well-designed studies that control for confounders such as the severity of illness at the time of admission suggest that CP are not associated with an increased incidence of adverse events. Currently available data do not support a selective approach to utilizing CP for some MDROs while not using CP for others. Current guidelines call for the use of CP for preventing MDRO transmission, including MRSA and VRE. Healthcare facilities need to focus on how to implement CP in a patient-centered manner, rather than abandoning CP for some MDROs.

Disclosures

Dr. Maragakis is a member of the Healthcare Infection Control Practices Advisory Committee for the Centers for Disease Control and Prevention. Dr. Jernigan is an employee of the Centers for Disease Control and Prevention.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Contact precautions (CP), the use of gowns and gloves as personal protective equipment when caring for patients who are colonized or infected with one or more multidrug-resistant organisms (MDROs), is an important infection prevention intervention utilized to prevent pathogens from being transmitted among patients in healthcare settings. Recently, certain healthcare facilities have taken steps to limit the use of CP for patients colonized or infected with MDROs that are considered to be endemic, namely methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococci (VRE). In this issue of the Journal of Hospital Medicine, authors Young et al. argue that CP for MRSA and VRE is an intervention that should be eliminated as part of the Choosing Wisely^® campaign because it is a “thing we do for no reason.”¹ We respectfully disagree with this characterization of CP for MRSA and VRE, and we assert instead that CP are a necessary practice that should be continued.

Young et al. refer to published studies and a recent meta-analysis that did not conclusively show a benefit of CP for MRSA and VRE.² The quasi-experimental studies cited have major methodological flaws that limit their ability to demonstrate the effect of CP. Most importantly, these studies fail to account for the fact that among patients who develop an infection following hospital-acquired MRSA colonization, approximately 70% of the infections are identified after discharge.³ When such studies do not restrict their outcome measure to include only those infections occurring among patients with hospital-acquired colonization, and do not take steps to accurately identify postdischarge infections that occur in such patients, their results are biased toward the null and difficult to interpret. Due to several serious challenges to study feasibility, including the need for an extremely large sample size, a very long period of follow-up, and the need to control for a variety of other concurrent infection prevention measures, there may never be a study that conclusively proves that CP, apart from other infection prevention interventions, has a significant impact. However, despite these limitations, one of the recent multicenter randomized controlled trials, cited by the authors as evidence against the use of CP, was able to demonstrate a significant reduction in MRSA transmission using universal gowns and gloves for all intensive care unit patients, even in sites that utilized other effective strategies, including chlorhexidine bathing.^4,5

In this issue of the Journal of Hospital Medicine^®, Young et al. acknowledge that CP are generally utilized as part of a comprehensive package of infection prevention approaches that also includes hand hygiene, environmental cleaning, antimicrobial stewardship, and evidence-based interventions to prevent device- and procedure-related infections. This multifaceted approach makes it more difficult to determine the attributable effect of CP alone. However, there is a strong rationale for using CP to prevent transmission, and there are numerous examples where the use of bundled approaches that include CP was associated with success. In the Netherlands, CP were part of an aggressive “search and destroy” approach to MRSA associated with almost total elimination of MRSA from hospitals in that country. The United Kingdom achieved an 80% decrease in MRSA bacteremia following a series of aggressive intervention policies designed to prevent MRSA transmission, including use of screening and CP.⁶ In the United States, the Veterans Affairs system utilizes this type of approach and reported a 62% decrease in MRSA rates. Subsequent analysis showed that the downward trend of hospital-onset MRSA infections was observed only among patients who were not carrying MRSA at the time of admission, suggesting that preventing transmission was an important contributor to the overall trends.^7,8 More broadly, healthcare-associated MRSA rates in the United States have decreased dramatically over the past decade,^9,10 a period during which more than 81% of hospitals reported using CP for patients colonized or infected with MRSA as part of the bundle of infection prevention approaches.¹¹ Given these decreases, and the potential role that CP played in achieving these results, we, along with others,¹² urge caution about the dangers of abandoning CP prematurely and without data to indicate that it is safe to stop.

Although some studies report adverse events associated with CP, including a reduced number of visits from healthcare personnel and increased anxiety and depression, these studies rarely control for important confounding variables such as the severity of illness or the presence of anxiety and depression at the time of hospital admission.^13-15 The highest-quality evidence in studies that control for severity of illness and the presence of depression at the time of admission suggests that CP are not associated with an increased incidence of adverse events.^16,17

Interestingly, Young et al. acknowledge that CP are important and should be continued for patients infected or colonized with certain MDROs, including carbapenem-resistant Enterobacteriaceae, multidrug-resistant Pseudomonas aeruginosa, and Candida auris. They even suggest continuing CP for patients with certain types of antimicrobial-resistant Staphylococcus aureus isolates that are resistant or intermediate to vancomycin (Vancomycin-resistant Staphylococcus aureus [VRSA] or Vancomycin-intermediate Staphylococcus aureus [VISA]) and for which transmission has rarely been documented in the United States. It is unclear why they believe that CP are indicated and useful to prevent transmission of these multidrug-resistant pathogens while advocating that CP are not useful or indicated to prevent transmission of MRSA and VRE. One must consider whether it makes sense to use such a selective approach to using CP for patients with some, but not all, MDROs.

The authors state that CP should be employed to help interrupt outbreaks and for patients with high-risk situations such as open wounds, uncontained secretions, or incontinent diarrhea. We agree that there is appeal to a risk-based approach in which CP are applied based on the likelihood that an individual patient may be carrying and shedding an MDRO. However, to our knowledge, there are no validated algorithms available for this purpose, and it appears likely that using such algorithms would result in an increase in the proportion of patients cared for using CP, rather than a decrease.

The use of CP when caring for patients colonized or infected with an MDRO is considered to be a standard of care. Based on experimental, clinical, and epidemiologic studies and a strong theoretical rationale, the use of CP is currently recommended by the United States Centers for Disease Control and Prevention (CDC), the Healthcare Infection Control Practices Advisory Committee (HICPAC),¹⁸ the Society for Healthcare Epidemiology of America (SHEA),¹⁹ and the Infectious Diseases Society of America.²⁰ Many healthcare facilities continue to employ CP for patients with a wide array of MDROs, including MRSA and VRE, and many infection prevention experts continue to support and utilize this approach. In response to the growing movement to discontinue CP, the CDC recently reaffirmed its support and recommendation for the use of CP when caring for patients colonized or infected with MRSA.²¹

In summary, a bundled, multifaceted approach to infection prevention and transmission of MDROs is extremely important, and we caution against stopping CP for MRSA and VRE before data are available on the potential harm of that approach. Study limitations make it difficult to demonstrate the individual contribution of CP, but CP are an important component of a comprehensive infection prevention MDRO bundle that has successfully reduced healthcare-associated MRSA. Well-designed studies that control for confounders such as the severity of illness at the time of admission suggest that CP are not associated with an increased incidence of adverse events. Currently available data do not support a selective approach to utilizing CP for some MDROs while not using CP for others. Current guidelines call for the use of CP for preventing MDRO transmission, including MRSA and VRE. Healthcare facilities need to focus on how to implement CP in a patient-centered manner, rather than abandoning CP for some MDROs.

Disclosures

Dr. Maragakis is a member of the Healthcare Infection Control Practices Advisory Committee for the Centers for Disease Control and Prevention. Dr. Jernigan is an employee of the Centers for Disease Control and Prevention.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Contact precautions (CP), the use of gowns and gloves as personal protective equipment when caring for patients who are colonized or infected with one or more multidrug-resistant organisms (MDROs), is an important infection prevention intervention utilized to prevent pathogens from being transmitted among patients in healthcare settings. Recently, certain healthcare facilities have taken steps to limit the use of CP for patients colonized or infected with MDROs that are considered to be endemic, namely methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococci (VRE). In this issue of the Journal of Hospital Medicine, authors Young et al. argue that CP for MRSA and VRE is an intervention that should be eliminated as part of the Choosing Wisely^® campaign because it is a “thing we do for no reason.”¹ We respectfully disagree with this characterization of CP for MRSA and VRE, and we assert instead that CP are a necessary practice that should be continued.

Young et al. refer to published studies and a recent meta-analysis that did not conclusively show a benefit of CP for MRSA and VRE.² The quasi-experimental studies cited have major methodological flaws that limit their ability to demonstrate the effect of CP. Most importantly, these studies fail to account for the fact that among patients who develop an infection following hospital-acquired MRSA colonization, approximately 70% of the infections are identified after discharge.³ When such studies do not restrict their outcome measure to include only those infections occurring among patients with hospital-acquired colonization, and do not take steps to accurately identify postdischarge infections that occur in such patients, their results are biased toward the null and difficult to interpret. Due to several serious challenges to study feasibility, including the need for an extremely large sample size, a very long period of follow-up, and the need to control for a variety of other concurrent infection prevention measures, there may never be a study that conclusively proves that CP, apart from other infection prevention interventions, has a significant impact. However, despite these limitations, one of the recent multicenter randomized controlled trials, cited by the authors as evidence against the use of CP, was able to demonstrate a significant reduction in MRSA transmission using universal gowns and gloves for all intensive care unit patients, even in sites that utilized other effective strategies, including chlorhexidine bathing.^4,5

In this issue of the Journal of Hospital Medicine^®, Young et al. acknowledge that CP are generally utilized as part of a comprehensive package of infection prevention approaches that also includes hand hygiene, environmental cleaning, antimicrobial stewardship, and evidence-based interventions to prevent device- and procedure-related infections. This multifaceted approach makes it more difficult to determine the attributable effect of CP alone. However, there is a strong rationale for using CP to prevent transmission, and there are numerous examples where the use of bundled approaches that include CP was associated with success. In the Netherlands, CP were part of an aggressive “search and destroy” approach to MRSA associated with almost total elimination of MRSA from hospitals in that country. The United Kingdom achieved an 80% decrease in MRSA bacteremia following a series of aggressive intervention policies designed to prevent MRSA transmission, including use of screening and CP.⁶ In the United States, the Veterans Affairs system utilizes this type of approach and reported a 62% decrease in MRSA rates. Subsequent analysis showed that the downward trend of hospital-onset MRSA infections was observed only among patients who were not carrying MRSA at the time of admission, suggesting that preventing transmission was an important contributor to the overall trends.^7,8 More broadly, healthcare-associated MRSA rates in the United States have decreased dramatically over the past decade,^9,10 a period during which more than 81% of hospitals reported using CP for patients colonized or infected with MRSA as part of the bundle of infection prevention approaches.¹¹ Given these decreases, and the potential role that CP played in achieving these results, we, along with others,¹² urge caution about the dangers of abandoning CP prematurely and without data to indicate that it is safe to stop.

Although some studies report adverse events associated with CP, including a reduced number of visits from healthcare personnel and increased anxiety and depression, these studies rarely control for important confounding variables such as the severity of illness or the presence of anxiety and depression at the time of hospital admission.^13-15 The highest-quality evidence in studies that control for severity of illness and the presence of depression at the time of admission suggests that CP are not associated with an increased incidence of adverse events.^16,17

Interestingly, Young et al. acknowledge that CP are important and should be continued for patients infected or colonized with certain MDROs, including carbapenem-resistant Enterobacteriaceae, multidrug-resistant Pseudomonas aeruginosa, and Candida auris. They even suggest continuing CP for patients with certain types of antimicrobial-resistant Staphylococcus aureus isolates that are resistant or intermediate to vancomycin (Vancomycin-resistant Staphylococcus aureus [VRSA] or Vancomycin-intermediate Staphylococcus aureus [VISA]) and for which transmission has rarely been documented in the United States. It is unclear why they believe that CP are indicated and useful to prevent transmission of these multidrug-resistant pathogens while advocating that CP are not useful or indicated to prevent transmission of MRSA and VRE. One must consider whether it makes sense to use such a selective approach to using CP for patients with some, but not all, MDROs.

The authors state that CP should be employed to help interrupt outbreaks and for patients with high-risk situations such as open wounds, uncontained secretions, or incontinent diarrhea. We agree that there is appeal to a risk-based approach in which CP are applied based on the likelihood that an individual patient may be carrying and shedding an MDRO. However, to our knowledge, there are no validated algorithms available for this purpose, and it appears likely that using such algorithms would result in an increase in the proportion of patients cared for using CP, rather than a decrease.

The use of CP when caring for patients colonized or infected with an MDRO is considered to be a standard of care. Based on experimental, clinical, and epidemiologic studies and a strong theoretical rationale, the use of CP is currently recommended by the United States Centers for Disease Control and Prevention (CDC), the Healthcare Infection Control Practices Advisory Committee (HICPAC),¹⁸ the Society for Healthcare Epidemiology of America (SHEA),¹⁹ and the Infectious Diseases Society of America.²⁰ Many healthcare facilities continue to employ CP for patients with a wide array of MDROs, including MRSA and VRE, and many infection prevention experts continue to support and utilize this approach. In response to the growing movement to discontinue CP, the CDC recently reaffirmed its support and recommendation for the use of CP when caring for patients colonized or infected with MRSA.²¹

In summary, a bundled, multifaceted approach to infection prevention and transmission of MDROs is extremely important, and we caution against stopping CP for MRSA and VRE before data are available on the potential harm of that approach. Study limitations make it difficult to demonstrate the individual contribution of CP, but CP are an important component of a comprehensive infection prevention MDRO bundle that has successfully reduced healthcare-associated MRSA. Well-designed studies that control for confounders such as the severity of illness at the time of admission suggest that CP are not associated with an increased incidence of adverse events. Currently available data do not support a selective approach to utilizing CP for some MDROs while not using CP for others. Current guidelines call for the use of CP for preventing MDRO transmission, including MRSA and VRE. Healthcare facilities need to focus on how to implement CP in a patient-centered manner, rather than abandoning CP for some MDROs.

Disclosures

Dr. Maragakis is a member of the Healthcare Infection Control Practices Advisory Committee for the Centers for Disease Control and Prevention. Dr. Jernigan is an employee of the Centers for Disease Control and Prevention.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

In 2015, the Centers for Medicare and Medicaid Services (CMS) tied 42% of Medicare payments to a value-based model of care.¹ Many of these models are designed to expand the scope of hospitals’ accountability to include care provided to patients postdischarge (eg, readmission penalties, bundled payments, accountable care organizations). With such a significant change in organizational incentives, one would expect to see activity as it relates to hospital-skilled nursing facility (SNF) integration, potentially including shared risk among providers.^2,3

Hospitals can choose from several different strategies when contemplating SNF integration, such as vertical integration with SNFs, which would involve acquiring and owning SNFs. However, despite the high level of incentive alignment and financial integration achieved through SNF acquisition, this strategy has not been widely adopted. Perhaps this is because hospitals can often attain a shorter length of stay and lower readmission rates without taking on the additional risk of owning a facility, except under particular market conditions.⁴ Hospitals can alternatively pursue virtual integration by developing preferred provider networks through contractual relationships or other formal processes, attempting to direct patients to SNF providers that have met predefined criteria, as described by Conway and colleagues in this issue of the Journal of Hospital Medicine^®.⁵ While hospitals have adopted this form of integration more widely than vertical integration, only those with additional financial motivations, such as those employing bundled payments, engaged in accountable care organizations (ACOs) or forward-thinking organizations preparing for looming global models of payment, have implemented such action. Finally, hospitals can focus on relational coordination through informal person-to-person communication and transition management. Given the high number of patients discharged to SNFs, the strategies above are not mutually exclusive, and enhanced relational coordination is most likely going to occur regardless of the type of—and perhaps even without—organizational-level integration.

For those hospitals choosing not to pursue integration with SNFs, there are several reasons to maintain the status quo. First, hospitals have different interpretations of provider choice (“beneficiary freedom to choose”), whereby many do not believe they can provide information to patients outside of facility names and addresses. As such, they will refrain from developing a SNF network due to their interpretation of hazy federal rules.⁶ Second, it is possible that the incremental benefit of establishing a network is viewed by many hospitals as not worth the cost, measured by the time and effort required and the potential risk of not adhering to choice requirements. This could be especially true for hospitals without additional financial motivations, such as participation in an ACO or bundled payment program.

As the landscape continues to evolve, more successful systems will embrace a more concordant partnership with local and regional SNF providers, and several market factors will support the trend. First, the Medicare Payment Advisory Commission (MedPAC) is discussing the idea of choice in the context of postacute discharge, potentially leading to hospitals relaxing their strict interpretations of choice and the level of information provided to patients.⁷ Second, the evidence supports better patient outcomes when hospitals develop SNF networks.^8,9 Finally, continued penetration of value-based payment models combined with CMS decisions regarding choice will continue to provide the additional motivation hospitals may need to change the cost-benefit calculation in favor of developing a network.

IMPLICATIONS FOR HOSPITALISTS

Traditionally, primary care physicians followed their patients through the acute- and postacute care continuum, but a variety of changes led to the growth of hospital medicine as fewer primary care physicians saw patients in the hospital.^10,11 This shift has challenged efforts to ensure continuity of care across settings, especially since most hospitalists have ceded control of postdischarge placement to case managers and therapists. Further, there has been little incentive to connect hospitalists to any other component or provider along the range of care, and compensation models rarely, if at all, consider any accountability for patient outcomes outside the hospital. Several factors can change this reality for hospitalists.

First, as more providers adopt team-based care approaches and as alternative payment models expand the scope of accountability, hospitalists will become an even more central component of the risk evaluation process for hospitalized patients as it relates to their discharge profile. This could mean that hospitalists are more involved in the postdischarge follow-up of patients sent home, to make sure patients adhere to discharge instructions. Alternatively, hospitalists may need to increase the level of physician-to-physician communication with SNF medical directors for patients discharged to SNF. This, in turn, could result in an increasing number of hospitalist groups recruiting SNFists to join their group or potentially assigning existing hospitalists or physician assistants to round on patients in the SNF. The 2018 Society of Hospital Medicine report showed an increase in activity among hospital medicine groups performing services in postacute-care facilities outside the hospital from 13% in 2016 to 25% in 2018.¹² Similarly, a 2017 study in JAMA Internal Medicine reported a 48.2% increase in the number of physicians classified as SNFists from 2007 to 2014.¹³

Second, hospitalists will be more involved in the discharge planning process through internal interdisciplinary team communications. Whereas case managers and therapists owned the discharge planning process historically, new teams will include hospitalists, case managers, physical therapists, and pharmacists. System leaders will task them with identifying the appropriate discharge destination (eg, SNF, home health), finalizing the medication reconciliation, scheduling follow-up appointments, and completing a warm handoff.

Finally, as the field matures and hospitalists learn more about postacute-care connections, they will continue to be held more accountable for patient outcomes postdischarge. Many hospitalists have already connected to community providers through checklists and use evidence-based discharge programs like ProjectRed or Project BOOST.^14,15 Organizations will need a similar strategy for SNFs, developing process measures, with the input of hospitalists, around those noteworthy areas that hospitalists can control. This will require greater alignment among constituents around overall organizational goals and, more importantly, entail the hospitalist to be attuned to overall patient goals beyond the care provided in the hospital setting.

As payment and care models continue to evolve, the status quo cannot be sustained. We anticipate that hospitalists will become more integrated into the patient discharge process, especially as it relates to discharge to SNFs before patients reconnect to their community physicians. Hospital systems will accelerate integration through the development of preferred SNF networks, and hospitalists stand to play a critical role in the success of these arrangements by enriching the benefits they create through these outward relationships.

For organizations engaged in embedded networks, they can realize gains via incentive alignment, trust, information transfer, mutual support, and coordination through virtual integration, without requiring vertical ownership.^3,16Thus, the opportunity exists for hospitalists to be critical drivers of network success, serving as intermediaries from which information, collaboration, and shared problem-solving flow between hospitals, SNFs, patients, and the entire care team. Opportunities to rebuild our system are long past; however, like all changing sectors in healthcare, the disaggregate acute and postacute settings must move in lockstep. Hospitals and postacute care facilities must find ways to alter their thinking to eradicate the obstructive and injurious invisible wall.

Disclosures

The authors have nothing to disclose.

In 2015, the Centers for Medicare and Medicaid Services (CMS) tied 42% of Medicare payments to a value-based model of care.¹ Many of these models are designed to expand the scope of hospitals’ accountability to include care provided to patients postdischarge (eg, readmission penalties, bundled payments, accountable care organizations). With such a significant change in organizational incentives, one would expect to see activity as it relates to hospital-skilled nursing facility (SNF) integration, potentially including shared risk among providers.^2,3

Hospitals can choose from several different strategies when contemplating SNF integration, such as vertical integration with SNFs, which would involve acquiring and owning SNFs. However, despite the high level of incentive alignment and financial integration achieved through SNF acquisition, this strategy has not been widely adopted. Perhaps this is because hospitals can often attain a shorter length of stay and lower readmission rates without taking on the additional risk of owning a facility, except under particular market conditions.⁴ Hospitals can alternatively pursue virtual integration by developing preferred provider networks through contractual relationships or other formal processes, attempting to direct patients to SNF providers that have met predefined criteria, as described by Conway and colleagues in this issue of the Journal of Hospital Medicine^®.⁵ While hospitals have adopted this form of integration more widely than vertical integration, only those with additional financial motivations, such as those employing bundled payments, engaged in accountable care organizations (ACOs) or forward-thinking organizations preparing for looming global models of payment, have implemented such action. Finally, hospitals can focus on relational coordination through informal person-to-person communication and transition management. Given the high number of patients discharged to SNFs, the strategies above are not mutually exclusive, and enhanced relational coordination is most likely going to occur regardless of the type of—and perhaps even without—organizational-level integration.

For those hospitals choosing not to pursue integration with SNFs, there are several reasons to maintain the status quo. First, hospitals have different interpretations of provider choice (“beneficiary freedom to choose”), whereby many do not believe they can provide information to patients outside of facility names and addresses. As such, they will refrain from developing a SNF network due to their interpretation of hazy federal rules.⁶ Second, it is possible that the incremental benefit of establishing a network is viewed by many hospitals as not worth the cost, measured by the time and effort required and the potential risk of not adhering to choice requirements. This could be especially true for hospitals without additional financial motivations, such as participation in an ACO or bundled payment program.

As the landscape continues to evolve, more successful systems will embrace a more concordant partnership with local and regional SNF providers, and several market factors will support the trend. First, the Medicare Payment Advisory Commission (MedPAC) is discussing the idea of choice in the context of postacute discharge, potentially leading to hospitals relaxing their strict interpretations of choice and the level of information provided to patients.⁷ Second, the evidence supports better patient outcomes when hospitals develop SNF networks.^8,9 Finally, continued penetration of value-based payment models combined with CMS decisions regarding choice will continue to provide the additional motivation hospitals may need to change the cost-benefit calculation in favor of developing a network.

IMPLICATIONS FOR HOSPITALISTS

Traditionally, primary care physicians followed their patients through the acute- and postacute care continuum, but a variety of changes led to the growth of hospital medicine as fewer primary care physicians saw patients in the hospital.^10,11 This shift has challenged efforts to ensure continuity of care across settings, especially since most hospitalists have ceded control of postdischarge placement to case managers and therapists. Further, there has been little incentive to connect hospitalists to any other component or provider along the range of care, and compensation models rarely, if at all, consider any accountability for patient outcomes outside the hospital. Several factors can change this reality for hospitalists.

First, as more providers adopt team-based care approaches and as alternative payment models expand the scope of accountability, hospitalists will become an even more central component of the risk evaluation process for hospitalized patients as it relates to their discharge profile. This could mean that hospitalists are more involved in the postdischarge follow-up of patients sent home, to make sure patients adhere to discharge instructions. Alternatively, hospitalists may need to increase the level of physician-to-physician communication with SNF medical directors for patients discharged to SNF. This, in turn, could result in an increasing number of hospitalist groups recruiting SNFists to join their group or potentially assigning existing hospitalists or physician assistants to round on patients in the SNF. The 2018 Society of Hospital Medicine report showed an increase in activity among hospital medicine groups performing services in postacute-care facilities outside the hospital from 13% in 2016 to 25% in 2018.¹² Similarly, a 2017 study in JAMA Internal Medicine reported a 48.2% increase in the number of physicians classified as SNFists from 2007 to 2014.¹³

Second, hospitalists will be more involved in the discharge planning process through internal interdisciplinary team communications. Whereas case managers and therapists owned the discharge planning process historically, new teams will include hospitalists, case managers, physical therapists, and pharmacists. System leaders will task them with identifying the appropriate discharge destination (eg, SNF, home health), finalizing the medication reconciliation, scheduling follow-up appointments, and completing a warm handoff.

Finally, as the field matures and hospitalists learn more about postacute-care connections, they will continue to be held more accountable for patient outcomes postdischarge. Many hospitalists have already connected to community providers through checklists and use evidence-based discharge programs like ProjectRed or Project BOOST.^14,15 Organizations will need a similar strategy for SNFs, developing process measures, with the input of hospitalists, around those noteworthy areas that hospitalists can control. This will require greater alignment among constituents around overall organizational goals and, more importantly, entail the hospitalist to be attuned to overall patient goals beyond the care provided in the hospital setting.

As payment and care models continue to evolve, the status quo cannot be sustained. We anticipate that hospitalists will become more integrated into the patient discharge process, especially as it relates to discharge to SNFs before patients reconnect to their community physicians. Hospital systems will accelerate integration through the development of preferred SNF networks, and hospitalists stand to play a critical role in the success of these arrangements by enriching the benefits they create through these outward relationships.

For organizations engaged in embedded networks, they can realize gains via incentive alignment, trust, information transfer, mutual support, and coordination through virtual integration, without requiring vertical ownership.^3,16Thus, the opportunity exists for hospitalists to be critical drivers of network success, serving as intermediaries from which information, collaboration, and shared problem-solving flow between hospitals, SNFs, patients, and the entire care team. Opportunities to rebuild our system are long past; however, like all changing sectors in healthcare, the disaggregate acute and postacute settings must move in lockstep. Hospitals and postacute care facilities must find ways to alter their thinking to eradicate the obstructive and injurious invisible wall.

Disclosures

The authors have nothing to disclose.

In 2015, the Centers for Medicare and Medicaid Services (CMS) tied 42% of Medicare payments to a value-based model of care.¹ Many of these models are designed to expand the scope of hospitals’ accountability to include care provided to patients postdischarge (eg, readmission penalties, bundled payments, accountable care organizations). With such a significant change in organizational incentives, one would expect to see activity as it relates to hospital-skilled nursing facility (SNF) integration, potentially including shared risk among providers.^2,3

Hospitals can choose from several different strategies when contemplating SNF integration, such as vertical integration with SNFs, which would involve acquiring and owning SNFs. However, despite the high level of incentive alignment and financial integration achieved through SNF acquisition, this strategy has not been widely adopted. Perhaps this is because hospitals can often attain a shorter length of stay and lower readmission rates without taking on the additional risk of owning a facility, except under particular market conditions.⁴ Hospitals can alternatively pursue virtual integration by developing preferred provider networks through contractual relationships or other formal processes, attempting to direct patients to SNF providers that have met predefined criteria, as described by Conway and colleagues in this issue of the Journal of Hospital Medicine^®.⁵ While hospitals have adopted this form of integration more widely than vertical integration, only those with additional financial motivations, such as those employing bundled payments, engaged in accountable care organizations (ACOs) or forward-thinking organizations preparing for looming global models of payment, have implemented such action. Finally, hospitals can focus on relational coordination through informal person-to-person communication and transition management. Given the high number of patients discharged to SNFs, the strategies above are not mutually exclusive, and enhanced relational coordination is most likely going to occur regardless of the type of—and perhaps even without—organizational-level integration.

For those hospitals choosing not to pursue integration with SNFs, there are several reasons to maintain the status quo. First, hospitals have different interpretations of provider choice (“beneficiary freedom to choose”), whereby many do not believe they can provide information to patients outside of facility names and addresses. As such, they will refrain from developing a SNF network due to their interpretation of hazy federal rules.⁶ Second, it is possible that the incremental benefit of establishing a network is viewed by many hospitals as not worth the cost, measured by the time and effort required and the potential risk of not adhering to choice requirements. This could be especially true for hospitals without additional financial motivations, such as participation in an ACO or bundled payment program.

As the landscape continues to evolve, more successful systems will embrace a more concordant partnership with local and regional SNF providers, and several market factors will support the trend. First, the Medicare Payment Advisory Commission (MedPAC) is discussing the idea of choice in the context of postacute discharge, potentially leading to hospitals relaxing their strict interpretations of choice and the level of information provided to patients.⁷ Second, the evidence supports better patient outcomes when hospitals develop SNF networks.^8,9 Finally, continued penetration of value-based payment models combined with CMS decisions regarding choice will continue to provide the additional motivation hospitals may need to change the cost-benefit calculation in favor of developing a network.

IMPLICATIONS FOR HOSPITALISTS

Traditionally, primary care physicians followed their patients through the acute- and postacute care continuum, but a variety of changes led to the growth of hospital medicine as fewer primary care physicians saw patients in the hospital.^10,11 This shift has challenged efforts to ensure continuity of care across settings, especially since most hospitalists have ceded control of postdischarge placement to case managers and therapists. Further, there has been little incentive to connect hospitalists to any other component or provider along the range of care, and compensation models rarely, if at all, consider any accountability for patient outcomes outside the hospital. Several factors can change this reality for hospitalists.

First, as more providers adopt team-based care approaches and as alternative payment models expand the scope of accountability, hospitalists will become an even more central component of the risk evaluation process for hospitalized patients as it relates to their discharge profile. This could mean that hospitalists are more involved in the postdischarge follow-up of patients sent home, to make sure patients adhere to discharge instructions. Alternatively, hospitalists may need to increase the level of physician-to-physician communication with SNF medical directors for patients discharged to SNF. This, in turn, could result in an increasing number of hospitalist groups recruiting SNFists to join their group or potentially assigning existing hospitalists or physician assistants to round on patients in the SNF. The 2018 Society of Hospital Medicine report showed an increase in activity among hospital medicine groups performing services in postacute-care facilities outside the hospital from 13% in 2016 to 25% in 2018.¹² Similarly, a 2017 study in JAMA Internal Medicine reported a 48.2% increase in the number of physicians classified as SNFists from 2007 to 2014.¹³

Second, hospitalists will be more involved in the discharge planning process through internal interdisciplinary team communications. Whereas case managers and therapists owned the discharge planning process historically, new teams will include hospitalists, case managers, physical therapists, and pharmacists. System leaders will task them with identifying the appropriate discharge destination (eg, SNF, home health), finalizing the medication reconciliation, scheduling follow-up appointments, and completing a warm handoff.

Finally, as the field matures and hospitalists learn more about postacute-care connections, they will continue to be held more accountable for patient outcomes postdischarge. Many hospitalists have already connected to community providers through checklists and use evidence-based discharge programs like ProjectRed or Project BOOST.^14,15 Organizations will need a similar strategy for SNFs, developing process measures, with the input of hospitalists, around those noteworthy areas that hospitalists can control. This will require greater alignment among constituents around overall organizational goals and, more importantly, entail the hospitalist to be attuned to overall patient goals beyond the care provided in the hospital setting.

As payment and care models continue to evolve, the status quo cannot be sustained. We anticipate that hospitalists will become more integrated into the patient discharge process, especially as it relates to discharge to SNFs before patients reconnect to their community physicians. Hospital systems will accelerate integration through the development of preferred SNF networks, and hospitalists stand to play a critical role in the success of these arrangements by enriching the benefits they create through these outward relationships.

For organizations engaged in embedded networks, they can realize gains via incentive alignment, trust, information transfer, mutual support, and coordination through virtual integration, without requiring vertical ownership.^3,16Thus, the opportunity exists for hospitalists to be critical drivers of network success, serving as intermediaries from which information, collaboration, and shared problem-solving flow between hospitals, SNFs, patients, and the entire care team. Opportunities to rebuild our system are long past; however, like all changing sectors in healthcare, the disaggregate acute and postacute settings must move in lockstep. Hospitals and postacute care facilities must find ways to alter their thinking to eradicate the obstructive and injurious invisible wall.

Disclosures

The authors have nothing to disclose.

Hospitals and health systems are under mounting financial pressure to shorten hospitalizations and reduce readmissions. These priorities have led to an ever-increasing focus on postacute care (PAC), and more specifically on improving transitions from the hospital.^1,2 According to a 2013 Institute of Medicine report, PAC is the source of 73% of the variation in Medicare spending³ and readmissions during the postacute episode nearly double the average Medicare payment.⁴ Within the PAC landscape, discharges to skilled nursing facilities (SNFs) have received particular focus due to the high rates of readmission and associated care costs.⁵

Hospitals, hospital physicians, PAC providers, and payers need to improve SNF transitions in care. Hospitals are increasingly responsible for patient care beyond their walls through several mechanisms including rehospitalization penalties, value-based reimbursement strategies (eg, bundled payments), and risk-based contracting on the total cost of care through relationships with accountable care organizations (ACOs) and Medicare Advantage plans. Similarly, hospital-employed physicians and PAC providers are more engaged in achieving value-based goals through increased alignment of provider compensation models^6,7 with risk-based contracting.

Current evidence suggests that rehospitalizations could be reduced by focusing on a concentrated referral network of preferred high-quality SNFs;^8,9 however, less is known about how to develop and operate such linkages at the administrative or clinical levels.⁸ In this article, we propose a collaborative framework for the establishment of a preferred PAC network.

One mechanism employed to improve transitions to SNFs and reduce associated readmissions is to create a preferred provider network. Increasing the concentration of hospital discharges to higher performing facilities is associated with lower rehospitalization rates, particularly during the critical days following discharge.¹⁰

While the criteria applied for preferred provider networks vary, there are several emerging themes.¹⁰ Quality metrics are often applied, generally starting with Centers for Medicare and Medicaid Services (CMS) quality star ratings and Long-Term Care Minimum Data Set (MDS) metrics with additional criteria frequently layered upon those. Some examples include the extent of physician coverage,¹¹ the extent of nursing coverage (eg, nursing ratios or 24/7 nursing care), geographic access, and flexible admission times (including weekends and nights).¹² In addition, several outcome measures may be used such as 30-day readmission rates, patient/family satisfaction ratings, ED visits, primary care follow-up within seven days of PAC discharge, or impact on the total cost of care.

Beyond the specified criteria, some hospitals choose to build upon existing relationships when developing their preferred network. By selecting historically high-volume facilities, they are able to leverage the existing name recognition amongst patients and providers.¹³ This minimizes retraining of discharge planners, maintains institutional relationships, and aligns with the patients’ geographic preferences.^2,13 While the high volume SNFs may not have the highest quality ratings, some hospitals find they can leverage the value of preferred partner status to push behavior change and improve performance.¹³

Here we propose a framework for the establishment of a preferred provider network for a hospital or health system based on the early experience of establishing an SNF Collaborative within Johns Hopkins Medicine (JHM). JHM is a large integrated health care system, which includes five hospitals within the region, including two large academic hospitals and three community hospitals serving patients in Maryland and the District of Columbia.¹⁴

JHM identified a need for improved coordination with PAC providers and saw opportunities to build upon successful individual hospital efforts to create a system-level approach with a PAC partnership sharing the goals of improving care and reducing costs. Additional opportunities exist given the unique Maryland all-payer Global Budget Revenue system managed by the Health Services Cost Review Commission. This system imposes hospital-level penalties for readmissions or poor quality measure performance and is moving to a new phase that will place hospitals directly at risk for the total Part A and Part B Medicare expenditures for a cohort of attributed Medicare patients, inclusive of their PAC expenses. This state-wide program is one example of a shift in payment structures from volume to value that is occurring throughout the healthcare sector.

Developing a formal collaboration inclusive of the five local hospitals, Johns Hopkins HealthCare (JHHC)—the managed care division of JHM—and the JHM ACO (Johns Hopkins Medicine Alliance for Patients, JMAP), we established a JHM SNF Collaborative. This group was tasked with improving the continuum of care for our patients discharged to PAC facilities. Given the number and diversity of entities involved, we sought to draw on efforts already managed and piloted locally, while disseminating best practices and providing added services at the collaborative level. We propose a collaborative multistakeholder model (Figure) that we anticipate will be adaptable to other health systems.

In the process of establishing the SNF Collaborative at JHM, there were a number of early challenges faced and lessons learned:

• In a large integrated delivery system, there is a need to balance the benefits of central coordination with the support for ongoing local efforts to promote partner engagement at the hospital and SNF level. The forums created within the collaborative governance structure can facilitate sharing of the prior health system, hospital or SNF initiatives to grow upon successes and avoid prior pitfalls.
• Early identification of risk-adjusted PAC data sources is central to the collaborative establishment and goal setting. This requires assessment of internal analytic resources, budget, and desired timeline for implementation to determine the optimal arrangement. Similarly, identification of available data sources to drive the analytic efforts is essential and should include a health information exchange, claims, and MDS among others.
• Partnering with local QIOs provides support for facility-level quality improvement efforts. They have the staff and onsite expertise to facilitate process implementation within individual SNFs.
• Larger preferred provider networks require considerable administrative support to facilitate communication with the entities, coordinate completion of network agreements, and manage the dissemination of SNF- and hospital-specific performance data.
• Legal and contractual support related to data sharing and HIPAA compliance is needed due to the complexity of the health system and SNF legal structure. Multiple JHM legal entities were involved in this collaborative as were a mixture of freestanding SNFs and corporate chains. There was a significant effort required to execute both data-sharing agreements as well as charters to enable QIO participation.
• Physician leadership and insight are key to implementing meaningful and broad change. When devising system-wide solutions, incorporation and respect for local processes and needs are paramount for provider engagement and behavior change. This process will likely identify gaps in understanding the PAC patient’s experience and needs. It may also reveal practice variability and foster opportunities for provider education on the needs of PAC teams and how to best facilitate quality transitions.

We proposed a framework for establishing a collaborative partnership with a preferred network of SNF providers. Depending on organizational readiness, significant upfront investment of time and resources could be needed to establish a coordinated network of SNF providers. However, once established, such networks can be leveraged to support ongoing process improvement efforts within a hospital or delivery system and can be used strategically by such health systems as they implement value-based health strategies. Furthermore, the lessons learned from transitions to SNFs can be applied more broadly in the PAC landscape including transitions to home from both the hospital and SNF.

Acknowledgments

The authors wish to acknowledge all the members and participants in the Johns Hopkins Medicine Skilled Nursing Facility Collaborative and the executive sponsors and JHM hospital presidents for their support of this work.

Disclosures

Michele Bellantoni receives intramural salary support for being the medical director of the JHM SNF Collaborative. Damien Doyle is a part-time geriatrician at the Hebrew Home of Greater Washington, a skilled nursing facility. He received travel expense support for GAPNA, a local Advanced Practice Nurse Association meeting.The authors otherwise have no potential conflicts of interest to disclose.

Funding

The authors state that there were no external sponsors for this work.

Hospitals and health systems are under mounting financial pressure to shorten hospitalizations and reduce readmissions. These priorities have led to an ever-increasing focus on postacute care (PAC), and more specifically on improving transitions from the hospital.^1,2 According to a 2013 Institute of Medicine report, PAC is the source of 73% of the variation in Medicare spending³ and readmissions during the postacute episode nearly double the average Medicare payment.⁴ Within the PAC landscape, discharges to skilled nursing facilities (SNFs) have received particular focus due to the high rates of readmission and associated care costs.⁵

Hospitals, hospital physicians, PAC providers, and payers need to improve SNF transitions in care. Hospitals are increasingly responsible for patient care beyond their walls through several mechanisms including rehospitalization penalties, value-based reimbursement strategies (eg, bundled payments), and risk-based contracting on the total cost of care through relationships with accountable care organizations (ACOs) and Medicare Advantage plans. Similarly, hospital-employed physicians and PAC providers are more engaged in achieving value-based goals through increased alignment of provider compensation models^6,7 with risk-based contracting.

Current evidence suggests that rehospitalizations could be reduced by focusing on a concentrated referral network of preferred high-quality SNFs;^8,9 however, less is known about how to develop and operate such linkages at the administrative or clinical levels.⁸ In this article, we propose a collaborative framework for the establishment of a preferred PAC network.

One mechanism employed to improve transitions to SNFs and reduce associated readmissions is to create a preferred provider network. Increasing the concentration of hospital discharges to higher performing facilities is associated with lower rehospitalization rates, particularly during the critical days following discharge.¹⁰

While the criteria applied for preferred provider networks vary, there are several emerging themes.¹⁰ Quality metrics are often applied, generally starting with Centers for Medicare and Medicaid Services (CMS) quality star ratings and Long-Term Care Minimum Data Set (MDS) metrics with additional criteria frequently layered upon those. Some examples include the extent of physician coverage,¹¹ the extent of nursing coverage (eg, nursing ratios or 24/7 nursing care), geographic access, and flexible admission times (including weekends and nights).¹² In addition, several outcome measures may be used such as 30-day readmission rates, patient/family satisfaction ratings, ED visits, primary care follow-up within seven days of PAC discharge, or impact on the total cost of care.

Beyond the specified criteria, some hospitals choose to build upon existing relationships when developing their preferred network. By selecting historically high-volume facilities, they are able to leverage the existing name recognition amongst patients and providers.¹³ This minimizes retraining of discharge planners, maintains institutional relationships, and aligns with the patients’ geographic preferences.^2,13 While the high volume SNFs may not have the highest quality ratings, some hospitals find they can leverage the value of preferred partner status to push behavior change and improve performance.¹³

Here we propose a framework for the establishment of a preferred provider network for a hospital or health system based on the early experience of establishing an SNF Collaborative within Johns Hopkins Medicine (JHM). JHM is a large integrated health care system, which includes five hospitals within the region, including two large academic hospitals and three community hospitals serving patients in Maryland and the District of Columbia.¹⁴

JHM identified a need for improved coordination with PAC providers and saw opportunities to build upon successful individual hospital efforts to create a system-level approach with a PAC partnership sharing the goals of improving care and reducing costs. Additional opportunities exist given the unique Maryland all-payer Global Budget Revenue system managed by the Health Services Cost Review Commission. This system imposes hospital-level penalties for readmissions or poor quality measure performance and is moving to a new phase that will place hospitals directly at risk for the total Part A and Part B Medicare expenditures for a cohort of attributed Medicare patients, inclusive of their PAC expenses. This state-wide program is one example of a shift in payment structures from volume to value that is occurring throughout the healthcare sector.

Developing a formal collaboration inclusive of the five local hospitals, Johns Hopkins HealthCare (JHHC)—the managed care division of JHM—and the JHM ACO (Johns Hopkins Medicine Alliance for Patients, JMAP), we established a JHM SNF Collaborative. This group was tasked with improving the continuum of care for our patients discharged to PAC facilities. Given the number and diversity of entities involved, we sought to draw on efforts already managed and piloted locally, while disseminating best practices and providing added services at the collaborative level. We propose a collaborative multistakeholder model (Figure) that we anticipate will be adaptable to other health systems.

In the process of establishing the SNF Collaborative at JHM, there were a number of early challenges faced and lessons learned:

• In a large integrated delivery system, there is a need to balance the benefits of central coordination with the support for ongoing local efforts to promote partner engagement at the hospital and SNF level. The forums created within the collaborative governance structure can facilitate sharing of the prior health system, hospital or SNF initiatives to grow upon successes and avoid prior pitfalls.
• Early identification of risk-adjusted PAC data sources is central to the collaborative establishment and goal setting. This requires assessment of internal analytic resources, budget, and desired timeline for implementation to determine the optimal arrangement. Similarly, identification of available data sources to drive the analytic efforts is essential and should include a health information exchange, claims, and MDS among others.
• Partnering with local QIOs provides support for facility-level quality improvement efforts. They have the staff and onsite expertise to facilitate process implementation within individual SNFs.
• Larger preferred provider networks require considerable administrative support to facilitate communication with the entities, coordinate completion of network agreements, and manage the dissemination of SNF- and hospital-specific performance data.
• Legal and contractual support related to data sharing and HIPAA compliance is needed due to the complexity of the health system and SNF legal structure. Multiple JHM legal entities were involved in this collaborative as were a mixture of freestanding SNFs and corporate chains. There was a significant effort required to execute both data-sharing agreements as well as charters to enable QIO participation.
• Physician leadership and insight are key to implementing meaningful and broad change. When devising system-wide solutions, incorporation and respect for local processes and needs are paramount for provider engagement and behavior change. This process will likely identify gaps in understanding the PAC patient’s experience and needs. It may also reveal practice variability and foster opportunities for provider education on the needs of PAC teams and how to best facilitate quality transitions.

We proposed a framework for establishing a collaborative partnership with a preferred network of SNF providers. Depending on organizational readiness, significant upfront investment of time and resources could be needed to establish a coordinated network of SNF providers. However, once established, such networks can be leveraged to support ongoing process improvement efforts within a hospital or delivery system and can be used strategically by such health systems as they implement value-based health strategies. Furthermore, the lessons learned from transitions to SNFs can be applied more broadly in the PAC landscape including transitions to home from both the hospital and SNF.

Acknowledgments

The authors wish to acknowledge all the members and participants in the Johns Hopkins Medicine Skilled Nursing Facility Collaborative and the executive sponsors and JHM hospital presidents for their support of this work.

Disclosures

Michele Bellantoni receives intramural salary support for being the medical director of the JHM SNF Collaborative. Damien Doyle is a part-time geriatrician at the Hebrew Home of Greater Washington, a skilled nursing facility. He received travel expense support for GAPNA, a local Advanced Practice Nurse Association meeting.The authors otherwise have no potential conflicts of interest to disclose.

Funding

The authors state that there were no external sponsors for this work.

Hospitals and health systems are under mounting financial pressure to shorten hospitalizations and reduce readmissions. These priorities have led to an ever-increasing focus on postacute care (PAC), and more specifically on improving transitions from the hospital.^1,2 According to a 2013 Institute of Medicine report, PAC is the source of 73% of the variation in Medicare spending³ and readmissions during the postacute episode nearly double the average Medicare payment.⁴ Within the PAC landscape, discharges to skilled nursing facilities (SNFs) have received particular focus due to the high rates of readmission and associated care costs.⁵

Hospitals, hospital physicians, PAC providers, and payers need to improve SNF transitions in care. Hospitals are increasingly responsible for patient care beyond their walls through several mechanisms including rehospitalization penalties, value-based reimbursement strategies (eg, bundled payments), and risk-based contracting on the total cost of care through relationships with accountable care organizations (ACOs) and Medicare Advantage plans. Similarly, hospital-employed physicians and PAC providers are more engaged in achieving value-based goals through increased alignment of provider compensation models^6,7 with risk-based contracting.

Current evidence suggests that rehospitalizations could be reduced by focusing on a concentrated referral network of preferred high-quality SNFs;^8,9 however, less is known about how to develop and operate such linkages at the administrative or clinical levels.⁸ In this article, we propose a collaborative framework for the establishment of a preferred PAC network.

One mechanism employed to improve transitions to SNFs and reduce associated readmissions is to create a preferred provider network. Increasing the concentration of hospital discharges to higher performing facilities is associated with lower rehospitalization rates, particularly during the critical days following discharge.¹⁰

While the criteria applied for preferred provider networks vary, there are several emerging themes.¹⁰ Quality metrics are often applied, generally starting with Centers for Medicare and Medicaid Services (CMS) quality star ratings and Long-Term Care Minimum Data Set (MDS) metrics with additional criteria frequently layered upon those. Some examples include the extent of physician coverage,¹¹ the extent of nursing coverage (eg, nursing ratios or 24/7 nursing care), geographic access, and flexible admission times (including weekends and nights).¹² In addition, several outcome measures may be used such as 30-day readmission rates, patient/family satisfaction ratings, ED visits, primary care follow-up within seven days of PAC discharge, or impact on the total cost of care.

Beyond the specified criteria, some hospitals choose to build upon existing relationships when developing their preferred network. By selecting historically high-volume facilities, they are able to leverage the existing name recognition amongst patients and providers.¹³ This minimizes retraining of discharge planners, maintains institutional relationships, and aligns with the patients’ geographic preferences.^2,13 While the high volume SNFs may not have the highest quality ratings, some hospitals find they can leverage the value of preferred partner status to push behavior change and improve performance.¹³

Here we propose a framework for the establishment of a preferred provider network for a hospital or health system based on the early experience of establishing an SNF Collaborative within Johns Hopkins Medicine (JHM). JHM is a large integrated health care system, which includes five hospitals within the region, including two large academic hospitals and three community hospitals serving patients in Maryland and the District of Columbia.¹⁴

JHM identified a need for improved coordination with PAC providers and saw opportunities to build upon successful individual hospital efforts to create a system-level approach with a PAC partnership sharing the goals of improving care and reducing costs. Additional opportunities exist given the unique Maryland all-payer Global Budget Revenue system managed by the Health Services Cost Review Commission. This system imposes hospital-level penalties for readmissions or poor quality measure performance and is moving to a new phase that will place hospitals directly at risk for the total Part A and Part B Medicare expenditures for a cohort of attributed Medicare patients, inclusive of their PAC expenses. This state-wide program is one example of a shift in payment structures from volume to value that is occurring throughout the healthcare sector.

Developing a formal collaboration inclusive of the five local hospitals, Johns Hopkins HealthCare (JHHC)—the managed care division of JHM—and the JHM ACO (Johns Hopkins Medicine Alliance for Patients, JMAP), we established a JHM SNF Collaborative. This group was tasked with improving the continuum of care for our patients discharged to PAC facilities. Given the number and diversity of entities involved, we sought to draw on efforts already managed and piloted locally, while disseminating best practices and providing added services at the collaborative level. We propose a collaborative multistakeholder model (Figure) that we anticipate will be adaptable to other health systems.

In the process of establishing the SNF Collaborative at JHM, there were a number of early challenges faced and lessons learned:

• In a large integrated delivery system, there is a need to balance the benefits of central coordination with the support for ongoing local efforts to promote partner engagement at the hospital and SNF level. The forums created within the collaborative governance structure can facilitate sharing of the prior health system, hospital or SNF initiatives to grow upon successes and avoid prior pitfalls.
• Early identification of risk-adjusted PAC data sources is central to the collaborative establishment and goal setting. This requires assessment of internal analytic resources, budget, and desired timeline for implementation to determine the optimal arrangement. Similarly, identification of available data sources to drive the analytic efforts is essential and should include a health information exchange, claims, and MDS among others.
• Partnering with local QIOs provides support for facility-level quality improvement efforts. They have the staff and onsite expertise to facilitate process implementation within individual SNFs.
• Larger preferred provider networks require considerable administrative support to facilitate communication with the entities, coordinate completion of network agreements, and manage the dissemination of SNF- and hospital-specific performance data.
• Legal and contractual support related to data sharing and HIPAA compliance is needed due to the complexity of the health system and SNF legal structure. Multiple JHM legal entities were involved in this collaborative as were a mixture of freestanding SNFs and corporate chains. There was a significant effort required to execute both data-sharing agreements as well as charters to enable QIO participation.
• Physician leadership and insight are key to implementing meaningful and broad change. When devising system-wide solutions, incorporation and respect for local processes and needs are paramount for provider engagement and behavior change. This process will likely identify gaps in understanding the PAC patient’s experience and needs. It may also reveal practice variability and foster opportunities for provider education on the needs of PAC teams and how to best facilitate quality transitions.

We proposed a framework for establishing a collaborative partnership with a preferred network of SNF providers. Depending on organizational readiness, significant upfront investment of time and resources could be needed to establish a coordinated network of SNF providers. However, once established, such networks can be leveraged to support ongoing process improvement efforts within a hospital or delivery system and can be used strategically by such health systems as they implement value-based health strategies. Furthermore, the lessons learned from transitions to SNFs can be applied more broadly in the PAC landscape including transitions to home from both the hospital and SNF.

Acknowledgments

The authors wish to acknowledge all the members and participants in the Johns Hopkins Medicine Skilled Nursing Facility Collaborative and the executive sponsors and JHM hospital presidents for their support of this work.

Disclosures

Michele Bellantoni receives intramural salary support for being the medical director of the JHM SNF Collaborative. Damien Doyle is a part-time geriatrician at the Hebrew Home of Greater Washington, a skilled nursing facility. He received travel expense support for GAPNA, a local Advanced Practice Nurse Association meeting.The authors otherwise have no potential conflicts of interest to disclose.

Funding

The authors state that there were no external sponsors for this work.

Ensuring the delivery of safe and cost-effective care is the core mission of hospitals,¹ but nearly 90% of unplanned patient transfers to critical care may be the result of a new or worsening condition.² The cost of treatment of sepsis, respiratory failure, and arrest, which are among the deadliest conditions for hospitalized patients,^3,4 are estimated to be $30.7 billion annually (8.1% of national hospital costs).⁵ As many as 44% of adverse events may be avoidable,⁶ and concerns about patient safety have motivated hospitals and health systems to find solutions to identify and treat deteriorating patients expeditiously. Evidence suggests that many hospitalized patients presenting with rapid decline showed warning signs 24-48 hours before the event.⁷ Therefore, ample time may be available for early identification and intervention in many patients.

As early as 1997, hospitals have used early warning systems (EWSs) to identify at-risk patients and proactively inform clinicians.⁸ EWSs can predict a proportion of patients who are at risk for clinical deterioration (this benefit is measured with sensitivity) with the tradeoff that some alerts are false (as measured with positive predictive value [PPV] or its inverse, workup-to-detection ratio [WDR]^9-11). Historically, EWS tools were paper-based instruments designed for fast manual calculation by hospital staff. Many aggregate-weighted EWS instruments continue to be used for research and practice, including the Modified Early Warning Systems (MEWS)¹² and National Early Warning System (NEWS).^13,14 Aggregate-weighted EWSs lack predictive precision because they use simple addition of a few clinical parameter scores, including vital signs and level of consciousness.¹⁵ Recently, a new category has emerged, which use multivariable regression or machine learning; we refer to this category as “EWSs using statistical modeling”. This type of EWS uses more computationally intensive risk stratification methods to predict risk¹⁶ by adjusting for a larger set of clinical covariates, thereby reducing the degree of unexplained variance. Although these EWSs are thought to be more precise and to generate fewer false positive alarms compared with others,^14,17-19 no review to date has systematically synthesized and compared their performance against aggregate-weighted EWSs.

Purpose

The purpose of this systematic review was to evaluate the recent literature regarding prognostic test accuracy and clinical workloads generated by EWSs using statistical modeling versus aggregate-weighted systems.

Search Strategy

Adhering to PRISMA protocol guidelines for systematic reviews, we searched the peer-reviewed literature in PubMed and CINAHL Plus, as well as conference proceedings and online repositories of patient safety organizations published between January 1, 2012 and September 15, 2018. We selected this timeframe because EWSs using statistical modeling are relatively new approaches compared with the body of evidence concerning aggregate-weighted EWSs. An expert PhD researcher confirmed the search results in a blinded independent query.

Inclusion and Exclusion Criteria

We included peer-reviewed articles reporting the area under the receiver operator curve (AUC),²⁰ or the equivalent c-statistic, of models predicting clinical deterioration (measured as the composite of transfer to intensive care unit (ICU) and/or mortality) among adult patients in general hospital wards. We excluded studies if they did not compare an EWS using statistical modeling with an aggregate-weighted EWS, did not report AUC, or only reported on an aggregate-weighted EWS. Excluded settings were pediatrics, obstetrics, emergency departments, ICUs, transitional care units, and oncology. We also excluded studies with samples limited to physiological monitoring, sepsis, or postsurgical subpopulations.

Data Abstraction

Following the TRIPOD guidelines for the reporting of predictive models,²¹ and the PRISMA and Cochrane Collaboration guidelines for systematic reviews,^22-24 we extracted study characteristics (Table 1), sample demographics (Appendix Table 4), model characteristics and performance (Appendix Table 5), and level of scientific evidence and risk of bias (Appendix Table 6). To address the potential for overfitting, we selected model performance results of the validation dataset rather than the derivation dataset, if reported. If studies reported multiple models in either EWS category, we selected the best-performing model for comparison.

Measures of Model Performance

Because predictive models can achieve good case identification at the expense of high clinical workloads, an assessment of model performance would be incomplete without measures of clinical utility. For clinicians, this aspect can be measured as the model’s PPV (the percentage of true positive alerts among all alerts), or more intelligibly, as the WDR, which equals 1/PPV. WDR indicates the number of patients requiring evaluation to identify and treat one true positive case.^9-11 It is known that differences in event rates (prevalence or pretest probability) influence a model’s PPV²⁵ and its reciprocal WDR. However, for systematic comparison, PPV and WDR can be standardized using a fixed representative event rate across studies.^24,26 We abstracted the reported PPV and WDR, and computed standardized PPV and WDR for an event rate of 4%.

Other measures included the area under the receiver operator curve (AUC),²⁰ sensitivity, and specificity. AUC plots a model’s false positive rate (x-axis) against its true positive rate (y-axis), with an ideal scenario of very high y-values and very low x-values.²⁷ Sensitivity (the model’s ability to detect a true positive case among all cases) and specificity (the model’s ability to detect a true noncase among all noncases²⁸) are influenced by chosen alert thresholds. It is incorrect to assume that a given model produces only one sensitivity/specificity result; for systematic comparison, we therefore selected results in the 50% sensitivity range, and separately, in the 92% specificity range for EWSs using statistical modeling. Then, we simulated a fixed sensitivity of 0.51 and assumed specificity of 0.87 in aggregate-weighted EWSs.

Search Results

The PubMed search for “early warning score OR early warning system AND deterioration OR predict transfer ICU” returned 285 peer-reviewed articles. A search on CINAHL Plus using the same filters and query terms returned 219 articles with no additional matches (Figure 1). Of the 285 articles, we excluded 269 during the abstract screen and 10 additional articles during full-text review (Figure 1). A final review of the reference lists of the six selected studies did not yield additional articles.

Study Characteristics

There were several similarities across the selected studies (Table 1). All occurred in the United States; all compared their model’s performance against at least one aggregate-weighted EWS model;^14,17-19,29 and all used retrospective cohort designs. Of the six studies, one took place in a single hospital;²⁹ three pooled data from five hospitals;^17,18,30 and two occurred in a large integrated healthcare delivery system using data from 14 and, subsequently, 21 hospitals.^14,19 The largest study¹⁴ included nearly 650,000 admissions, while the smallest study²⁹ reported slightly less than 7,500 admissions. Of the six studies, four used multivariable regression,^14,17,19,29 and two used machine learning techniques for outcome prediction.^18,30

Outcome Variables

The primary outcome for inclusion in this review was clinical deterioration measured by the composite of transfer to ICU and some measure of mortality. Churpek et al.^10,11 and Green et al.³⁰ also included cardiac arrest, and Alvarez et al.²² included respiratory compromise in their outcome composite.

Researchers used varying definitions of mortality, including “death outside the ICU in a patient whose care directive was full code;”^14,19 “death on the wards without attempted resuscitation;”^17,18 “an in-hospital death in patients without a DNR order at admission that occurred on the medical ward or in ICU within 24 hours after transfer;”²⁹ or “death within 24 hours.”³⁰

Predictor Variables

We observed a broad assortment of predictor variables. All models included vital signs (heart rate, respiratory rate, blood pressure, and venous oxygen saturation); mental state; laboratory data; age; and sex. Additional variables included comorbidity, shock index,³¹ severity of illness score, length of stay, event time of day, season, admission category, and length of stay,^14,19 among others.

Model Performance

Reported PPV ranged from 0.16 to 0.42 (mean = 0.27) in EWSs using statistical modeling and 0.15 to 0.28 (mean = 0.19) in aggregate-weighted EWS models. The weighted mean standardized PPV, adjusted for an event rate of 4% across studies (Table 2), was 0.21 in EWSs using statistical modeling versus 0.14 in aggregate-weighted EWS models (simulated at 0.51 sensitivity and 0.87 specificity).

Only two studies^14,19 reported the WDR metric (alerts generated to identify one true positive case) explicitly. Based on the above PPV results, EWSs using statistical modeling generated a standardized WDR of 4.9 in models using statistical modeling versus 7.1 in aggregate-weighted models (Figure 2). The delta of 2.2 evaluations to find and treat one true positive case equals a 45% relative increase in RRT evaluation workloads using aggregate-weighted EWSs.

AUC values ranged from 0.77 to 0.85 (weighted mean = 0.80) in EWSs using statistical modeling, indicating good model discrimination. AUCs of aggregate-weighted EWSs ranged from 0.70 to 0.76 (weighted mean = 0.73), indicating fair model discrimination (Figure 2). The overall AUC delta was 0.07. However, our estimates may possibly be favoring EWSs that use statistical modeling by virtue of their derivation in an original research population compared with aggregate-weighted EWSs that were derived externally. For example, sensitivity analysis of eCART,¹⁸ an EWS using machine learning, showed an AUC drop of 1% in a large external patient population,¹⁴ while NEWS AUCs¹³ dropped between 11% and 15% in two large external populations (Appendix Table 7).^14,30 For hospitals adopting an externally developed EWS using statistical modeling, these results suggest that an AUC delta of approximately 5% can be expected and 7% for an internally developed EWS.

Risk of Bias Assessment

We scored the studies by adapting the Cochrane Collaboration tool for assessing risk of bias ³² (Appendix Table 5). Of the six studies, five received total scores between 1.0 and 2.0 (indicating relatively low bias risk), and one study had a score of 3.5 (indicating higher bias risk). Low bias studies^14,17-19,30 used large samples across multiple hospitals, discussed the choice of predictor variables and outcomes more precisely, and reported their measurement approaches and analytic methods in more detail, including imputation of missing data and model calibration.

In this systematic review, we assessed the predictive ability of EWSs using statistical modeling versus aggregate-weighted EWS models to detect clinical deterioration risk in hospitalized adults in general wards. From 2007 to 2018, at least five systematic reviews examined aggregate-weighted EWSs in adult inpatient settings.^33-37 No systematic review, however, has synthesized the evidence of EWSs using statistical modeling.

The recent evidence is limited to six studies, of which five had favorable risk of bias scores. All studies included in this review demonstrated superior model performance of the EWSs using statistical modeling compared with an aggregate-weighted EWS, and at least five of the six studies employed rigor in design, measurement, and analytic method. The AUC absolute difference between EWSs using statistical modeling and aggregate-weighted EWSs was 7% overall, moving model performance from fair to good (Table 2; Figure 2). Although this increase in discriminative power may appear modest, it translates into avoiding a 45% increase in WDR workload generated by an aggregate-weighted EWS, approximately two patient evaluations for each true positive case.

Results of our review suggest that EWSs using statistical modeling predict clinical deterioration risk with better precision. This is an important finding for the following reasons: (1) Better risk prediction can support the activation of rescue; (2) Given federal mandates to curb spending, the elimination of some resource-intensive false positive evaluations supports high-value care;³⁸ and (3) The Quadruple Aim³⁹ accounts for clinician wellbeing. EWSs using statistical modeling may offer benefits in terms of clinician satisfaction with the human–system interface because better discrimination reduces the daily evaluation workload/cognitive burden and because the reduction of false positive alerts may reduce alert fatigue.^40,41

Still, an important issue with risk detection is that it is unknown which percentage of patients are uniquely identified by an EWS and not already under evaluation by the clinical team. For example, a recent study by Bedoya et al.⁴² found that using NEWS did not improve clinical outcomes and nurses frequently disregarded the alert. Another study⁴³ found that the combined clinical judgment of physicians and nurses had an AUC of 0.90 in predicting mortality. These results suggest that at certain times, an EWS alert may not add new useful information for clinicians even when it correctly identifies deterioration risk. It remains difficult to define exactly how many patients an EWS would have to uniquely identify to have clinical utility.

Even EWSs that use statistical modeling cannot detect all true deterioration cases perfectly, and they may at times trigger an alert only when the clinical team is already aware of a patient’s clinical decline. Consequently, EWSs using statistical modeling can at best augment and support—but not replace—RRT rounding, physician workup, and vigilant frontline staff. However, clinicians, too, are not perfect, and the failure-to-rescue literature suggests that certain human factors are antecedents to patient crises (eg, stress and distraction,^44-46 judging by precedent/experience,^44,47 and innate limitations of human cognition⁴⁷). Because neither clinicians nor EWSs can predict deterioration perfectly, the best possible rescue response combines clinical vigilance, RRT rounding, and EWSs using statistical modeling as complementary solutions.

Our findings suggest that predictive models cannot be judged purely on AUC (in fact, it would be ill-advised) but also by their clinical utility (expressed in WDR and PPV): How many patients does a clinician need to evaluate?^9-11 Precision is not meaningful if it comes at the expense of unmanageable evaluation workloads, and our findings suggest that clinicians should evaluate models based on their clinical utility. Hospitals considering adoption of an EWS using statistical modeling should consider that externally developed EWSs appear to experience a performance drop when applied to a new patient population; a slightly higher WDR and slightly lower AUC can be expected. EWSs using statistical modeling appear to perform best when tailored to the targeted patient population (or are derived in-house). Model depreciation over time will likely require recalibration. In addition, adoption of a machine learning algorithm may mean that original model results are obscured by the black box output of the algorithm.^48-50

Findings from this systematic review are subject to several limitations. First, we applied strict inclusion criteria, which led us to exclude studies that offered findings in specialty units and specific patient subpopulations, among others. In the interest of systematic comparison, our findings are limited to general wards. We also restricted our search to recent studies that reported on models predicting clinical deterioration, which we defined as the composite of ICU transfer and/or death. Clinically, deteriorating patients in general wards either die or are transferred to ICU. This criterion resulted in exclusion of the Rothman Index,⁵¹ which predicts “death within 24 hours” but not ICU transfer. The AUC in this study was higher than those selected in this review (0.93 compared to 0.82 for MEWS; AUC delta: 0.09). The higher AUC may be a function of the outcome definition (30-day mortality would be more challenging to predict). Therefore, hospitals or health systems interested in purchasing an EWS using statistical modeling should carefully consider the outcome selection and definition.

Second, as is true for systematic reviews in general,⁵² the degree of clinical and methodological heterogeneity across the selected studies may limit our findings. Studies occurred in various settings (university hospital, teaching hospitals, and community hospitals), which may serve diverging patient populations. We observed that studies in university-based settings had a higher event rate ranging from 5.6% to 7.8%, which may result in higher PPV results in these settings. However, this increase would apply to both EWS types equally. To arrive at a “true” reflection of model performance, the simulations for PPV and WDR have used a more conservative event rate of 4%. We observed heterogenous mortality definitions, which did not always account for the reality that a patient’s death may be an appropriate outcome (ie, it was concordant with treatment wishes in the context of severe illness or an end-of-life trajectory). Studies also used different sampling procedures; some allowed multiple observations although most did not. The variation in sampling may change PPV and limit our systematic comparison. However, regardless of methodological differences, our review suggests that EWSs using statistical modeling perform better than aggregate-weighted EWSs in each of the selected studies.

Third, systematic reviews may be subject to the issue of publication bias because they can only compare published results and could possibly omit an unknown number of unpublished studies. However, the selected studies uniformly demonstrated similar model improvements, which are plausibly related to the larger number of covariates, statistical methods, and shrinkage of random error.

Finally, this review was limited to the comparison of observational studies, which aimed to answer how the two EWS classes compared. These studies did not address whether an alert had an impact on clinical care and patient outcomes. Results from at least one randomized nonblinded controlled trial suggest that alert-driven RRT activation may reduce the length of stay by 24 hours and use of oximetry, but has no impact on mortality, ICU transfer, and ICU length of stay.⁵³

Our findings point to three areas of need for the field of predictive EWS research: (1) a standardized set of clinical deterioration outcome measures, (2) a standardized set of measures capturing clinical evaluation workload and alert frequency, and (3) cost estimates of clinical workloads with and without deployment of an EWS using statistical modeling. Given the present divergence of outcome definitions, EWS research may benefit from a common “clinical deterioration” outcome standard, including transfer to ICU, inpatient/30-day/90-day mortality, and death with DNR, comfort care, or hospice. The field is lacking a standardized clinical workload measure and an understanding of the net percentage of patients uniquely identified by an EWS.

By using predictive analytics, health systems may be better able to achieve the goals of high-value care and patient safety and support the Quadruple Aim. Still, gaps in knowledge exist regarding the measurement of the clinical processes triggered by EWSs, evaluation workloads, alert fatigue, clinician burnout associated with the human-alert interface, and costs versus benefits. Future research should evaluate the degree to which EWSs can identify risk among patients who are not already under evaluation by the clinical team, assess the balanced treatment effects of RRT interventions between decedents and survivors, and investigate clinical process times relative to the time of an EWS alert using statistical modeling.

Acknowledgments

The authors would like to thank Ms. Jill Pope at the Kaiser Permanente Center for Health Research in Portland, OR for her assistance with manuscript preparation. Daniel Linnen would like to thank Dr. Linda Franck, PhD, RN, FAAN, Professor at the University of California, San Francisco, School of Nursing for reviewing the manuscript.

Disclosures

The authors declare no conflicts of interest.

Funding

The Maribelle & Stephen Leavitt Scholarship, the Jonas Nurse Scholars Scholarship at the University of California, San Francisco, and the Nurse Scholars Academy Predoctoral Research Fellowship at Kaiser Permanente Northern California supported this study during Daniel Linnen’s doctoral training at the University of California, San Francisco. Dr. Vincent Liu was funded by National Institute of General Medical Sciences Grant K23GM112018.

Ensuring the delivery of safe and cost-effective care is the core mission of hospitals,¹ but nearly 90% of unplanned patient transfers to critical care may be the result of a new or worsening condition.² The cost of treatment of sepsis, respiratory failure, and arrest, which are among the deadliest conditions for hospitalized patients,^3,4 are estimated to be $30.7 billion annually (8.1% of national hospital costs).⁵ As many as 44% of adverse events may be avoidable,⁶ and concerns about patient safety have motivated hospitals and health systems to find solutions to identify and treat deteriorating patients expeditiously. Evidence suggests that many hospitalized patients presenting with rapid decline showed warning signs 24-48 hours before the event.⁷ Therefore, ample time may be available for early identification and intervention in many patients.

As early as 1997, hospitals have used early warning systems (EWSs) to identify at-risk patients and proactively inform clinicians.⁸ EWSs can predict a proportion of patients who are at risk for clinical deterioration (this benefit is measured with sensitivity) with the tradeoff that some alerts are false (as measured with positive predictive value [PPV] or its inverse, workup-to-detection ratio [WDR]^9-11). Historically, EWS tools were paper-based instruments designed for fast manual calculation by hospital staff. Many aggregate-weighted EWS instruments continue to be used for research and practice, including the Modified Early Warning Systems (MEWS)¹² and National Early Warning System (NEWS).^13,14 Aggregate-weighted EWSs lack predictive precision because they use simple addition of a few clinical parameter scores, including vital signs and level of consciousness.¹⁵ Recently, a new category has emerged, which use multivariable regression or machine learning; we refer to this category as “EWSs using statistical modeling”. This type of EWS uses more computationally intensive risk stratification methods to predict risk¹⁶ by adjusting for a larger set of clinical covariates, thereby reducing the degree of unexplained variance. Although these EWSs are thought to be more precise and to generate fewer false positive alarms compared with others,^14,17-19 no review to date has systematically synthesized and compared their performance against aggregate-weighted EWSs.

Purpose

The purpose of this systematic review was to evaluate the recent literature regarding prognostic test accuracy and clinical workloads generated by EWSs using statistical modeling versus aggregate-weighted systems.

Search Strategy

Adhering to PRISMA protocol guidelines for systematic reviews, we searched the peer-reviewed literature in PubMed and CINAHL Plus, as well as conference proceedings and online repositories of patient safety organizations published between January 1, 2012 and September 15, 2018. We selected this timeframe because EWSs using statistical modeling are relatively new approaches compared with the body of evidence concerning aggregate-weighted EWSs. An expert PhD researcher confirmed the search results in a blinded independent query.

Inclusion and Exclusion Criteria

We included peer-reviewed articles reporting the area under the receiver operator curve (AUC),²⁰ or the equivalent c-statistic, of models predicting clinical deterioration (measured as the composite of transfer to intensive care unit (ICU) and/or mortality) among adult patients in general hospital wards. We excluded studies if they did not compare an EWS using statistical modeling with an aggregate-weighted EWS, did not report AUC, or only reported on an aggregate-weighted EWS. Excluded settings were pediatrics, obstetrics, emergency departments, ICUs, transitional care units, and oncology. We also excluded studies with samples limited to physiological monitoring, sepsis, or postsurgical subpopulations.

Data Abstraction

Following the TRIPOD guidelines for the reporting of predictive models,²¹ and the PRISMA and Cochrane Collaboration guidelines for systematic reviews,^22-24 we extracted study characteristics (Table 1), sample demographics (Appendix Table 4), model characteristics and performance (Appendix Table 5), and level of scientific evidence and risk of bias (Appendix Table 6). To address the potential for overfitting, we selected model performance results of the validation dataset rather than the derivation dataset, if reported. If studies reported multiple models in either EWS category, we selected the best-performing model for comparison.

Measures of Model Performance

Because predictive models can achieve good case identification at the expense of high clinical workloads, an assessment of model performance would be incomplete without measures of clinical utility. For clinicians, this aspect can be measured as the model’s PPV (the percentage of true positive alerts among all alerts), or more intelligibly, as the WDR, which equals 1/PPV. WDR indicates the number of patients requiring evaluation to identify and treat one true positive case.^9-11 It is known that differences in event rates (prevalence or pretest probability) influence a model’s PPV²⁵ and its reciprocal WDR. However, for systematic comparison, PPV and WDR can be standardized using a fixed representative event rate across studies.^24,26 We abstracted the reported PPV and WDR, and computed standardized PPV and WDR for an event rate of 4%.

Other measures included the area under the receiver operator curve (AUC),²⁰ sensitivity, and specificity. AUC plots a model’s false positive rate (x-axis) against its true positive rate (y-axis), with an ideal scenario of very high y-values and very low x-values.²⁷ Sensitivity (the model’s ability to detect a true positive case among all cases) and specificity (the model’s ability to detect a true noncase among all noncases²⁸) are influenced by chosen alert thresholds. It is incorrect to assume that a given model produces only one sensitivity/specificity result; for systematic comparison, we therefore selected results in the 50% sensitivity range, and separately, in the 92% specificity range for EWSs using statistical modeling. Then, we simulated a fixed sensitivity of 0.51 and assumed specificity of 0.87 in aggregate-weighted EWSs.

Search Results

The PubMed search for “early warning score OR early warning system AND deterioration OR predict transfer ICU” returned 285 peer-reviewed articles. A search on CINAHL Plus using the same filters and query terms returned 219 articles with no additional matches (Figure 1). Of the 285 articles, we excluded 269 during the abstract screen and 10 additional articles during full-text review (Figure 1). A final review of the reference lists of the six selected studies did not yield additional articles.

Study Characteristics

There were several similarities across the selected studies (Table 1). All occurred in the United States; all compared their model’s performance against at least one aggregate-weighted EWS model;^14,17-19,29 and all used retrospective cohort designs. Of the six studies, one took place in a single hospital;²⁹ three pooled data from five hospitals;^17,18,30 and two occurred in a large integrated healthcare delivery system using data from 14 and, subsequently, 21 hospitals.^14,19 The largest study¹⁴ included nearly 650,000 admissions, while the smallest study²⁹ reported slightly less than 7,500 admissions. Of the six studies, four used multivariable regression,^14,17,19,29 and two used machine learning techniques for outcome prediction.^18,30

Outcome Variables

The primary outcome for inclusion in this review was clinical deterioration measured by the composite of transfer to ICU and some measure of mortality. Churpek et al.^10,11 and Green et al.³⁰ also included cardiac arrest, and Alvarez et al.²² included respiratory compromise in their outcome composite.

Researchers used varying definitions of mortality, including “death outside the ICU in a patient whose care directive was full code;”^14,19 “death on the wards without attempted resuscitation;”^17,18 “an in-hospital death in patients without a DNR order at admission that occurred on the medical ward or in ICU within 24 hours after transfer;”²⁹ or “death within 24 hours.”³⁰

Predictor Variables

We observed a broad assortment of predictor variables. All models included vital signs (heart rate, respiratory rate, blood pressure, and venous oxygen saturation); mental state; laboratory data; age; and sex. Additional variables included comorbidity, shock index,³¹ severity of illness score, length of stay, event time of day, season, admission category, and length of stay,^14,19 among others.

Model Performance

Reported PPV ranged from 0.16 to 0.42 (mean = 0.27) in EWSs using statistical modeling and 0.15 to 0.28 (mean = 0.19) in aggregate-weighted EWS models. The weighted mean standardized PPV, adjusted for an event rate of 4% across studies (Table 2), was 0.21 in EWSs using statistical modeling versus 0.14 in aggregate-weighted EWS models (simulated at 0.51 sensitivity and 0.87 specificity).

Only two studies^14,19 reported the WDR metric (alerts generated to identify one true positive case) explicitly. Based on the above PPV results, EWSs using statistical modeling generated a standardized WDR of 4.9 in models using statistical modeling versus 7.1 in aggregate-weighted models (Figure 2). The delta of 2.2 evaluations to find and treat one true positive case equals a 45% relative increase in RRT evaluation workloads using aggregate-weighted EWSs.

AUC values ranged from 0.77 to 0.85 (weighted mean = 0.80) in EWSs using statistical modeling, indicating good model discrimination. AUCs of aggregate-weighted EWSs ranged from 0.70 to 0.76 (weighted mean = 0.73), indicating fair model discrimination (Figure 2). The overall AUC delta was 0.07. However, our estimates may possibly be favoring EWSs that use statistical modeling by virtue of their derivation in an original research population compared with aggregate-weighted EWSs that were derived externally. For example, sensitivity analysis of eCART,¹⁸ an EWS using machine learning, showed an AUC drop of 1% in a large external patient population,¹⁴ while NEWS AUCs¹³ dropped between 11% and 15% in two large external populations (Appendix Table 7).^14,30 For hospitals adopting an externally developed EWS using statistical modeling, these results suggest that an AUC delta of approximately 5% can be expected and 7% for an internally developed EWS.

Risk of Bias Assessment

We scored the studies by adapting the Cochrane Collaboration tool for assessing risk of bias ³² (Appendix Table 5). Of the six studies, five received total scores between 1.0 and 2.0 (indicating relatively low bias risk), and one study had a score of 3.5 (indicating higher bias risk). Low bias studies^14,17-19,30 used large samples across multiple hospitals, discussed the choice of predictor variables and outcomes more precisely, and reported their measurement approaches and analytic methods in more detail, including imputation of missing data and model calibration.

In this systematic review, we assessed the predictive ability of EWSs using statistical modeling versus aggregate-weighted EWS models to detect clinical deterioration risk in hospitalized adults in general wards. From 2007 to 2018, at least five systematic reviews examined aggregate-weighted EWSs in adult inpatient settings.^33-37 No systematic review, however, has synthesized the evidence of EWSs using statistical modeling.

The recent evidence is limited to six studies, of which five had favorable risk of bias scores. All studies included in this review demonstrated superior model performance of the EWSs using statistical modeling compared with an aggregate-weighted EWS, and at least five of the six studies employed rigor in design, measurement, and analytic method. The AUC absolute difference between EWSs using statistical modeling and aggregate-weighted EWSs was 7% overall, moving model performance from fair to good (Table 2; Figure 2). Although this increase in discriminative power may appear modest, it translates into avoiding a 45% increase in WDR workload generated by an aggregate-weighted EWS, approximately two patient evaluations for each true positive case.

Results of our review suggest that EWSs using statistical modeling predict clinical deterioration risk with better precision. This is an important finding for the following reasons: (1) Better risk prediction can support the activation of rescue; (2) Given federal mandates to curb spending, the elimination of some resource-intensive false positive evaluations supports high-value care;³⁸ and (3) The Quadruple Aim³⁹ accounts for clinician wellbeing. EWSs using statistical modeling may offer benefits in terms of clinician satisfaction with the human–system interface because better discrimination reduces the daily evaluation workload/cognitive burden and because the reduction of false positive alerts may reduce alert fatigue.^40,41

Still, an important issue with risk detection is that it is unknown which percentage of patients are uniquely identified by an EWS and not already under evaluation by the clinical team. For example, a recent study by Bedoya et al.⁴² found that using NEWS did not improve clinical outcomes and nurses frequently disregarded the alert. Another study⁴³ found that the combined clinical judgment of physicians and nurses had an AUC of 0.90 in predicting mortality. These results suggest that at certain times, an EWS alert may not add new useful information for clinicians even when it correctly identifies deterioration risk. It remains difficult to define exactly how many patients an EWS would have to uniquely identify to have clinical utility.

Even EWSs that use statistical modeling cannot detect all true deterioration cases perfectly, and they may at times trigger an alert only when the clinical team is already aware of a patient’s clinical decline. Consequently, EWSs using statistical modeling can at best augment and support—but not replace—RRT rounding, physician workup, and vigilant frontline staff. However, clinicians, too, are not perfect, and the failure-to-rescue literature suggests that certain human factors are antecedents to patient crises (eg, stress and distraction,^44-46 judging by precedent/experience,^44,47 and innate limitations of human cognition⁴⁷). Because neither clinicians nor EWSs can predict deterioration perfectly, the best possible rescue response combines clinical vigilance, RRT rounding, and EWSs using statistical modeling as complementary solutions.

Our findings suggest that predictive models cannot be judged purely on AUC (in fact, it would be ill-advised) but also by their clinical utility (expressed in WDR and PPV): How many patients does a clinician need to evaluate?^9-11 Precision is not meaningful if it comes at the expense of unmanageable evaluation workloads, and our findings suggest that clinicians should evaluate models based on their clinical utility. Hospitals considering adoption of an EWS using statistical modeling should consider that externally developed EWSs appear to experience a performance drop when applied to a new patient population; a slightly higher WDR and slightly lower AUC can be expected. EWSs using statistical modeling appear to perform best when tailored to the targeted patient population (or are derived in-house). Model depreciation over time will likely require recalibration. In addition, adoption of a machine learning algorithm may mean that original model results are obscured by the black box output of the algorithm.^48-50

Findings from this systematic review are subject to several limitations. First, we applied strict inclusion criteria, which led us to exclude studies that offered findings in specialty units and specific patient subpopulations, among others. In the interest of systematic comparison, our findings are limited to general wards. We also restricted our search to recent studies that reported on models predicting clinical deterioration, which we defined as the composite of ICU transfer and/or death. Clinically, deteriorating patients in general wards either die or are transferred to ICU. This criterion resulted in exclusion of the Rothman Index,⁵¹ which predicts “death within 24 hours” but not ICU transfer. The AUC in this study was higher than those selected in this review (0.93 compared to 0.82 for MEWS; AUC delta: 0.09). The higher AUC may be a function of the outcome definition (30-day mortality would be more challenging to predict). Therefore, hospitals or health systems interested in purchasing an EWS using statistical modeling should carefully consider the outcome selection and definition.

Second, as is true for systematic reviews in general,⁵² the degree of clinical and methodological heterogeneity across the selected studies may limit our findings. Studies occurred in various settings (university hospital, teaching hospitals, and community hospitals), which may serve diverging patient populations. We observed that studies in university-based settings had a higher event rate ranging from 5.6% to 7.8%, which may result in higher PPV results in these settings. However, this increase would apply to both EWS types equally. To arrive at a “true” reflection of model performance, the simulations for PPV and WDR have used a more conservative event rate of 4%. We observed heterogenous mortality definitions, which did not always account for the reality that a patient’s death may be an appropriate outcome (ie, it was concordant with treatment wishes in the context of severe illness or an end-of-life trajectory). Studies also used different sampling procedures; some allowed multiple observations although most did not. The variation in sampling may change PPV and limit our systematic comparison. However, regardless of methodological differences, our review suggests that EWSs using statistical modeling perform better than aggregate-weighted EWSs in each of the selected studies.

Third, systematic reviews may be subject to the issue of publication bias because they can only compare published results and could possibly omit an unknown number of unpublished studies. However, the selected studies uniformly demonstrated similar model improvements, which are plausibly related to the larger number of covariates, statistical methods, and shrinkage of random error.

Finally, this review was limited to the comparison of observational studies, which aimed to answer how the two EWS classes compared. These studies did not address whether an alert had an impact on clinical care and patient outcomes. Results from at least one randomized nonblinded controlled trial suggest that alert-driven RRT activation may reduce the length of stay by 24 hours and use of oximetry, but has no impact on mortality, ICU transfer, and ICU length of stay.⁵³

Our findings point to three areas of need for the field of predictive EWS research: (1) a standardized set of clinical deterioration outcome measures, (2) a standardized set of measures capturing clinical evaluation workload and alert frequency, and (3) cost estimates of clinical workloads with and without deployment of an EWS using statistical modeling. Given the present divergence of outcome definitions, EWS research may benefit from a common “clinical deterioration” outcome standard, including transfer to ICU, inpatient/30-day/90-day mortality, and death with DNR, comfort care, or hospice. The field is lacking a standardized clinical workload measure and an understanding of the net percentage of patients uniquely identified by an EWS.

By using predictive analytics, health systems may be better able to achieve the goals of high-value care and patient safety and support the Quadruple Aim. Still, gaps in knowledge exist regarding the measurement of the clinical processes triggered by EWSs, evaluation workloads, alert fatigue, clinician burnout associated with the human-alert interface, and costs versus benefits. Future research should evaluate the degree to which EWSs can identify risk among patients who are not already under evaluation by the clinical team, assess the balanced treatment effects of RRT interventions between decedents and survivors, and investigate clinical process times relative to the time of an EWS alert using statistical modeling.

Acknowledgments

The authors would like to thank Ms. Jill Pope at the Kaiser Permanente Center for Health Research in Portland, OR for her assistance with manuscript preparation. Daniel Linnen would like to thank Dr. Linda Franck, PhD, RN, FAAN, Professor at the University of California, San Francisco, School of Nursing for reviewing the manuscript.

Disclosures

The authors declare no conflicts of interest.

Funding

The Maribelle & Stephen Leavitt Scholarship, the Jonas Nurse Scholars Scholarship at the University of California, San Francisco, and the Nurse Scholars Academy Predoctoral Research Fellowship at Kaiser Permanente Northern California supported this study during Daniel Linnen’s doctoral training at the University of California, San Francisco. Dr. Vincent Liu was funded by National Institute of General Medical Sciences Grant K23GM112018.

Ensuring the delivery of safe and cost-effective care is the core mission of hospitals,¹ but nearly 90% of unplanned patient transfers to critical care may be the result of a new or worsening condition.² The cost of treatment of sepsis, respiratory failure, and arrest, which are among the deadliest conditions for hospitalized patients,^3,4 are estimated to be $30.7 billion annually (8.1% of national hospital costs).⁵ As many as 44% of adverse events may be avoidable,⁶ and concerns about patient safety have motivated hospitals and health systems to find solutions to identify and treat deteriorating patients expeditiously. Evidence suggests that many hospitalized patients presenting with rapid decline showed warning signs 24-48 hours before the event.⁷ Therefore, ample time may be available for early identification and intervention in many patients.

As early as 1997, hospitals have used early warning systems (EWSs) to identify at-risk patients and proactively inform clinicians.⁸ EWSs can predict a proportion of patients who are at risk for clinical deterioration (this benefit is measured with sensitivity) with the tradeoff that some alerts are false (as measured with positive predictive value [PPV] or its inverse, workup-to-detection ratio [WDR]^9-11). Historically, EWS tools were paper-based instruments designed for fast manual calculation by hospital staff. Many aggregate-weighted EWS instruments continue to be used for research and practice, including the Modified Early Warning Systems (MEWS)¹² and National Early Warning System (NEWS).^13,14 Aggregate-weighted EWSs lack predictive precision because they use simple addition of a few clinical parameter scores, including vital signs and level of consciousness.¹⁵ Recently, a new category has emerged, which use multivariable regression or machine learning; we refer to this category as “EWSs using statistical modeling”. This type of EWS uses more computationally intensive risk stratification methods to predict risk¹⁶ by adjusting for a larger set of clinical covariates, thereby reducing the degree of unexplained variance. Although these EWSs are thought to be more precise and to generate fewer false positive alarms compared with others,^14,17-19 no review to date has systematically synthesized and compared their performance against aggregate-weighted EWSs.

Purpose

The purpose of this systematic review was to evaluate the recent literature regarding prognostic test accuracy and clinical workloads generated by EWSs using statistical modeling versus aggregate-weighted systems.

Search Strategy

Adhering to PRISMA protocol guidelines for systematic reviews, we searched the peer-reviewed literature in PubMed and CINAHL Plus, as well as conference proceedings and online repositories of patient safety organizations published between January 1, 2012 and September 15, 2018. We selected this timeframe because EWSs using statistical modeling are relatively new approaches compared with the body of evidence concerning aggregate-weighted EWSs. An expert PhD researcher confirmed the search results in a blinded independent query.

Inclusion and Exclusion Criteria

We included peer-reviewed articles reporting the area under the receiver operator curve (AUC),²⁰ or the equivalent c-statistic, of models predicting clinical deterioration (measured as the composite of transfer to intensive care unit (ICU) and/or mortality) among adult patients in general hospital wards. We excluded studies if they did not compare an EWS using statistical modeling with an aggregate-weighted EWS, did not report AUC, or only reported on an aggregate-weighted EWS. Excluded settings were pediatrics, obstetrics, emergency departments, ICUs, transitional care units, and oncology. We also excluded studies with samples limited to physiological monitoring, sepsis, or postsurgical subpopulations.

Data Abstraction

Following the TRIPOD guidelines for the reporting of predictive models,²¹ and the PRISMA and Cochrane Collaboration guidelines for systematic reviews,^22-24 we extracted study characteristics (Table 1), sample demographics (Appendix Table 4), model characteristics and performance (Appendix Table 5), and level of scientific evidence and risk of bias (Appendix Table 6). To address the potential for overfitting, we selected model performance results of the validation dataset rather than the derivation dataset, if reported. If studies reported multiple models in either EWS category, we selected the best-performing model for comparison.

Measures of Model Performance

Because predictive models can achieve good case identification at the expense of high clinical workloads, an assessment of model performance would be incomplete without measures of clinical utility. For clinicians, this aspect can be measured as the model’s PPV (the percentage of true positive alerts among all alerts), or more intelligibly, as the WDR, which equals 1/PPV. WDR indicates the number of patients requiring evaluation to identify and treat one true positive case.^9-11 It is known that differences in event rates (prevalence or pretest probability) influence a model’s PPV²⁵ and its reciprocal WDR. However, for systematic comparison, PPV and WDR can be standardized using a fixed representative event rate across studies.^24,26 We abstracted the reported PPV and WDR, and computed standardized PPV and WDR for an event rate of 4%.

Other measures included the area under the receiver operator curve (AUC),²⁰ sensitivity, and specificity. AUC plots a model’s false positive rate (x-axis) against its true positive rate (y-axis), with an ideal scenario of very high y-values and very low x-values.²⁷ Sensitivity (the model’s ability to detect a true positive case among all cases) and specificity (the model’s ability to detect a true noncase among all noncases²⁸) are influenced by chosen alert thresholds. It is incorrect to assume that a given model produces only one sensitivity/specificity result; for systematic comparison, we therefore selected results in the 50% sensitivity range, and separately, in the 92% specificity range for EWSs using statistical modeling. Then, we simulated a fixed sensitivity of 0.51 and assumed specificity of 0.87 in aggregate-weighted EWSs.

Search Results

The PubMed search for “early warning score OR early warning system AND deterioration OR predict transfer ICU” returned 285 peer-reviewed articles. A search on CINAHL Plus using the same filters and query terms returned 219 articles with no additional matches (Figure 1). Of the 285 articles, we excluded 269 during the abstract screen and 10 additional articles during full-text review (Figure 1). A final review of the reference lists of the six selected studies did not yield additional articles.

Study Characteristics

There were several similarities across the selected studies (Table 1). All occurred in the United States; all compared their model’s performance against at least one aggregate-weighted EWS model;^14,17-19,29 and all used retrospective cohort designs. Of the six studies, one took place in a single hospital;²⁹ three pooled data from five hospitals;^17,18,30 and two occurred in a large integrated healthcare delivery system using data from 14 and, subsequently, 21 hospitals.^14,19 The largest study¹⁴ included nearly 650,000 admissions, while the smallest study²⁹ reported slightly less than 7,500 admissions. Of the six studies, four used multivariable regression,^14,17,19,29 and two used machine learning techniques for outcome prediction.^18,30

Outcome Variables

The primary outcome for inclusion in this review was clinical deterioration measured by the composite of transfer to ICU and some measure of mortality. Churpek et al.^10,11 and Green et al.³⁰ also included cardiac arrest, and Alvarez et al.²² included respiratory compromise in their outcome composite.

Researchers used varying definitions of mortality, including “death outside the ICU in a patient whose care directive was full code;”^14,19 “death on the wards without attempted resuscitation;”^17,18 “an in-hospital death in patients without a DNR order at admission that occurred on the medical ward or in ICU within 24 hours after transfer;”²⁹ or “death within 24 hours.”³⁰

Predictor Variables

We observed a broad assortment of predictor variables. All models included vital signs (heart rate, respiratory rate, blood pressure, and venous oxygen saturation); mental state; laboratory data; age; and sex. Additional variables included comorbidity, shock index,³¹ severity of illness score, length of stay, event time of day, season, admission category, and length of stay,^14,19 among others.

Model Performance

Reported PPV ranged from 0.16 to 0.42 (mean = 0.27) in EWSs using statistical modeling and 0.15 to 0.28 (mean = 0.19) in aggregate-weighted EWS models. The weighted mean standardized PPV, adjusted for an event rate of 4% across studies (Table 2), was 0.21 in EWSs using statistical modeling versus 0.14 in aggregate-weighted EWS models (simulated at 0.51 sensitivity and 0.87 specificity).

Only two studies^14,19 reported the WDR metric (alerts generated to identify one true positive case) explicitly. Based on the above PPV results, EWSs using statistical modeling generated a standardized WDR of 4.9 in models using statistical modeling versus 7.1 in aggregate-weighted models (Figure 2). The delta of 2.2 evaluations to find and treat one true positive case equals a 45% relative increase in RRT evaluation workloads using aggregate-weighted EWSs.

AUC values ranged from 0.77 to 0.85 (weighted mean = 0.80) in EWSs using statistical modeling, indicating good model discrimination. AUCs of aggregate-weighted EWSs ranged from 0.70 to 0.76 (weighted mean = 0.73), indicating fair model discrimination (Figure 2). The overall AUC delta was 0.07. However, our estimates may possibly be favoring EWSs that use statistical modeling by virtue of their derivation in an original research population compared with aggregate-weighted EWSs that were derived externally. For example, sensitivity analysis of eCART,¹⁸ an EWS using machine learning, showed an AUC drop of 1% in a large external patient population,¹⁴ while NEWS AUCs¹³ dropped between 11% and 15% in two large external populations (Appendix Table 7).^14,30 For hospitals adopting an externally developed EWS using statistical modeling, these results suggest that an AUC delta of approximately 5% can be expected and 7% for an internally developed EWS.

Risk of Bias Assessment

We scored the studies by adapting the Cochrane Collaboration tool for assessing risk of bias ³² (Appendix Table 5). Of the six studies, five received total scores between 1.0 and 2.0 (indicating relatively low bias risk), and one study had a score of 3.5 (indicating higher bias risk). Low bias studies^14,17-19,30 used large samples across multiple hospitals, discussed the choice of predictor variables and outcomes more precisely, and reported their measurement approaches and analytic methods in more detail, including imputation of missing data and model calibration.

In this systematic review, we assessed the predictive ability of EWSs using statistical modeling versus aggregate-weighted EWS models to detect clinical deterioration risk in hospitalized adults in general wards. From 2007 to 2018, at least five systematic reviews examined aggregate-weighted EWSs in adult inpatient settings.^33-37 No systematic review, however, has synthesized the evidence of EWSs using statistical modeling.

The recent evidence is limited to six studies, of which five had favorable risk of bias scores. All studies included in this review demonstrated superior model performance of the EWSs using statistical modeling compared with an aggregate-weighted EWS, and at least five of the six studies employed rigor in design, measurement, and analytic method. The AUC absolute difference between EWSs using statistical modeling and aggregate-weighted EWSs was 7% overall, moving model performance from fair to good (Table 2; Figure 2). Although this increase in discriminative power may appear modest, it translates into avoiding a 45% increase in WDR workload generated by an aggregate-weighted EWS, approximately two patient evaluations for each true positive case.

Results of our review suggest that EWSs using statistical modeling predict clinical deterioration risk with better precision. This is an important finding for the following reasons: (1) Better risk prediction can support the activation of rescue; (2) Given federal mandates to curb spending, the elimination of some resource-intensive false positive evaluations supports high-value care;³⁸ and (3) The Quadruple Aim³⁹ accounts for clinician wellbeing. EWSs using statistical modeling may offer benefits in terms of clinician satisfaction with the human–system interface because better discrimination reduces the daily evaluation workload/cognitive burden and because the reduction of false positive alerts may reduce alert fatigue.^40,41

Still, an important issue with risk detection is that it is unknown which percentage of patients are uniquely identified by an EWS and not already under evaluation by the clinical team. For example, a recent study by Bedoya et al.⁴² found that using NEWS did not improve clinical outcomes and nurses frequently disregarded the alert. Another study⁴³ found that the combined clinical judgment of physicians and nurses had an AUC of 0.90 in predicting mortality. These results suggest that at certain times, an EWS alert may not add new useful information for clinicians even when it correctly identifies deterioration risk. It remains difficult to define exactly how many patients an EWS would have to uniquely identify to have clinical utility.

Even EWSs that use statistical modeling cannot detect all true deterioration cases perfectly, and they may at times trigger an alert only when the clinical team is already aware of a patient’s clinical decline. Consequently, EWSs using statistical modeling can at best augment and support—but not replace—RRT rounding, physician workup, and vigilant frontline staff. However, clinicians, too, are not perfect, and the failure-to-rescue literature suggests that certain human factors are antecedents to patient crises (eg, stress and distraction,^44-46 judging by precedent/experience,^44,47 and innate limitations of human cognition⁴⁷). Because neither clinicians nor EWSs can predict deterioration perfectly, the best possible rescue response combines clinical vigilance, RRT rounding, and EWSs using statistical modeling as complementary solutions.

Our findings suggest that predictive models cannot be judged purely on AUC (in fact, it would be ill-advised) but also by their clinical utility (expressed in WDR and PPV): How many patients does a clinician need to evaluate?^9-11 Precision is not meaningful if it comes at the expense of unmanageable evaluation workloads, and our findings suggest that clinicians should evaluate models based on their clinical utility. Hospitals considering adoption of an EWS using statistical modeling should consider that externally developed EWSs appear to experience a performance drop when applied to a new patient population; a slightly higher WDR and slightly lower AUC can be expected. EWSs using statistical modeling appear to perform best when tailored to the targeted patient population (or are derived in-house). Model depreciation over time will likely require recalibration. In addition, adoption of a machine learning algorithm may mean that original model results are obscured by the black box output of the algorithm.^48-50

Findings from this systematic review are subject to several limitations. First, we applied strict inclusion criteria, which led us to exclude studies that offered findings in specialty units and specific patient subpopulations, among others. In the interest of systematic comparison, our findings are limited to general wards. We also restricted our search to recent studies that reported on models predicting clinical deterioration, which we defined as the composite of ICU transfer and/or death. Clinically, deteriorating patients in general wards either die or are transferred to ICU. This criterion resulted in exclusion of the Rothman Index,⁵¹ which predicts “death within 24 hours” but not ICU transfer. The AUC in this study was higher than those selected in this review (0.93 compared to 0.82 for MEWS; AUC delta: 0.09). The higher AUC may be a function of the outcome definition (30-day mortality would be more challenging to predict). Therefore, hospitals or health systems interested in purchasing an EWS using statistical modeling should carefully consider the outcome selection and definition.

Second, as is true for systematic reviews in general,⁵² the degree of clinical and methodological heterogeneity across the selected studies may limit our findings. Studies occurred in various settings (university hospital, teaching hospitals, and community hospitals), which may serve diverging patient populations. We observed that studies in university-based settings had a higher event rate ranging from 5.6% to 7.8%, which may result in higher PPV results in these settings. However, this increase would apply to both EWS types equally. To arrive at a “true” reflection of model performance, the simulations for PPV and WDR have used a more conservative event rate of 4%. We observed heterogenous mortality definitions, which did not always account for the reality that a patient’s death may be an appropriate outcome (ie, it was concordant with treatment wishes in the context of severe illness or an end-of-life trajectory). Studies also used different sampling procedures; some allowed multiple observations although most did not. The variation in sampling may change PPV and limit our systematic comparison. However, regardless of methodological differences, our review suggests that EWSs using statistical modeling perform better than aggregate-weighted EWSs in each of the selected studies.

Third, systematic reviews may be subject to the issue of publication bias because they can only compare published results and could possibly omit an unknown number of unpublished studies. However, the selected studies uniformly demonstrated similar model improvements, which are plausibly related to the larger number of covariates, statistical methods, and shrinkage of random error.

Finally, this review was limited to the comparison of observational studies, which aimed to answer how the two EWS classes compared. These studies did not address whether an alert had an impact on clinical care and patient outcomes. Results from at least one randomized nonblinded controlled trial suggest that alert-driven RRT activation may reduce the length of stay by 24 hours and use of oximetry, but has no impact on mortality, ICU transfer, and ICU length of stay.⁵³

Our findings point to three areas of need for the field of predictive EWS research: (1) a standardized set of clinical deterioration outcome measures, (2) a standardized set of measures capturing clinical evaluation workload and alert frequency, and (3) cost estimates of clinical workloads with and without deployment of an EWS using statistical modeling. Given the present divergence of outcome definitions, EWS research may benefit from a common “clinical deterioration” outcome standard, including transfer to ICU, inpatient/30-day/90-day mortality, and death with DNR, comfort care, or hospice. The field is lacking a standardized clinical workload measure and an understanding of the net percentage of patients uniquely identified by an EWS.

By using predictive analytics, health systems may be better able to achieve the goals of high-value care and patient safety and support the Quadruple Aim. Still, gaps in knowledge exist regarding the measurement of the clinical processes triggered by EWSs, evaluation workloads, alert fatigue, clinician burnout associated with the human-alert interface, and costs versus benefits. Future research should evaluate the degree to which EWSs can identify risk among patients who are not already under evaluation by the clinical team, assess the balanced treatment effects of RRT interventions between decedents and survivors, and investigate clinical process times relative to the time of an EWS alert using statistical modeling.

Acknowledgments

The authors would like to thank Ms. Jill Pope at the Kaiser Permanente Center for Health Research in Portland, OR for her assistance with manuscript preparation. Daniel Linnen would like to thank Dr. Linda Franck, PhD, RN, FAAN, Professor at the University of California, San Francisco, School of Nursing for reviewing the manuscript.

Disclosures

The authors declare no conflicts of interest.

Funding

The Maribelle & Stephen Leavitt Scholarship, the Jonas Nurse Scholars Scholarship at the University of California, San Francisco, and the Nurse Scholars Academy Predoctoral Research Fellowship at Kaiser Permanente Northern California supported this study during Daniel Linnen’s doctoral training at the University of California, San Francisco. Dr. Vincent Liu was funded by National Institute of General Medical Sciences Grant K23GM112018.