“Canst thou not minister to a mind diseased, pluck from the memory a rooted sorrow, raze out the written troubles of the brain, and with some sweet oblivious antidote cleanse the stuffed bosom of that perilous stuff which weighs upon the heart?”
– William Shakespeare, Macbeth

For many years, the United States has been experiencing a shortage of psychiatrists. Currently, there are only 28,000 to 33,000 psychiatrists in active patient care practice in the United States.^1,2 The lack of psychiatrists is pronounced in many areas of the country, including rural regions, some urban neighborhoods, and community health centers. In approximately half of US counties, there are no psychiatrists at all.³

While patients with mental illnesses often are treated in primary care settings, the need for qualified mental health clinicians remains acute. Two-thirds of primary care physicians report difficulty in referring patients for mental health care, due to the shortage of clinicians and long wait times for patients to be seen.⁴ In the Department of Veterans Affairs (VA), the shortage of qualified psychiatrists is even more acute, due to ongoing combat operations and an increased number of missions and manpower requirements to complete them, which also has increased veterans’ mental health needs during life after their service.⁵

The outlook for providing adequate numbers of psychiatrists in the future is even more concerning. Based on a population analysis, Satiani et al⁶ predicts an extreme shortage of psychiatrists for the next 30 years, with the availability of psychiatrists per population expected to reach an all-time low by 2024. Based on ratios from the Department of Health and Human Services, this would mean a shortage of 14,000 to 31,000 psychiatrists over the next 5 to 6 years alone. This is due primarily to the expected retirement of more than 25,000 psychiatrists age >55 during the next 5 years. With mental illness becoming the costliest medical condition in the United States, at $201 billion annually, the potential impact of this shortage is alarming.⁶

Addressing the shortage

Efforts aimed at increasing the number of psychiatrists, improving access to care, and improving efficiency of care have focused on expanding recruitment and training capacity in psychiatry residency programs, utilizing new models such as telepsychiatry and Assertive Community Treatment (ACT) teams, increasing the number of Certified Community Behavioral Health Clinics, and embedding psychiatrists in large primary care practices.⁷ Another avenue for addressing the psychiatrist shortage has been the training and hiring of more advanced practice clinicians, including physician assistants (PAs) and nurse practitioners (NPs). Approximately 11,000 PAs and NPs specialize in psychiatry in the United States; this number represents more than one-third of the current psychiatrist workforce. This substantially adds to the available mental health professionals who can assess, monitor, and treat mental health conditions, and creates an opportunity for further growth to help make up for the expected shortfall of psychiatrists. In an analysis of a data set that included information on 90% of physician practices in the United States, Martsolf et al⁸ found that in 2016, 28% of specialty practices employed advanced practice clinicians.

Physician assistants and NPs make up the largest group of non-physician mental health professionals who can prescribe medications. Physician assistant training is most closely aligned with the allopathic training model of physicians.⁹ Some typical duties of PAs working in a psychiatric setting are outlined in the Table.

How many PAs elect to specialize in psychiatry, compared with the percentage of physicians who choose psychiatry as a career? Data from the National Commission on Certification of Physician Assistants (NCCPA) revealed that in 2018 there were 1,470 PAs working in psychiatry, or approximately 1.5% of all PAs in practice.¹⁰ In comparison, approximately 5% of physicians complete residency training in psychiatry.²

Continue to: Although the need for more...

Although the need for more mental health professionals—especially those who can prescribe—is well documented, PA practice in psychiatry has been underrepresented, with PAs choosing to work in the field at a rate just over one-fourth that of physicians. While there is no clear explanation for the lack of PAs in psychiatry, PA programs’ training model has been to produce generalist clinicians who can work in numerous settings, particularly primary care. However, during the past several decades, PA practice choice has shifted largely from primary care to specialty care. In 1974, an estimated 68.8% of PAs worked in primary care settings (family medicine/general practice, general internal medicine, and general pediatrics), while the remainder worked in specialty areas.¹¹ In contrast, by 2018, only 25.8% of PAs worked in primary care settings.¹⁰ Despite more PAs choosing to work in medical specialties, the number choosing psychiatry remains very low. With the great need for well-trained mental health prescribers, the opportunity for growth in this area of medicine and increased salary incentive should serve as an impetus for PAs to consider psychiatry. Like their physician counterparts, PAs working in specialty areas of medicine tend to be paid more, sometimes substantially more.¹²

Training requirements

What is the level of training and experience for PAs who choose to work in psychiatry? Physician assistant program applicants generally come from a pre-med background with a Bachelor’s degree in a hard science, and often have medical experience as a nurse, paramedic, emergency medical technician, or other health profession. Physician assistants are trained in the same medical model of care as physicians, although their training is structured over an average 27-month cycle, with 1 year devoted to didactic education and 1 year or more devoted to clinical training.¹³ They are qualified to “go to work” soon after graduating and passing the NCCPA Physician Assistant National Certifying Examination (PANCE), and may require a state license. Upon graduation, PAs have received approximately 1,000 hours of didactic and 2,000 hours of clinical training across the general spectrum of medicine.

Physician assistants who choose to specialize in psychiatry may complete a residency/fellowship in psychiatry of approximately 1 year, and/or obtain the Certificate of Added Qualification (CAQ) in psychiatry from the NCCPA. Most PAs who work in psychiatry have done so through “on-the-job” training, where their knowledge and skills have expanded through working with their supervising physician(s) and gaining experience from their clinical practice and self-study. For many years, there were only 1 or 2 PA residency/fellowship opportunities in psychiatry in the United States for PAs wanting to acquire additional formal didactic and clinical knowledge and skills in psychiatry. Fortunately, there has been a growing number of PA residencies/fellowships in psychiatry. These programs are typically 1 year in length and can provide a PA who wants to specialize in psychiatry with an additional 300 to >500 didactic and 1,500 to 2,000 clinical hours of training in the assessment, diagnosis, and treatment of the spectrum of psychiatric conditions. Currently, there are 10 to 12 programs in the United States that offer this training to PAs, producing approximately 18 to 20 residency-trained psychiatric PAs each year. Almost one-half of PAs who are residency-trained in psychiatry are being trained in VA facilities and affiliated institutions sponsored by the VA Office of Academic Affiliations. Along with the PA’s basic education, the additional knowledge and skills acquired in residency prepare the PA to be a highly capable psychiatric clinician, with a combined 1,500 didactic and 4,000 clinical hours of training in general medicine plus psychiatry. The addition of the CAQ demonstrates the PA’s commitment to additional learning in psychiatry, as the added work experience requirements, the additional postgraduate continuing medical education requirements in psychiatry, and the psychiatry board exam clearly show dedication to a higher level of knowledge and skill in the specialty.

Because PAs have been trained as generalists who are able to work in any setting or specialty, they have a broad range of knowledge in medicine and surgery. This can be especially helpful when working in a psychiatric practice, where they can provide an added medical focus to patient care when needed. As more PAs are choosing to work in a specialty area for much or all of their practice, they are able to gain significant knowledge and skills in that specialty.

Getting more PAs into psychiatry

So what does the future hold for PAs in psychiatry? The increased need and opportunity in mental health will likely draw a higher percentage of PAs to this specialty. Hopefully, an increase in the number of PA psychiatry residencies or advanced mental health training opportunities, and the continued goal of obtaining the CAQ in psychiatry, will serve to increase the number of psychiatric PAs.¹⁴

Continue to:  The NCCPA has also recognized...

The NCCPA has also recognized the importance of increasing PA knowledge and integration in mental health care by establishing a PArtners in Mental Health Steering Committee, composed of leaders from the largest PA organizations, including the American Academy of Physician Assistants, the Accreditation Review Commission on Education for the Physician Assistant, the Physician Assistant Education Association, the Physician Assistant Foundation, and other PA and interprofessional members. The NCCPA’s PArtners in Mental Health Initiative: Stakeholders Report 2018 outlines an ongoing strategy to increase PA engagement in and awareness of mental health among the PA community and future providers via outreach to member organizations, state societies, PA programs, and those at the state and national level who legislate and reimburse PA services for mental health care.¹⁵ The Steering Committee’s recommendations include:

• enhancing PA educational approaches in mental health
• strengthening the PA practice environment to address mental health needs and foster integration
• promoting national campaigns to raise the profile of PAs addressing mental health across disciplines
• creating an organizational structure that incorporates current participants, offers backbone support to this movement, and plans for communication and financing.¹⁶

The role of PA educators

In the end, PA leaders and educators will play a substantial role in influencing future PAs to seek a career in psychiatry. Currently, psychiatry education varies among PA programs. Some offer robust didactic and clinical education and training, while other programs are limited in the number of hours of psychiatric didactic education, and may offer psychiatry clinical opportunities only in the context of a primary care setting, rather than in a dedicated psychiatric setting. Additionally, the mission of training PAs for generalist, primary care practice may limit many PAs from considering psychiatry because they do not necessarily view psychiatry as closely aligning with primary care generalist practice the way cardiology, pulmonology, gastroenterology, or other internal medicine specialties do.

In terms of PA postgraduate education, many PAs have completed residencies in surgical specialties or emergency medicine. Coincidentally, surgery and emergency medicine residencies are the most prolific of the postgraduate residency programs, adding a significant number of well-trained PAs to these specialties. The NCCPA also offers the CAQ for Cardiovascular and Thoracic Surgery, Orthopedic Surgery, and Emergency Medicine, which may attract PAs into these specialties, with or without completing a residency.

Because the NCCPA also offers the CAQ in Psychiatry, it would be reasonable and attractive for PAs who complete a psychiatry residency to obtain this certification. In fact, the PA psychiatry residency at our own institution trains our residents to be fully prepared and board-eligible to take the CAQ in Psychiatry upon completing residency. To date, every PA residency graduate who has completed our program and taken the CAQ in Psychiatry exam has passed and been awarded the CAQ in Psychiatry. They have proven themselves to the program and the NCCPA, and have impressed their employers with their clinical abilities and medical knowledge.

For psychiatrists, the addition of a well-trained or willing-to-be-trained PA to the practice can provide an economic advantage and strong team partnership that ensures optimal care for patients in this time of shortage of skilled mental health clinicians. The need is clear and will continue. Physician assistant educators must provide adequate didactic and clinical training in psychiatry to PA students, and support students interested in pursuing a career path in this specialty. Physician assistant organizations must meet the challenge of increasing the number of PAs in psychiatry, and encourage the establishment of additional post-graduate residency programs in psychiatry for PAs. Lastly, more PAs need to be made aware that psychiatry is an in-demand specialty that offers broad autonomy and rewarding clinical work.

Continue to: Bottom Line

Physician assistants (PAs) who choose to specialize in psychiatry will find enormous opportunity, as the need for well-trained and knowledgeable mental health providers is acute. Those PAs who obtain additional training and/or certification in psychiatry will be highly valued and sought-after, with an abundance of job opportunities. Physician assistant programs should continue to improve didactic and clinical training for their students in psychiatry, and encourage increased numbers of PAs to consider psychiatry as a career path.

Related Resources

• National Commission on Certification of Physician Assistants. Psychiatry Certificate of Added Qualifications. https://www.nccpa.net/psychiatry.
• Association of Physician Assistants in Psychiatry. http://psychpa.com/.

“Canst thou not minister to a mind diseased, pluck from the memory a rooted sorrow, raze out the written troubles of the brain, and with some sweet oblivious antidote cleanse the stuffed bosom of that perilous stuff which weighs upon the heart?”
– William Shakespeare, Macbeth

For many years, the United States has been experiencing a shortage of psychiatrists. Currently, there are only 28,000 to 33,000 psychiatrists in active patient care practice in the United States.^1,2 The lack of psychiatrists is pronounced in many areas of the country, including rural regions, some urban neighborhoods, and community health centers. In approximately half of US counties, there are no psychiatrists at all.³

While patients with mental illnesses often are treated in primary care settings, the need for qualified mental health clinicians remains acute. Two-thirds of primary care physicians report difficulty in referring patients for mental health care, due to the shortage of clinicians and long wait times for patients to be seen.⁴ In the Department of Veterans Affairs (VA), the shortage of qualified psychiatrists is even more acute, due to ongoing combat operations and an increased number of missions and manpower requirements to complete them, which also has increased veterans’ mental health needs during life after their service.⁵

The outlook for providing adequate numbers of psychiatrists in the future is even more concerning. Based on a population analysis, Satiani et al⁶ predicts an extreme shortage of psychiatrists for the next 30 years, with the availability of psychiatrists per population expected to reach an all-time low by 2024. Based on ratios from the Department of Health and Human Services, this would mean a shortage of 14,000 to 31,000 psychiatrists over the next 5 to 6 years alone. This is due primarily to the expected retirement of more than 25,000 psychiatrists age >55 during the next 5 years. With mental illness becoming the costliest medical condition in the United States, at $201 billion annually, the potential impact of this shortage is alarming.⁶

Addressing the shortage

Efforts aimed at increasing the number of psychiatrists, improving access to care, and improving efficiency of care have focused on expanding recruitment and training capacity in psychiatry residency programs, utilizing new models such as telepsychiatry and Assertive Community Treatment (ACT) teams, increasing the number of Certified Community Behavioral Health Clinics, and embedding psychiatrists in large primary care practices.⁷ Another avenue for addressing the psychiatrist shortage has been the training and hiring of more advanced practice clinicians, including physician assistants (PAs) and nurse practitioners (NPs). Approximately 11,000 PAs and NPs specialize in psychiatry in the United States; this number represents more than one-third of the current psychiatrist workforce. This substantially adds to the available mental health professionals who can assess, monitor, and treat mental health conditions, and creates an opportunity for further growth to help make up for the expected shortfall of psychiatrists. In an analysis of a data set that included information on 90% of physician practices in the United States, Martsolf et al⁸ found that in 2016, 28% of specialty practices employed advanced practice clinicians.

Physician assistants and NPs make up the largest group of non-physician mental health professionals who can prescribe medications. Physician assistant training is most closely aligned with the allopathic training model of physicians.⁹ Some typical duties of PAs working in a psychiatric setting are outlined in the Table.

How many PAs elect to specialize in psychiatry, compared with the percentage of physicians who choose psychiatry as a career? Data from the National Commission on Certification of Physician Assistants (NCCPA) revealed that in 2018 there were 1,470 PAs working in psychiatry, or approximately 1.5% of all PAs in practice.¹⁰ In comparison, approximately 5% of physicians complete residency training in psychiatry.²

Continue to: Although the need for more...

Although the need for more mental health professionals—especially those who can prescribe—is well documented, PA practice in psychiatry has been underrepresented, with PAs choosing to work in the field at a rate just over one-fourth that of physicians. While there is no clear explanation for the lack of PAs in psychiatry, PA programs’ training model has been to produce generalist clinicians who can work in numerous settings, particularly primary care. However, during the past several decades, PA practice choice has shifted largely from primary care to specialty care. In 1974, an estimated 68.8% of PAs worked in primary care settings (family medicine/general practice, general internal medicine, and general pediatrics), while the remainder worked in specialty areas.¹¹ In contrast, by 2018, only 25.8% of PAs worked in primary care settings.¹⁰ Despite more PAs choosing to work in medical specialties, the number choosing psychiatry remains very low. With the great need for well-trained mental health prescribers, the opportunity for growth in this area of medicine and increased salary incentive should serve as an impetus for PAs to consider psychiatry. Like their physician counterparts, PAs working in specialty areas of medicine tend to be paid more, sometimes substantially more.¹²

Training requirements

What is the level of training and experience for PAs who choose to work in psychiatry? Physician assistant program applicants generally come from a pre-med background with a Bachelor’s degree in a hard science, and often have medical experience as a nurse, paramedic, emergency medical technician, or other health profession. Physician assistants are trained in the same medical model of care as physicians, although their training is structured over an average 27-month cycle, with 1 year devoted to didactic education and 1 year or more devoted to clinical training.¹³ They are qualified to “go to work” soon after graduating and passing the NCCPA Physician Assistant National Certifying Examination (PANCE), and may require a state license. Upon graduation, PAs have received approximately 1,000 hours of didactic and 2,000 hours of clinical training across the general spectrum of medicine.

Physician assistants who choose to specialize in psychiatry may complete a residency/fellowship in psychiatry of approximately 1 year, and/or obtain the Certificate of Added Qualification (CAQ) in psychiatry from the NCCPA. Most PAs who work in psychiatry have done so through “on-the-job” training, where their knowledge and skills have expanded through working with their supervising physician(s) and gaining experience from their clinical practice and self-study. For many years, there were only 1 or 2 PA residency/fellowship opportunities in psychiatry in the United States for PAs wanting to acquire additional formal didactic and clinical knowledge and skills in psychiatry. Fortunately, there has been a growing number of PA residencies/fellowships in psychiatry. These programs are typically 1 year in length and can provide a PA who wants to specialize in psychiatry with an additional 300 to >500 didactic and 1,500 to 2,000 clinical hours of training in the assessment, diagnosis, and treatment of the spectrum of psychiatric conditions. Currently, there are 10 to 12 programs in the United States that offer this training to PAs, producing approximately 18 to 20 residency-trained psychiatric PAs each year. Almost one-half of PAs who are residency-trained in psychiatry are being trained in VA facilities and affiliated institutions sponsored by the VA Office of Academic Affiliations. Along with the PA’s basic education, the additional knowledge and skills acquired in residency prepare the PA to be a highly capable psychiatric clinician, with a combined 1,500 didactic and 4,000 clinical hours of training in general medicine plus psychiatry. The addition of the CAQ demonstrates the PA’s commitment to additional learning in psychiatry, as the added work experience requirements, the additional postgraduate continuing medical education requirements in psychiatry, and the psychiatry board exam clearly show dedication to a higher level of knowledge and skill in the specialty.

Because PAs have been trained as generalists who are able to work in any setting or specialty, they have a broad range of knowledge in medicine and surgery. This can be especially helpful when working in a psychiatric practice, where they can provide an added medical focus to patient care when needed. As more PAs are choosing to work in a specialty area for much or all of their practice, they are able to gain significant knowledge and skills in that specialty.

Getting more PAs into psychiatry

So what does the future hold for PAs in psychiatry? The increased need and opportunity in mental health will likely draw a higher percentage of PAs to this specialty. Hopefully, an increase in the number of PA psychiatry residencies or advanced mental health training opportunities, and the continued goal of obtaining the CAQ in psychiatry, will serve to increase the number of psychiatric PAs.¹⁴

Continue to:  The NCCPA has also recognized...

The NCCPA has also recognized the importance of increasing PA knowledge and integration in mental health care by establishing a PArtners in Mental Health Steering Committee, composed of leaders from the largest PA organizations, including the American Academy of Physician Assistants, the Accreditation Review Commission on Education for the Physician Assistant, the Physician Assistant Education Association, the Physician Assistant Foundation, and other PA and interprofessional members. The NCCPA’s PArtners in Mental Health Initiative: Stakeholders Report 2018 outlines an ongoing strategy to increase PA engagement in and awareness of mental health among the PA community and future providers via outreach to member organizations, state societies, PA programs, and those at the state and national level who legislate and reimburse PA services for mental health care.¹⁵ The Steering Committee’s recommendations include:

• enhancing PA educational approaches in mental health
• strengthening the PA practice environment to address mental health needs and foster integration
• promoting national campaigns to raise the profile of PAs addressing mental health across disciplines
• creating an organizational structure that incorporates current participants, offers backbone support to this movement, and plans for communication and financing.¹⁶

The role of PA educators

In the end, PA leaders and educators will play a substantial role in influencing future PAs to seek a career in psychiatry. Currently, psychiatry education varies among PA programs. Some offer robust didactic and clinical education and training, while other programs are limited in the number of hours of psychiatric didactic education, and may offer psychiatry clinical opportunities only in the context of a primary care setting, rather than in a dedicated psychiatric setting. Additionally, the mission of training PAs for generalist, primary care practice may limit many PAs from considering psychiatry because they do not necessarily view psychiatry as closely aligning with primary care generalist practice the way cardiology, pulmonology, gastroenterology, or other internal medicine specialties do.

In terms of PA postgraduate education, many PAs have completed residencies in surgical specialties or emergency medicine. Coincidentally, surgery and emergency medicine residencies are the most prolific of the postgraduate residency programs, adding a significant number of well-trained PAs to these specialties. The NCCPA also offers the CAQ for Cardiovascular and Thoracic Surgery, Orthopedic Surgery, and Emergency Medicine, which may attract PAs into these specialties, with or without completing a residency.

Because the NCCPA also offers the CAQ in Psychiatry, it would be reasonable and attractive for PAs who complete a psychiatry residency to obtain this certification. In fact, the PA psychiatry residency at our own institution trains our residents to be fully prepared and board-eligible to take the CAQ in Psychiatry upon completing residency. To date, every PA residency graduate who has completed our program and taken the CAQ in Psychiatry exam has passed and been awarded the CAQ in Psychiatry. They have proven themselves to the program and the NCCPA, and have impressed their employers with their clinical abilities and medical knowledge.

For psychiatrists, the addition of a well-trained or willing-to-be-trained PA to the practice can provide an economic advantage and strong team partnership that ensures optimal care for patients in this time of shortage of skilled mental health clinicians. The need is clear and will continue. Physician assistant educators must provide adequate didactic and clinical training in psychiatry to PA students, and support students interested in pursuing a career path in this specialty. Physician assistant organizations must meet the challenge of increasing the number of PAs in psychiatry, and encourage the establishment of additional post-graduate residency programs in psychiatry for PAs. Lastly, more PAs need to be made aware that psychiatry is an in-demand specialty that offers broad autonomy and rewarding clinical work.

Continue to: Bottom Line

Physician assistants (PAs) who choose to specialize in psychiatry will find enormous opportunity, as the need for well-trained and knowledgeable mental health providers is acute. Those PAs who obtain additional training and/or certification in psychiatry will be highly valued and sought-after, with an abundance of job opportunities. Physician assistant programs should continue to improve didactic and clinical training for their students in psychiatry, and encourage increased numbers of PAs to consider psychiatry as a career path.

Related Resources

• National Commission on Certification of Physician Assistants. Psychiatry Certificate of Added Qualifications. https://www.nccpa.net/psychiatry.
• Association of Physician Assistants in Psychiatry. http://psychpa.com/.

“Canst thou not minister to a mind diseased, pluck from the memory a rooted sorrow, raze out the written troubles of the brain, and with some sweet oblivious antidote cleanse the stuffed bosom of that perilous stuff which weighs upon the heart?”
– William Shakespeare, Macbeth

For many years, the United States has been experiencing a shortage of psychiatrists. Currently, there are only 28,000 to 33,000 psychiatrists in active patient care practice in the United States.^1,2 The lack of psychiatrists is pronounced in many areas of the country, including rural regions, some urban neighborhoods, and community health centers. In approximately half of US counties, there are no psychiatrists at all.³

While patients with mental illnesses often are treated in primary care settings, the need for qualified mental health clinicians remains acute. Two-thirds of primary care physicians report difficulty in referring patients for mental health care, due to the shortage of clinicians and long wait times for patients to be seen.⁴ In the Department of Veterans Affairs (VA), the shortage of qualified psychiatrists is even more acute, due to ongoing combat operations and an increased number of missions and manpower requirements to complete them, which also has increased veterans’ mental health needs during life after their service.⁵

The outlook for providing adequate numbers of psychiatrists in the future is even more concerning. Based on a population analysis, Satiani et al⁶ predicts an extreme shortage of psychiatrists for the next 30 years, with the availability of psychiatrists per population expected to reach an all-time low by 2024. Based on ratios from the Department of Health and Human Services, this would mean a shortage of 14,000 to 31,000 psychiatrists over the next 5 to 6 years alone. This is due primarily to the expected retirement of more than 25,000 psychiatrists age >55 during the next 5 years. With mental illness becoming the costliest medical condition in the United States, at $201 billion annually, the potential impact of this shortage is alarming.⁶

Addressing the shortage

Efforts aimed at increasing the number of psychiatrists, improving access to care, and improving efficiency of care have focused on expanding recruitment and training capacity in psychiatry residency programs, utilizing new models such as telepsychiatry and Assertive Community Treatment (ACT) teams, increasing the number of Certified Community Behavioral Health Clinics, and embedding psychiatrists in large primary care practices.⁷ Another avenue for addressing the psychiatrist shortage has been the training and hiring of more advanced practice clinicians, including physician assistants (PAs) and nurse practitioners (NPs). Approximately 11,000 PAs and NPs specialize in psychiatry in the United States; this number represents more than one-third of the current psychiatrist workforce. This substantially adds to the available mental health professionals who can assess, monitor, and treat mental health conditions, and creates an opportunity for further growth to help make up for the expected shortfall of psychiatrists. In an analysis of a data set that included information on 90% of physician practices in the United States, Martsolf et al⁸ found that in 2016, 28% of specialty practices employed advanced practice clinicians.

Physician assistants and NPs make up the largest group of non-physician mental health professionals who can prescribe medications. Physician assistant training is most closely aligned with the allopathic training model of physicians.⁹ Some typical duties of PAs working in a psychiatric setting are outlined in the Table.

How many PAs elect to specialize in psychiatry, compared with the percentage of physicians who choose psychiatry as a career? Data from the National Commission on Certification of Physician Assistants (NCCPA) revealed that in 2018 there were 1,470 PAs working in psychiatry, or approximately 1.5% of all PAs in practice.¹⁰ In comparison, approximately 5% of physicians complete residency training in psychiatry.²

Continue to: Although the need for more...

Although the need for more mental health professionals—especially those who can prescribe—is well documented, PA practice in psychiatry has been underrepresented, with PAs choosing to work in the field at a rate just over one-fourth that of physicians. While there is no clear explanation for the lack of PAs in psychiatry, PA programs’ training model has been to produce generalist clinicians who can work in numerous settings, particularly primary care. However, during the past several decades, PA practice choice has shifted largely from primary care to specialty care. In 1974, an estimated 68.8% of PAs worked in primary care settings (family medicine/general practice, general internal medicine, and general pediatrics), while the remainder worked in specialty areas.¹¹ In contrast, by 2018, only 25.8% of PAs worked in primary care settings.¹⁰ Despite more PAs choosing to work in medical specialties, the number choosing psychiatry remains very low. With the great need for well-trained mental health prescribers, the opportunity for growth in this area of medicine and increased salary incentive should serve as an impetus for PAs to consider psychiatry. Like their physician counterparts, PAs working in specialty areas of medicine tend to be paid more, sometimes substantially more.¹²

Training requirements

What is the level of training and experience for PAs who choose to work in psychiatry? Physician assistant program applicants generally come from a pre-med background with a Bachelor’s degree in a hard science, and often have medical experience as a nurse, paramedic, emergency medical technician, or other health profession. Physician assistants are trained in the same medical model of care as physicians, although their training is structured over an average 27-month cycle, with 1 year devoted to didactic education and 1 year or more devoted to clinical training.¹³ They are qualified to “go to work” soon after graduating and passing the NCCPA Physician Assistant National Certifying Examination (PANCE), and may require a state license. Upon graduation, PAs have received approximately 1,000 hours of didactic and 2,000 hours of clinical training across the general spectrum of medicine.

Physician assistants who choose to specialize in psychiatry may complete a residency/fellowship in psychiatry of approximately 1 year, and/or obtain the Certificate of Added Qualification (CAQ) in psychiatry from the NCCPA. Most PAs who work in psychiatry have done so through “on-the-job” training, where their knowledge and skills have expanded through working with their supervising physician(s) and gaining experience from their clinical practice and self-study. For many years, there were only 1 or 2 PA residency/fellowship opportunities in psychiatry in the United States for PAs wanting to acquire additional formal didactic and clinical knowledge and skills in psychiatry. Fortunately, there has been a growing number of PA residencies/fellowships in psychiatry. These programs are typically 1 year in length and can provide a PA who wants to specialize in psychiatry with an additional 300 to >500 didactic and 1,500 to 2,000 clinical hours of training in the assessment, diagnosis, and treatment of the spectrum of psychiatric conditions. Currently, there are 10 to 12 programs in the United States that offer this training to PAs, producing approximately 18 to 20 residency-trained psychiatric PAs each year. Almost one-half of PAs who are residency-trained in psychiatry are being trained in VA facilities and affiliated institutions sponsored by the VA Office of Academic Affiliations. Along with the PA’s basic education, the additional knowledge and skills acquired in residency prepare the PA to be a highly capable psychiatric clinician, with a combined 1,500 didactic and 4,000 clinical hours of training in general medicine plus psychiatry. The addition of the CAQ demonstrates the PA’s commitment to additional learning in psychiatry, as the added work experience requirements, the additional postgraduate continuing medical education requirements in psychiatry, and the psychiatry board exam clearly show dedication to a higher level of knowledge and skill in the specialty.

Because PAs have been trained as generalists who are able to work in any setting or specialty, they have a broad range of knowledge in medicine and surgery. This can be especially helpful when working in a psychiatric practice, where they can provide an added medical focus to patient care when needed. As more PAs are choosing to work in a specialty area for much or all of their practice, they are able to gain significant knowledge and skills in that specialty.

Getting more PAs into psychiatry

So what does the future hold for PAs in psychiatry? The increased need and opportunity in mental health will likely draw a higher percentage of PAs to this specialty. Hopefully, an increase in the number of PA psychiatry residencies or advanced mental health training opportunities, and the continued goal of obtaining the CAQ in psychiatry, will serve to increase the number of psychiatric PAs.¹⁴

Continue to:  The NCCPA has also recognized...

The NCCPA has also recognized the importance of increasing PA knowledge and integration in mental health care by establishing a PArtners in Mental Health Steering Committee, composed of leaders from the largest PA organizations, including the American Academy of Physician Assistants, the Accreditation Review Commission on Education for the Physician Assistant, the Physician Assistant Education Association, the Physician Assistant Foundation, and other PA and interprofessional members. The NCCPA’s PArtners in Mental Health Initiative: Stakeholders Report 2018 outlines an ongoing strategy to increase PA engagement in and awareness of mental health among the PA community and future providers via outreach to member organizations, state societies, PA programs, and those at the state and national level who legislate and reimburse PA services for mental health care.¹⁵ The Steering Committee’s recommendations include:

• enhancing PA educational approaches in mental health
• strengthening the PA practice environment to address mental health needs and foster integration
• promoting national campaigns to raise the profile of PAs addressing mental health across disciplines
• creating an organizational structure that incorporates current participants, offers backbone support to this movement, and plans for communication and financing.¹⁶

The role of PA educators

In the end, PA leaders and educators will play a substantial role in influencing future PAs to seek a career in psychiatry. Currently, psychiatry education varies among PA programs. Some offer robust didactic and clinical education and training, while other programs are limited in the number of hours of psychiatric didactic education, and may offer psychiatry clinical opportunities only in the context of a primary care setting, rather than in a dedicated psychiatric setting. Additionally, the mission of training PAs for generalist, primary care practice may limit many PAs from considering psychiatry because they do not necessarily view psychiatry as closely aligning with primary care generalist practice the way cardiology, pulmonology, gastroenterology, or other internal medicine specialties do.

In terms of PA postgraduate education, many PAs have completed residencies in surgical specialties or emergency medicine. Coincidentally, surgery and emergency medicine residencies are the most prolific of the postgraduate residency programs, adding a significant number of well-trained PAs to these specialties. The NCCPA also offers the CAQ for Cardiovascular and Thoracic Surgery, Orthopedic Surgery, and Emergency Medicine, which may attract PAs into these specialties, with or without completing a residency.

Because the NCCPA also offers the CAQ in Psychiatry, it would be reasonable and attractive for PAs who complete a psychiatry residency to obtain this certification. In fact, the PA psychiatry residency at our own institution trains our residents to be fully prepared and board-eligible to take the CAQ in Psychiatry upon completing residency. To date, every PA residency graduate who has completed our program and taken the CAQ in Psychiatry exam has passed and been awarded the CAQ in Psychiatry. They have proven themselves to the program and the NCCPA, and have impressed their employers with their clinical abilities and medical knowledge.

For psychiatrists, the addition of a well-trained or willing-to-be-trained PA to the practice can provide an economic advantage and strong team partnership that ensures optimal care for patients in this time of shortage of skilled mental health clinicians. The need is clear and will continue. Physician assistant educators must provide adequate didactic and clinical training in psychiatry to PA students, and support students interested in pursuing a career path in this specialty. Physician assistant organizations must meet the challenge of increasing the number of PAs in psychiatry, and encourage the establishment of additional post-graduate residency programs in psychiatry for PAs. Lastly, more PAs need to be made aware that psychiatry is an in-demand specialty that offers broad autonomy and rewarding clinical work.

Continue to: Bottom Line

Physician assistants (PAs) who choose to specialize in psychiatry will find enormous opportunity, as the need for well-trained and knowledgeable mental health providers is acute. Those PAs who obtain additional training and/or certification in psychiatry will be highly valued and sought-after, with an abundance of job opportunities. Physician assistant programs should continue to improve didactic and clinical training for their students in psychiatry, and encourage increased numbers of PAs to consider psychiatry as a career path.

Related Resources

• National Commission on Certification of Physician Assistants. Psychiatry Certificate of Added Qualifications. https://www.nccpa.net/psychiatry.
• Association of Physician Assistants in Psychiatry. http://psychpa.com/.

Major depressive disorder (MDD) is a significant pediatric health problem, with a lifetime prevalence as high as 20% by the end of adolescence.^1-3 Major depressive disorder in adolescence is associated with significant morbidity, including poor social functioning, school difficulties, early pregnancy, and increased risk of physical illness and substance abuse.^4-6 It is also linked with significant mortality, with increased risk for suicide, which is now the second leading cause of death in individuals age 10 to 24 years.^1,7,8

As their name suggests, antidepressants comprise a group of medications that are used to treat MDD; they are also, however, first-line agents for generalized anxiety disorder (GAD), posttraumatic stress disorder (PTSD), and obsessive-compulsive disorder (OCD) in adults. Anxiety disorders (including GAD and other anxiety diagnoses) and PTSD are also common in childhood and adolescence with a combined lifetime prevalence ranging from 15% to 30%.^9,10 These disorders are also associated with increased risk of suicide.¹¹ For all of these disorders, depending on the severity of presentation and the preference of the patient, treatments are often a combination of psychotherapy and psychopharmacology.

Clinicians face several challenges when considering antidepressants for pediatric patients. Pediatricians and psychiatrists need to understand whether these medications work in children and adolescents, and whether there are unique developmental safety and tolerability issues. The evidence base in child psychiatry is considerably smaller compared with that of adult psychiatry. From this more limited evidence base also came the controversial “black-box” warning regarding a risk of emergent suicidality when starting antidepressants that accompanies all antidepressants for pediatric, but not adult, patients. This warning has had major effects on clinical encounters with children experiencing depression, including altering clinician prescribing behavior.¹²

In this article, we review the current evidence for antidepressant efficacy, tolerability, and safety in pediatric patients. We also suggest ways in which clinicians might choose, start, and stop antidepressants in children, as well as how to talk with parents about benefits, risks, and the black-box warning.

Do antidepressants work in children?

Selective serotonin reuptake inhibitors. Selective serotonin reuptake inhibitors (SSRIs) are the most commonly used class of antidepressants in both children and adults.¹³ While only a few SSRIs are FDA-approved for pediatric indications, the lack of FDA approval is typically related to a lack of sufficient testing in randomized controlled trials (RCTs) for specific pediatric indications, rather than to demonstrable differences in efficacy between antidepressant agents. Since there is currently no data to suggest inferiority of one agent compared to another in children or adults,^14,15 efficacy data will be discussed here as applied to the class of SSRIs, generalizing from RCTs conducted on individual drugs. Table 1 lists FDA indications and dosing information for individual antidepressants.

There is strong evidence that SSRIs are effective for treating pediatric anxiety disorders (eg, social anxiety disorder and GAD)¹⁶ and OCD,¹⁷ with numbers needed to treat (NNT) between 3 and 5. For both of these disorders, SSRIs combined with cognitive-behavioral therapy (CBT) have the highest likelihood of improving symptoms or achieving remission.^17,18

Selective serotonin reuptake inhibitors are also effective for treating pediatric MDD; however, the literature is more complex for this disorder compared to GAD and OCD as there are considerable differences in effect sizes between National Institute of Mental Health (NIMH)–funded studies and industry-sponsored trials.¹³ The major NIMH-sponsored adolescent depression trial, TADS (Treatment for Adolescents and Depression Study), showed that SSRIs (fluoxetine in this case) were quite effective, with an NNT of 4 over the acute phase (12 weeks).¹⁹ Ultimately, approximately 80% of adolescents improved over 9 months. Many industry-sponsored trials for MDD in pediatric patients had large placebo response rates (approximately 60%), which resulted in smaller between-group differences, and estimates of an NNT closer to 12,¹³ which has muddied the waters in meta-analyses that include all trials.²⁰ Improvement in depressive symptoms also appears to be bolstered by concomitant CBT in MDD,¹⁹ but not as robustly as in GAD and OCD. While the full benefit of SSRIs for depression may take as long as 8 weeks, a meta-analysis of depression studies of pediatric patients suggests that significant benefits from placebo are observed as early as 2 weeks, and that further treatment gains are minimal after 4 weeks.¹⁵ Thus, we recommend at least a 4- to 6-week trial at therapeutic dosing before deeming a medication a treatment failure.

Continue to: Posttraumatic stress disorder...

Posttraumatic stress disorder is a fourth disorder in which SSRIs are a first-line treatment in adults. The data for using SSRIs to treat pediatric patients with PTSD is scant, with only a few RCTs, and no large NIMH-funded trials. Randomized controlled trials have not demonstrated significant differences between SSRIs and placebo^21,22 and thus the current first-line recommendation in pediatric PTSD remains trauma-focused therapy, with good evidence for trauma-focused CBT.²³ Practically speaking, there can be considerable overlap of PTSD, depression, and anxiety symptoms in children,²³ and children with a history of trauma who also have comorbid MDD may benefit from medication if their symptoms persist despite an adequate trial of psychotherapy.

Taken together, the current evidence suggests that SSRIs are often effective in pediatric GAD, OCD, and MDD, with low NNTs (ranging from 3 to 5 based on NIMH-funded trials) for all of these disorders; there is not yet sufficient evidence of efficacy in pediatric patients with PTSD.

Fluoxetine has been studied more intensively than other SSRIs (for example, it was the antidepressant used in the TADS trial), and thus has the largest evidence base. For this reason, fluoxetine is often considered the first of the first-line options. Additionally, fluoxetine has a longer half-life than other antidepressants, which may make it more effective in situations where patients are likely to miss doses, and results in a lower risk of withdrawal symptoms when stopped due to “self-tapering.”

SNRIs and atypical antidepressants. Other antidepressants commonly used in pediatric patients but with far less evidence of efficacy include serotonin-norepinephrine reuptake inhibitors (SNRIs) and the atypical antidepressants bupropion and mirtazapine. The SNRI duloxetine is FDA-approved for treating GAD in children age 7 to 17, but there are no other pediatric indications for duloxetine, or for the other SNRIs.

In general, adverse effect profiles are worse for SNRIs compared to SSRIs, further limiting their utility. While there are no pediatric studies demonstrating SNRI efficacy for neuropathic pain, good data exists in adults.²⁴ Thus, an SNRI could be a reasonable option if a pediatric patient has failed prior adequate SSRI trials and also has comorbid neuropathic pain.

Continue to: Neither bupropion nor mirtazapine...

Neither bupropion nor mirtazapine have undergone rigorous testing in pediatric patients, and therefore these agents should generally be considered only once other first-line treatments have failed. Bupropion has been evaluated for attention-deficit/hyperactivity disorder (ADHD)²⁵ and for adolescent smoking cessation.²⁶ However, the evidence is weak, and bupropion is not considered a first-line option for children and adolescents.

Tricyclic antidepressants. Randomized controlled trials have demonstrated that tricyclic antidepressants (TCAs) are efficacious for treating several pediatric conditions; however, their significant side effect profile, their monitoring requirements, as well as their lethality in overdose has left them replaced by SSRIs in most cases. That said, they can be appropriate in refractory ADHD (desipramine^27,28) and refractory OCD (clomipramine is FDA-approved for this indication²⁹); they are considered a third-line treatment for enuresis.³⁰

Why did my patient stop the medication?

Common adverse effects. Although the greatest benefit of antidepressant medications compared with placebo is achieved relatively early on in treatment, it generally takes time for these benefits to accrue and become clinically apparent.^15,31 By contrast, most adverse effects of antidepressants present and are at their most severe early in treatment. The combination of early adverse effects and delayed efficacy leads many patients, families, and clinicians to discontinue medications before they have an adequate chance to work. Thus, it is imperative to provide psychoeducation before starting a medication about the typical time-course of improvement and adverse effects (Table 2).

Adverse effects of SSRIs often appear or worsen transiently during initiation of a medication, during a dose increase,³² or, theoretically, with the addition of a medication that interferes with SSRI metabolism (eg, cimetidine inhibition of cytochrome P450 2D6).³³ If families are prepared for this phenomenon and the therapeutic alliance is adequate, adverse effects can be tolerated to allow for a full medication trial. Common adverse effects of SSRIs include sleep problems (insomnia/sedation), gastrointestinal upset, sexual dysfunction, dry mouth, and hyperhidrosis. Although SSRIs differ somewhat in the frequency of these effects, as a class, they are more similar than different. Adequate psychoeducation is especially imperative in the treatment of OCD and anxiety disorders, where there is limited evidence of efficacy for any non-serotonergic antidepressants.

Serotonin-norepinephrine reuptake inhibitors are not considered first-line medications because of the reduced evidence base compared to SSRIs and their enhanced adverse effect profiles. Because SNRIs partially share a mechanism of action with SSRIs, they also share portions of the adverse effects profile. However, SNRIs have the additional adverse effect of hypertension, which is related to their noradrenergic activity. Thus, it is reasonable to obtain a baseline blood pressure before initiating an SNRI, as well as periodically after initiation and during dose increases, particularly if the patient has other risk factors for hypertension.³⁴

Continue to: Although TCAs have efficacy...

Although TCAs have efficacy in some pediatric disorders,^27-29,35 their adverse effect profile limits their use. Tricyclic antidepressants are highly anticholinergic (causing dizziness secondary to orthostatic hypotension, dry mouth, and urinary retention) and antihistaminergic (causing sedation and weight gain). Additionally, TCAs lower the seizure threshold and have adverse cardiac effects relating to their anti-alpha-1 adrenergic activity, resulting in dose-dependent increases in the QTc and cardiac toxicity in overdose that could lead to arrhythmia and death. These medications have their place, but their use requires careful informed consent, clear treatment goals, and baseline and periodic cardiac monitoring (via electrocardiogram).

Serious adverse effects. Clinicians may be hesitant to prescribe antidepressants for pediatric patients because of the potential for more serious adverse effects, including severe behavioral activation syndromes, serotonin syndrome, and emergent suicidality. However, current FDA-approved antidepressants arguably have one of the most positive risk/benefit profiles of any orally-administered medication approved for pediatric patients. Having a strong understanding of the evidence is critical to evaluating when it is appropriate to prescribe an antidepressant, how to properly monitor the patient, and how to obtain accurate informed consent.

Pediatric behavioral activation syndrome. Many clinicians report that children receiving antidepressants experience a pediatric behavioral activation syndrome, which exists along a spectrum from mild activation, increased energy, insomnia, or irritability up through more severe presentations of agitation, hyperactivity, or possibly mania. A recent meta-analysis suggested a positive association between antidepressant use and activation events on the milder end of this spectrum in pediatric patients with non-OCD anxiety disorders,¹⁶ and it is thought that compared with adolescents, younger children are more susceptible to activation adverse effects.³⁶ The likelihood of activation events has been associated with higher antidepressant plasma levels,³⁷ suggesting that dose or individual differences in metabolism may play a role. At the severe end of the spectrum, the risk of induction of mania in pediatric patients with depression or anxiety is relatively rare (<2%) and not statistically different from placebo in RCTs of pediatric participants.³⁸ Meta-analyses of larger randomized, placebo-controlled trials of adults do not support the idea that SSRIs and other second-generation antidepressants carry an increased risk of mania compared with placebo.^39,40 Children or adolescents with bona fide bipolar disorder (ie, patients who have had observed mania that meets all DSM-5 criteria) should be treated with a mood-stabilizing agent or antipsychotic if prescribed an antidepressant.⁴¹ These clear-cut cases are, however, relatively rare, and more often clinicians are confronted with ambiguous cases that include a family history of bipolar disorder along with “softer” symptoms of irritability, intrusiveness, or aggression. In these children, SSRIs may be appropriate for depressive, OCD, or anxiety symptoms, and should be strongly considered before prescribing antipsychotics or mood stabilizers, as long as initiated with proper monitoring.

Serotonin syndrome is a life-threatening condition caused by excess synaptic serotonin. It is characterized by confusion, sweating, diarrhea, hypertension, hyperthermia, and tachycardia. At its most severe, serotonin syndrome can result in seizures, arrhythmias, and death. The risk of serotonin syndrome is very low when using an SSRI as monotherapy. Risk increases with polypharmacy, particularly unexamined polypharmacy when multiple serotonergic agents are inadvertently on board. Commonly used serotonergic agents include other antidepressants, migraine medications (eg, triptans), some pain medications, and the cough suppressant dextromethorphan.

The easiest way to mitigate the risk of serotonin syndrome is to use an interaction index computer program, which can help ensure that the interacting agents are not prescribed without first discussing the risks and benefits. It is important to teach adolescents that certain recreational drugs are highly serotonergic and can cause serious interactions with antidepressants. For example, recreational use of dextrometh­orphan or 3,4-methylenedioxymethamphetamine (MDMA; commonly known as “ecstasy”) has been associated with serotonin syndrome in adolescents taking antidepressant medications.^42,43

Continue to: Suicidality

Suicidality. The black-box warning regarding a risk of emergent suicidality when starting antidepressant treatment in children is controversial.⁴⁴ The prospect that a medication intended to ameliorate depression might instead risk increasing suicidal thinking is alarming to parents and clinicians alike. To appropriately weigh and discuss the risks and benefits with families, it is important to understand the data upon which the warning is based.

In 2004, the FDA commissioned a review of 23 antidepressant trials, both published and unpublished, pooling studies across multiple indications (MDD, OCD, anxiety, and ADHD) and multiple antidepressant classes. This meta-analysis, which included nearly 4,400 pediatric patients, found a small but statistically significant increase in spontaneously-reported suicidal thoughts or actions, with a risk difference of 1% (95% confidence interval [CI], 1% to 2%).⁴⁵ These data suggest that if one treats 100 pediatric patients, 1 to 2 of them may experience short-term increases in suicidal thinking or behavior.⁴⁵ There were no differences in suicidal thinking when assessed systematically (ie, when all subjects reported symptoms of suicidal ideation on structured rating scales), and there were no completed suicides.⁴⁵ A subsequent analysis that included 27 pediatric trials suggested an even lower, although still significant, risk difference (<1%), yielding a number needed to harm (NNH) of 143.⁴⁶ Thus, with low NNT for efficacy (3 to 6) and relatively high NNH for emergent suicidal thoughts or behaviors (100 to 143), for many patients the benefits will outweigh the risks.

Figure 1, Figure 2, and Figure 3 are Cates plots that depict the absolute benefits of antidepressants compared with the risk of suicidality for pediatric patients with MDD, OCD, and anxiety disorders. Recent meta-analyses have suggested that the increased risk of suicidality in antidepressant trials is specific to studies that included children and adolescents, and is not observed in adult studies. A meta-analysis of 70 trials involving 18,526 participants suggested that the odds ratio of suicidality in trials of children and adolescents was 2.39 (95% CI, 1.31 to 4.33) compared with 0.81 (95% CI, 0.51 to 1.28) in adults.⁴⁷ Additionally, a network meta-analysis exclusively focusing on pediatric antidepressant trials in MDD reported significantly higher suicidality-related adverse events in venlafaxine trials compared with placebo, duloxetine, and several SSRIs (fluoxetine, paroxetine, and escitalopram).²⁰ These data should be interpreted with caution as differences in suicidality detected between agents is quite possibly related to differences in the method of assessment between trials, as opposed to actual differences in risk between agents.

Epidemiologic data further support the use of antidepressants in pediatric patients, showing that antidepressant use is associated with decreased teen suicide attempts and completions,⁴⁸ and the decline in prescriptions that occurred following the black-box warning was accompanied by a 14% increase in teen suicides.⁴⁹ Multiple hypotheses have been proposed to explain the pediatric clinical trial findings. One idea is that potential adverse effects of activation, or the intended effects of restoring the motivation, energy, and social engagement that is often impaired in depression, increases the likelihood of thinking about suicide or acting on thoughts. Another theory is that reporting of suicidality may be increased, rather than increased de novo suicidality itself. Antidepressants are effective for treating pediatric anxiety disorders, including social anxiety disorder,¹⁶ which could result in more willingness to report. Also, the manner in which adverse effects are generally ascertained in trials might have led to increased spontaneous reporting. In many trials, investigators ask whether participants have any adverse effects in general, and inquire about specific adverse effects only if the family answers affirmatively. Thus, the increased rate of other adverse effects associated with antidepressants (sleep problems, gastrointestinal upset, dry mouth, etc.) might trigger a specific question regarding suicidal ideation, which the child or family then may be more likely to report. Alternatively, any type of psychiatric treatment could increase an individual’s propensity to report; in adolescent psychotherapy trials, non-medicated participants have reported emergent suicidality at similar frequencies as those described in drug trials.⁵⁰ Regardless of the mechanism, the possibility of treatment-emergent suicidality is a low-frequency but serious event that necessitates careful monitoring when starting medication. Current guidelines suggest seeing children weekly for the first month after medication initiation, every 2 weeks for the following month, and monthly thereafter.⁵¹

Continue to: How long should the antidepressant be continued?

How long should the antidepressant be continued?

Many patients are concerned about how long they may be taking medication, and whether they will be taking an antidepressant “forever.” A treatment course can be broken into an acute phase, wherein remission is achieved and maintained for 6 to 8 weeks. This is followed by a continuation phase, with the goal of relapse prevention, lasting 16 to 20 weeks. The length of the last phase—the maintenance phase—depends both on the child’s history, the underlying therapeutic indication, the adverse effect burden experienced, and the family’s preferences/values. In general, for a first depressive episode, after treating for 1 year, a trial of discontinuation can be attempted with close monitoring. For a second depressive episode, we recommend 2 years of remission on antidepressant therapy before attempting discontinuation. In patients who have had 3 depressive episodes, or have had episodes of high severity, we recommend continuing antidepressant treatment indefinitely. Although much less well studied, the risk of relapse following SSRI discontinuation appears much more significant in OCD, whereas anxiety disorders and MDD have a relatively comparable risk.⁵²

In general, stopping an antidepressant should be done carefully and slowly. The speed with which a specific antidepressant can be stopped is largely related to its half-life. Agents with very long half-lives, such as fluoxetine (half-life of 5 days for the parent compound and 9 days for active metabolite), can often be stopped altogether, being “auto-tapered” by the long half-life. One might still consider a taper if the patient were taking high doses. Medications with shorter half-lives must be more carefully tapered to avoid discontinuation syndromes. Discontinuation syndromes are characterized by flu-like symptoms (nausea, myalgias, fatigue, dizziness) and worsening mood. Medications with short half-lives (eg, paroxetine and venlafaxine) have the highest potential for this syndrome in children,⁵³ and thus are used less frequently.

What to do when first-line treatments fail

When a child does not experience sufficient improvement from first-line treatments, it is crucial to determine whether they have experienced an adequate dosing, duration, and quality of medication and psychotherapy.

Adequate psychotherapy? To determine whether children are receiving adequate CBT, ask:

1. if the child receives homework from psychotherapy
2. if the parents are included in treatment
3. if therapy has involved identifying thought patterns that may be contributing to the child’s illness, and
4. if the therapist has ever exposed the child to a challenge likely to produce anxiety or distress in a supervised environment and has developed an exposure hierarchy (for conditions with primarily exposure-based therapies, such as OCD or anxiety disorders).

If a family is not receiving most of these elements in psychotherapy, this is a good indicator that they may not be receiving evidence-based CBT.

Continue to: Adequate pharmacotherapy?

Adequate pharmacotherapy? Similarly, when determining the adequacy of previous pharmacotherapy, it is critical to determine whether the child received an adequate dose of medications (at least the FDA-recommended minimum dose) for an adequate duration of time at therapeutic dosing (at least 6 weeks for MDD, 8 weeks for anxiety disorders, and 8 to 12 weeks for pediatric patients with OCD), and that the child actually took the medication regularly during that period. Patient compliance can typically be tracked through checking refill requests or intervals through the patient’s pharmacy. Ensuring proper delivery of first-line treatments is imperative because (1) the adverse effects associated with second-line treatments are often more substantial; (2) the cost in terms of time and money is considerably higher with second-line treatments, and; (3) the evidence regarding the benefits of these treatments is much less certain.

Inadequate dosing is a common reason for non-response in pediatric patients. Therapeutic dose ranges for common antidepressants are displayed in Table 1. Many clinicians underdose antidepressants for pediatric patients initially (and often throughout treatment) because they fear that the typical dose titration used in clinical trials will increase the risk of adverse effects compared with more conservative dosing. There is limited evidence to suggest that this underdosing strategy is likely to be successful; adverse effects attributable to these medications are modest, and most that are experienced early in treatment (eg, headache, increased anxiety or irritability, sleep problems, gastrointestinal upset) are self-limiting and may be coincidental rather than medication-induced. Furthermore, there is no evidence for efficacy of subtherapeutic dosing in children in the acute phase of treatment or for preventing relapse.¹⁴ Thus, from an efficacy standpoint, a medication trial has not officially begun until the therapeutic dose range is reached.

Once dosing is within the therapeutic range, however, pediatric data differs from the adult literature. In most adult psychi­atric conditions, higher doses of SSRIs within the therapeutic range are associated with an increased response rate.^14,54 In pediatrics, there are few fixed dose trials, and once within the recommended therapeutic range, minimal data supports an association between higher dosing and higher efficacy.¹⁴ In general, pediatric guidelines are silent regarding optimal dosing of SSRIs within the recommended dose range, and higher antidepressant doses often result in a more significant adverse effect burden for children. One exception is pediatric OCD, where, similar to adults, the guidelines suggest that higher dosing of SSRIs often is required to induce a therapeutic response as compared to MDD and GAD.^31,55

If a child does not respond to adequate first-line treatment (or has a treatment history that cannot be fully verified), repeating these first-line interventions carries little risk and can be quite effective. For example, 60% of adolescents with MDD who did not initially respond to an SSRI demonstrated a significant response when prescribed a second SSRI or venlafaxine (with or without CBT).⁵⁶

When pediatric patients continue to experience significantly distressing and/or debilitating symptoms (particularly in MDD) despite multiple trials of antidepressants and psychotherapy, practitioners should consider a careful referral to interventional psychiatry services, which can include the more intensive treatments of electroconvulsive therapy, repetitive transcranial magnetic stimulation, or ketamine (see Box 1). Given the substantial morbidity and mortality associated with adolescent depression, interventional psychiatry treatments are under-researched and under-utilized clinically in pediatric populations.

Continue to: Antidepressants in general...

Antidepressants in general, and SSRIs in particular, are the first-line pharmacotherapy for pediatric anxiety, OCD, and MDD. For PTSD, although they are a first-line treatment in adults, their efficacy has not been demonstrated in children and adolescents. Antidepressants are generally safe, well-tolerated, and effective, with low NNTs (3 to 5 for anxiety and OCD; 4 to 12 in MDD, depending on whether industry trials are included). It is important that clinicians and families be educated about possible adverse effects and their time course in order to anticipate difficulties, ensure adequate informed consent, and monitor appropriately. The black-box warning regarding treatment-emergent suicidal thoughts or behaviors must be discussed (for suggested talking points, see Box 2). The NNH is large (100 to 143), and for many patients, the benefits will outweigh the risks. For pediatric patients who fail to respond to multiple adequate trials of antidepressants and psychotherapy, referrals for interventional psychiatry consultation should be considered.

Bottom Line

Although the evidence base for prescribing antidepressants for children and adolescents is smaller compared to the adult literature, properly understanding and prescribing these agents, and explaining their risks and benefits to families, can make a major difference in patient compliance, satisfaction, and outcomes. Antidepressants (particularly selective serotonin reuptake inhibitors) are the firstline pharmacologic intervention for pediatric patients with anxiety disorders, obsessive-compulsive disorder, or major depressive disorder.

Related Resource

• Bonin L, Moreland CS. Overview of prevention and treatment for pediatric depression. https://www.uptodate.com/contents/overview-of-prevention-and-treatment-for-pediatric-depression. Last reviewed July 2019.

Drug Brand Names

Bupropion • Wellbutrin, Zyban
Cimetidine • Tagamet
Citalopram • Celexa
Clomipramine • Anafranil
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Escitalopram • Lexapro
Fluoxetine • Prozac
Fluvoxamine • Luvox
Imipramine • Tofranil
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Sertraline • Zoloft
Venlafaxine • Effexor
Vilazodone • Viibryd
Vortioxetine • Trintellix

Box 1

Box 2

Major depressive disorder (MDD) is a significant pediatric health problem, with a lifetime prevalence as high as 20% by the end of adolescence.^1-3 Major depressive disorder in adolescence is associated with significant morbidity, including poor social functioning, school difficulties, early pregnancy, and increased risk of physical illness and substance abuse.^4-6 It is also linked with significant mortality, with increased risk for suicide, which is now the second leading cause of death in individuals age 10 to 24 years.^1,7,8

As their name suggests, antidepressants comprise a group of medications that are used to treat MDD; they are also, however, first-line agents for generalized anxiety disorder (GAD), posttraumatic stress disorder (PTSD), and obsessive-compulsive disorder (OCD) in adults. Anxiety disorders (including GAD and other anxiety diagnoses) and PTSD are also common in childhood and adolescence with a combined lifetime prevalence ranging from 15% to 30%.^9,10 These disorders are also associated with increased risk of suicide.¹¹ For all of these disorders, depending on the severity of presentation and the preference of the patient, treatments are often a combination of psychotherapy and psychopharmacology.

Clinicians face several challenges when considering antidepressants for pediatric patients. Pediatricians and psychiatrists need to understand whether these medications work in children and adolescents, and whether there are unique developmental safety and tolerability issues. The evidence base in child psychiatry is considerably smaller compared with that of adult psychiatry. From this more limited evidence base also came the controversial “black-box” warning regarding a risk of emergent suicidality when starting antidepressants that accompanies all antidepressants for pediatric, but not adult, patients. This warning has had major effects on clinical encounters with children experiencing depression, including altering clinician prescribing behavior.¹²

In this article, we review the current evidence for antidepressant efficacy, tolerability, and safety in pediatric patients. We also suggest ways in which clinicians might choose, start, and stop antidepressants in children, as well as how to talk with parents about benefits, risks, and the black-box warning.

Do antidepressants work in children?

Selective serotonin reuptake inhibitors. Selective serotonin reuptake inhibitors (SSRIs) are the most commonly used class of antidepressants in both children and adults.¹³ While only a few SSRIs are FDA-approved for pediatric indications, the lack of FDA approval is typically related to a lack of sufficient testing in randomized controlled trials (RCTs) for specific pediatric indications, rather than to demonstrable differences in efficacy between antidepressant agents. Since there is currently no data to suggest inferiority of one agent compared to another in children or adults,^14,15 efficacy data will be discussed here as applied to the class of SSRIs, generalizing from RCTs conducted on individual drugs. Table 1 lists FDA indications and dosing information for individual antidepressants.

There is strong evidence that SSRIs are effective for treating pediatric anxiety disorders (eg, social anxiety disorder and GAD)¹⁶ and OCD,¹⁷ with numbers needed to treat (NNT) between 3 and 5. For both of these disorders, SSRIs combined with cognitive-behavioral therapy (CBT) have the highest likelihood of improving symptoms or achieving remission.^17,18

Selective serotonin reuptake inhibitors are also effective for treating pediatric MDD; however, the literature is more complex for this disorder compared to GAD and OCD as there are considerable differences in effect sizes between National Institute of Mental Health (NIMH)–funded studies and industry-sponsored trials.¹³ The major NIMH-sponsored adolescent depression trial, TADS (Treatment for Adolescents and Depression Study), showed that SSRIs (fluoxetine in this case) were quite effective, with an NNT of 4 over the acute phase (12 weeks).¹⁹ Ultimately, approximately 80% of adolescents improved over 9 months. Many industry-sponsored trials for MDD in pediatric patients had large placebo response rates (approximately 60%), which resulted in smaller between-group differences, and estimates of an NNT closer to 12,¹³ which has muddied the waters in meta-analyses that include all trials.²⁰ Improvement in depressive symptoms also appears to be bolstered by concomitant CBT in MDD,¹⁹ but not as robustly as in GAD and OCD. While the full benefit of SSRIs for depression may take as long as 8 weeks, a meta-analysis of depression studies of pediatric patients suggests that significant benefits from placebo are observed as early as 2 weeks, and that further treatment gains are minimal after 4 weeks.¹⁵ Thus, we recommend at least a 4- to 6-week trial at therapeutic dosing before deeming a medication a treatment failure.

Continue to: Posttraumatic stress disorder...

Posttraumatic stress disorder is a fourth disorder in which SSRIs are a first-line treatment in adults. The data for using SSRIs to treat pediatric patients with PTSD is scant, with only a few RCTs, and no large NIMH-funded trials. Randomized controlled trials have not demonstrated significant differences between SSRIs and placebo^21,22 and thus the current first-line recommendation in pediatric PTSD remains trauma-focused therapy, with good evidence for trauma-focused CBT.²³ Practically speaking, there can be considerable overlap of PTSD, depression, and anxiety symptoms in children,²³ and children with a history of trauma who also have comorbid MDD may benefit from medication if their symptoms persist despite an adequate trial of psychotherapy.

Taken together, the current evidence suggests that SSRIs are often effective in pediatric GAD, OCD, and MDD, with low NNTs (ranging from 3 to 5 based on NIMH-funded trials) for all of these disorders; there is not yet sufficient evidence of efficacy in pediatric patients with PTSD.

Fluoxetine has been studied more intensively than other SSRIs (for example, it was the antidepressant used in the TADS trial), and thus has the largest evidence base. For this reason, fluoxetine is often considered the first of the first-line options. Additionally, fluoxetine has a longer half-life than other antidepressants, which may make it more effective in situations where patients are likely to miss doses, and results in a lower risk of withdrawal symptoms when stopped due to “self-tapering.”

SNRIs and atypical antidepressants. Other antidepressants commonly used in pediatric patients but with far less evidence of efficacy include serotonin-norepinephrine reuptake inhibitors (SNRIs) and the atypical antidepressants bupropion and mirtazapine. The SNRI duloxetine is FDA-approved for treating GAD in children age 7 to 17, but there are no other pediatric indications for duloxetine, or for the other SNRIs.

In general, adverse effect profiles are worse for SNRIs compared to SSRIs, further limiting their utility. While there are no pediatric studies demonstrating SNRI efficacy for neuropathic pain, good data exists in adults.²⁴ Thus, an SNRI could be a reasonable option if a pediatric patient has failed prior adequate SSRI trials and also has comorbid neuropathic pain.

Continue to: Neither bupropion nor mirtazapine...

Neither bupropion nor mirtazapine have undergone rigorous testing in pediatric patients, and therefore these agents should generally be considered only once other first-line treatments have failed. Bupropion has been evaluated for attention-deficit/hyperactivity disorder (ADHD)²⁵ and for adolescent smoking cessation.²⁶ However, the evidence is weak, and bupropion is not considered a first-line option for children and adolescents.

Tricyclic antidepressants. Randomized controlled trials have demonstrated that tricyclic antidepressants (TCAs) are efficacious for treating several pediatric conditions; however, their significant side effect profile, their monitoring requirements, as well as their lethality in overdose has left them replaced by SSRIs in most cases. That said, they can be appropriate in refractory ADHD (desipramine^27,28) and refractory OCD (clomipramine is FDA-approved for this indication²⁹); they are considered a third-line treatment for enuresis.³⁰

Why did my patient stop the medication?

Common adverse effects. Although the greatest benefit of antidepressant medications compared with placebo is achieved relatively early on in treatment, it generally takes time for these benefits to accrue and become clinically apparent.^15,31 By contrast, most adverse effects of antidepressants present and are at their most severe early in treatment. The combination of early adverse effects and delayed efficacy leads many patients, families, and clinicians to discontinue medications before they have an adequate chance to work. Thus, it is imperative to provide psychoeducation before starting a medication about the typical time-course of improvement and adverse effects (Table 2).

Adverse effects of SSRIs often appear or worsen transiently during initiation of a medication, during a dose increase,³² or, theoretically, with the addition of a medication that interferes with SSRI metabolism (eg, cimetidine inhibition of cytochrome P450 2D6).³³ If families are prepared for this phenomenon and the therapeutic alliance is adequate, adverse effects can be tolerated to allow for a full medication trial. Common adverse effects of SSRIs include sleep problems (insomnia/sedation), gastrointestinal upset, sexual dysfunction, dry mouth, and hyperhidrosis. Although SSRIs differ somewhat in the frequency of these effects, as a class, they are more similar than different. Adequate psychoeducation is especially imperative in the treatment of OCD and anxiety disorders, where there is limited evidence of efficacy for any non-serotonergic antidepressants.

Serotonin-norepinephrine reuptake inhibitors are not considered first-line medications because of the reduced evidence base compared to SSRIs and their enhanced adverse effect profiles. Because SNRIs partially share a mechanism of action with SSRIs, they also share portions of the adverse effects profile. However, SNRIs have the additional adverse effect of hypertension, which is related to their noradrenergic activity. Thus, it is reasonable to obtain a baseline blood pressure before initiating an SNRI, as well as periodically after initiation and during dose increases, particularly if the patient has other risk factors for hypertension.³⁴

Continue to: Although TCAs have efficacy...

Although TCAs have efficacy in some pediatric disorders,^27-29,35 their adverse effect profile limits their use. Tricyclic antidepressants are highly anticholinergic (causing dizziness secondary to orthostatic hypotension, dry mouth, and urinary retention) and antihistaminergic (causing sedation and weight gain). Additionally, TCAs lower the seizure threshold and have adverse cardiac effects relating to their anti-alpha-1 adrenergic activity, resulting in dose-dependent increases in the QTc and cardiac toxicity in overdose that could lead to arrhythmia and death. These medications have their place, but their use requires careful informed consent, clear treatment goals, and baseline and periodic cardiac monitoring (via electrocardiogram).

Serious adverse effects. Clinicians may be hesitant to prescribe antidepressants for pediatric patients because of the potential for more serious adverse effects, including severe behavioral activation syndromes, serotonin syndrome, and emergent suicidality. However, current FDA-approved antidepressants arguably have one of the most positive risk/benefit profiles of any orally-administered medication approved for pediatric patients. Having a strong understanding of the evidence is critical to evaluating when it is appropriate to prescribe an antidepressant, how to properly monitor the patient, and how to obtain accurate informed consent.

Pediatric behavioral activation syndrome. Many clinicians report that children receiving antidepressants experience a pediatric behavioral activation syndrome, which exists along a spectrum from mild activation, increased energy, insomnia, or irritability up through more severe presentations of agitation, hyperactivity, or possibly mania. A recent meta-analysis suggested a positive association between antidepressant use and activation events on the milder end of this spectrum in pediatric patients with non-OCD anxiety disorders,¹⁶ and it is thought that compared with adolescents, younger children are more susceptible to activation adverse effects.³⁶ The likelihood of activation events has been associated with higher antidepressant plasma levels,³⁷ suggesting that dose or individual differences in metabolism may play a role. At the severe end of the spectrum, the risk of induction of mania in pediatric patients with depression or anxiety is relatively rare (<2%) and not statistically different from placebo in RCTs of pediatric participants.³⁸ Meta-analyses of larger randomized, placebo-controlled trials of adults do not support the idea that SSRIs and other second-generation antidepressants carry an increased risk of mania compared with placebo.^39,40 Children or adolescents with bona fide bipolar disorder (ie, patients who have had observed mania that meets all DSM-5 criteria) should be treated with a mood-stabilizing agent or antipsychotic if prescribed an antidepressant.⁴¹ These clear-cut cases are, however, relatively rare, and more often clinicians are confronted with ambiguous cases that include a family history of bipolar disorder along with “softer” symptoms of irritability, intrusiveness, or aggression. In these children, SSRIs may be appropriate for depressive, OCD, or anxiety symptoms, and should be strongly considered before prescribing antipsychotics or mood stabilizers, as long as initiated with proper monitoring.

Serotonin syndrome is a life-threatening condition caused by excess synaptic serotonin. It is characterized by confusion, sweating, diarrhea, hypertension, hyperthermia, and tachycardia. At its most severe, serotonin syndrome can result in seizures, arrhythmias, and death. The risk of serotonin syndrome is very low when using an SSRI as monotherapy. Risk increases with polypharmacy, particularly unexamined polypharmacy when multiple serotonergic agents are inadvertently on board. Commonly used serotonergic agents include other antidepressants, migraine medications (eg, triptans), some pain medications, and the cough suppressant dextromethorphan.

The easiest way to mitigate the risk of serotonin syndrome is to use an interaction index computer program, which can help ensure that the interacting agents are not prescribed without first discussing the risks and benefits. It is important to teach adolescents that certain recreational drugs are highly serotonergic and can cause serious interactions with antidepressants. For example, recreational use of dextrometh­orphan or 3,4-methylenedioxymethamphetamine (MDMA; commonly known as “ecstasy”) has been associated with serotonin syndrome in adolescents taking antidepressant medications.^42,43

Continue to: Suicidality

Suicidality. The black-box warning regarding a risk of emergent suicidality when starting antidepressant treatment in children is controversial.⁴⁴ The prospect that a medication intended to ameliorate depression might instead risk increasing suicidal thinking is alarming to parents and clinicians alike. To appropriately weigh and discuss the risks and benefits with families, it is important to understand the data upon which the warning is based.

In 2004, the FDA commissioned a review of 23 antidepressant trials, both published and unpublished, pooling studies across multiple indications (MDD, OCD, anxiety, and ADHD) and multiple antidepressant classes. This meta-analysis, which included nearly 4,400 pediatric patients, found a small but statistically significant increase in spontaneously-reported suicidal thoughts or actions, with a risk difference of 1% (95% confidence interval [CI], 1% to 2%).⁴⁵ These data suggest that if one treats 100 pediatric patients, 1 to 2 of them may experience short-term increases in suicidal thinking or behavior.⁴⁵ There were no differences in suicidal thinking when assessed systematically (ie, when all subjects reported symptoms of suicidal ideation on structured rating scales), and there were no completed suicides.⁴⁵ A subsequent analysis that included 27 pediatric trials suggested an even lower, although still significant, risk difference (<1%), yielding a number needed to harm (NNH) of 143.⁴⁶ Thus, with low NNT for efficacy (3 to 6) and relatively high NNH for emergent suicidal thoughts or behaviors (100 to 143), for many patients the benefits will outweigh the risks.

Figure 1, Figure 2, and Figure 3 are Cates plots that depict the absolute benefits of antidepressants compared with the risk of suicidality for pediatric patients with MDD, OCD, and anxiety disorders. Recent meta-analyses have suggested that the increased risk of suicidality in antidepressant trials is specific to studies that included children and adolescents, and is not observed in adult studies. A meta-analysis of 70 trials involving 18,526 participants suggested that the odds ratio of suicidality in trials of children and adolescents was 2.39 (95% CI, 1.31 to 4.33) compared with 0.81 (95% CI, 0.51 to 1.28) in adults.⁴⁷ Additionally, a network meta-analysis exclusively focusing on pediatric antidepressant trials in MDD reported significantly higher suicidality-related adverse events in venlafaxine trials compared with placebo, duloxetine, and several SSRIs (fluoxetine, paroxetine, and escitalopram).²⁰ These data should be interpreted with caution as differences in suicidality detected between agents is quite possibly related to differences in the method of assessment between trials, as opposed to actual differences in risk between agents.

Epidemiologic data further support the use of antidepressants in pediatric patients, showing that antidepressant use is associated with decreased teen suicide attempts and completions,⁴⁸ and the decline in prescriptions that occurred following the black-box warning was accompanied by a 14% increase in teen suicides.⁴⁹ Multiple hypotheses have been proposed to explain the pediatric clinical trial findings. One idea is that potential adverse effects of activation, or the intended effects of restoring the motivation, energy, and social engagement that is often impaired in depression, increases the likelihood of thinking about suicide or acting on thoughts. Another theory is that reporting of suicidality may be increased, rather than increased de novo suicidality itself. Antidepressants are effective for treating pediatric anxiety disorders, including social anxiety disorder,¹⁶ which could result in more willingness to report. Also, the manner in which adverse effects are generally ascertained in trials might have led to increased spontaneous reporting. In many trials, investigators ask whether participants have any adverse effects in general, and inquire about specific adverse effects only if the family answers affirmatively. Thus, the increased rate of other adverse effects associated with antidepressants (sleep problems, gastrointestinal upset, dry mouth, etc.) might trigger a specific question regarding suicidal ideation, which the child or family then may be more likely to report. Alternatively, any type of psychiatric treatment could increase an individual’s propensity to report; in adolescent psychotherapy trials, non-medicated participants have reported emergent suicidality at similar frequencies as those described in drug trials.⁵⁰ Regardless of the mechanism, the possibility of treatment-emergent suicidality is a low-frequency but serious event that necessitates careful monitoring when starting medication. Current guidelines suggest seeing children weekly for the first month after medication initiation, every 2 weeks for the following month, and monthly thereafter.⁵¹

Continue to: How long should the antidepressant be continued?

How long should the antidepressant be continued?

Many patients are concerned about how long they may be taking medication, and whether they will be taking an antidepressant “forever.” A treatment course can be broken into an acute phase, wherein remission is achieved and maintained for 6 to 8 weeks. This is followed by a continuation phase, with the goal of relapse prevention, lasting 16 to 20 weeks. The length of the last phase—the maintenance phase—depends both on the child’s history, the underlying therapeutic indication, the adverse effect burden experienced, and the family’s preferences/values. In general, for a first depressive episode, after treating for 1 year, a trial of discontinuation can be attempted with close monitoring. For a second depressive episode, we recommend 2 years of remission on antidepressant therapy before attempting discontinuation. In patients who have had 3 depressive episodes, or have had episodes of high severity, we recommend continuing antidepressant treatment indefinitely. Although much less well studied, the risk of relapse following SSRI discontinuation appears much more significant in OCD, whereas anxiety disorders and MDD have a relatively comparable risk.⁵²

In general, stopping an antidepressant should be done carefully and slowly. The speed with which a specific antidepressant can be stopped is largely related to its half-life. Agents with very long half-lives, such as fluoxetine (half-life of 5 days for the parent compound and 9 days for active metabolite), can often be stopped altogether, being “auto-tapered” by the long half-life. One might still consider a taper if the patient were taking high doses. Medications with shorter half-lives must be more carefully tapered to avoid discontinuation syndromes. Discontinuation syndromes are characterized by flu-like symptoms (nausea, myalgias, fatigue, dizziness) and worsening mood. Medications with short half-lives (eg, paroxetine and venlafaxine) have the highest potential for this syndrome in children,⁵³ and thus are used less frequently.

What to do when first-line treatments fail

When a child does not experience sufficient improvement from first-line treatments, it is crucial to determine whether they have experienced an adequate dosing, duration, and quality of medication and psychotherapy.

Adequate psychotherapy? To determine whether children are receiving adequate CBT, ask:

1. if the child receives homework from psychotherapy
2. if the parents are included in treatment
3. if therapy has involved identifying thought patterns that may be contributing to the child’s illness, and
4. if the therapist has ever exposed the child to a challenge likely to produce anxiety or distress in a supervised environment and has developed an exposure hierarchy (for conditions with primarily exposure-based therapies, such as OCD or anxiety disorders).

If a family is not receiving most of these elements in psychotherapy, this is a good indicator that they may not be receiving evidence-based CBT.

Continue to: Adequate pharmacotherapy?

Adequate pharmacotherapy? Similarly, when determining the adequacy of previous pharmacotherapy, it is critical to determine whether the child received an adequate dose of medications (at least the FDA-recommended minimum dose) for an adequate duration of time at therapeutic dosing (at least 6 weeks for MDD, 8 weeks for anxiety disorders, and 8 to 12 weeks for pediatric patients with OCD), and that the child actually took the medication regularly during that period. Patient compliance can typically be tracked through checking refill requests or intervals through the patient’s pharmacy. Ensuring proper delivery of first-line treatments is imperative because (1) the adverse effects associated with second-line treatments are often more substantial; (2) the cost in terms of time and money is considerably higher with second-line treatments, and; (3) the evidence regarding the benefits of these treatments is much less certain.

Inadequate dosing is a common reason for non-response in pediatric patients. Therapeutic dose ranges for common antidepressants are displayed in Table 1. Many clinicians underdose antidepressants for pediatric patients initially (and often throughout treatment) because they fear that the typical dose titration used in clinical trials will increase the risk of adverse effects compared with more conservative dosing. There is limited evidence to suggest that this underdosing strategy is likely to be successful; adverse effects attributable to these medications are modest, and most that are experienced early in treatment (eg, headache, increased anxiety or irritability, sleep problems, gastrointestinal upset) are self-limiting and may be coincidental rather than medication-induced. Furthermore, there is no evidence for efficacy of subtherapeutic dosing in children in the acute phase of treatment or for preventing relapse.¹⁴ Thus, from an efficacy standpoint, a medication trial has not officially begun until the therapeutic dose range is reached.

Once dosing is within the therapeutic range, however, pediatric data differs from the adult literature. In most adult psychi­atric conditions, higher doses of SSRIs within the therapeutic range are associated with an increased response rate.^14,54 In pediatrics, there are few fixed dose trials, and once within the recommended therapeutic range, minimal data supports an association between higher dosing and higher efficacy.¹⁴ In general, pediatric guidelines are silent regarding optimal dosing of SSRIs within the recommended dose range, and higher antidepressant doses often result in a more significant adverse effect burden for children. One exception is pediatric OCD, where, similar to adults, the guidelines suggest that higher dosing of SSRIs often is required to induce a therapeutic response as compared to MDD and GAD.^31,55

If a child does not respond to adequate first-line treatment (or has a treatment history that cannot be fully verified), repeating these first-line interventions carries little risk and can be quite effective. For example, 60% of adolescents with MDD who did not initially respond to an SSRI demonstrated a significant response when prescribed a second SSRI or venlafaxine (with or without CBT).⁵⁶

When pediatric patients continue to experience significantly distressing and/or debilitating symptoms (particularly in MDD) despite multiple trials of antidepressants and psychotherapy, practitioners should consider a careful referral to interventional psychiatry services, which can include the more intensive treatments of electroconvulsive therapy, repetitive transcranial magnetic stimulation, or ketamine (see Box 1). Given the substantial morbidity and mortality associated with adolescent depression, interventional psychiatry treatments are under-researched and under-utilized clinically in pediatric populations.

Continue to: Antidepressants in general...

Antidepressants in general, and SSRIs in particular, are the first-line pharmacotherapy for pediatric anxiety, OCD, and MDD. For PTSD, although they are a first-line treatment in adults, their efficacy has not been demonstrated in children and adolescents. Antidepressants are generally safe, well-tolerated, and effective, with low NNTs (3 to 5 for anxiety and OCD; 4 to 12 in MDD, depending on whether industry trials are included). It is important that clinicians and families be educated about possible adverse effects and their time course in order to anticipate difficulties, ensure adequate informed consent, and monitor appropriately. The black-box warning regarding treatment-emergent suicidal thoughts or behaviors must be discussed (for suggested talking points, see Box 2). The NNH is large (100 to 143), and for many patients, the benefits will outweigh the risks. For pediatric patients who fail to respond to multiple adequate trials of antidepressants and psychotherapy, referrals for interventional psychiatry consultation should be considered.

Bottom Line

Although the evidence base for prescribing antidepressants for children and adolescents is smaller compared to the adult literature, properly understanding and prescribing these agents, and explaining their risks and benefits to families, can make a major difference in patient compliance, satisfaction, and outcomes. Antidepressants (particularly selective serotonin reuptake inhibitors) are the firstline pharmacologic intervention for pediatric patients with anxiety disorders, obsessive-compulsive disorder, or major depressive disorder.

Related Resource

• Bonin L, Moreland CS. Overview of prevention and treatment for pediatric depression. https://www.uptodate.com/contents/overview-of-prevention-and-treatment-for-pediatric-depression. Last reviewed July 2019.

Drug Brand Names

Bupropion • Wellbutrin, Zyban
Cimetidine • Tagamet
Citalopram • Celexa
Clomipramine • Anafranil
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Escitalopram • Lexapro
Fluoxetine • Prozac
Fluvoxamine • Luvox
Imipramine • Tofranil
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Sertraline • Zoloft
Venlafaxine • Effexor
Vilazodone • Viibryd
Vortioxetine • Trintellix

Box 1

Box 2

Major depressive disorder (MDD) is a significant pediatric health problem, with a lifetime prevalence as high as 20% by the end of adolescence.^1-3 Major depressive disorder in adolescence is associated with significant morbidity, including poor social functioning, school difficulties, early pregnancy, and increased risk of physical illness and substance abuse.^4-6 It is also linked with significant mortality, with increased risk for suicide, which is now the second leading cause of death in individuals age 10 to 24 years.^1,7,8

As their name suggests, antidepressants comprise a group of medications that are used to treat MDD; they are also, however, first-line agents for generalized anxiety disorder (GAD), posttraumatic stress disorder (PTSD), and obsessive-compulsive disorder (OCD) in adults. Anxiety disorders (including GAD and other anxiety diagnoses) and PTSD are also common in childhood and adolescence with a combined lifetime prevalence ranging from 15% to 30%.^9,10 These disorders are also associated with increased risk of suicide.¹¹ For all of these disorders, depending on the severity of presentation and the preference of the patient, treatments are often a combination of psychotherapy and psychopharmacology.

Clinicians face several challenges when considering antidepressants for pediatric patients. Pediatricians and psychiatrists need to understand whether these medications work in children and adolescents, and whether there are unique developmental safety and tolerability issues. The evidence base in child psychiatry is considerably smaller compared with that of adult psychiatry. From this more limited evidence base also came the controversial “black-box” warning regarding a risk of emergent suicidality when starting antidepressants that accompanies all antidepressants for pediatric, but not adult, patients. This warning has had major effects on clinical encounters with children experiencing depression, including altering clinician prescribing behavior.¹²

In this article, we review the current evidence for antidepressant efficacy, tolerability, and safety in pediatric patients. We also suggest ways in which clinicians might choose, start, and stop antidepressants in children, as well as how to talk with parents about benefits, risks, and the black-box warning.

Do antidepressants work in children?

Selective serotonin reuptake inhibitors. Selective serotonin reuptake inhibitors (SSRIs) are the most commonly used class of antidepressants in both children and adults.¹³ While only a few SSRIs are FDA-approved for pediatric indications, the lack of FDA approval is typically related to a lack of sufficient testing in randomized controlled trials (RCTs) for specific pediatric indications, rather than to demonstrable differences in efficacy between antidepressant agents. Since there is currently no data to suggest inferiority of one agent compared to another in children or adults,^14,15 efficacy data will be discussed here as applied to the class of SSRIs, generalizing from RCTs conducted on individual drugs. Table 1 lists FDA indications and dosing information for individual antidepressants.

There is strong evidence that SSRIs are effective for treating pediatric anxiety disorders (eg, social anxiety disorder and GAD)¹⁶ and OCD,¹⁷ with numbers needed to treat (NNT) between 3 and 5. For both of these disorders, SSRIs combined with cognitive-behavioral therapy (CBT) have the highest likelihood of improving symptoms or achieving remission.^17,18

Selective serotonin reuptake inhibitors are also effective for treating pediatric MDD; however, the literature is more complex for this disorder compared to GAD and OCD as there are considerable differences in effect sizes between National Institute of Mental Health (NIMH)–funded studies and industry-sponsored trials.¹³ The major NIMH-sponsored adolescent depression trial, TADS (Treatment for Adolescents and Depression Study), showed that SSRIs (fluoxetine in this case) were quite effective, with an NNT of 4 over the acute phase (12 weeks).¹⁹ Ultimately, approximately 80% of adolescents improved over 9 months. Many industry-sponsored trials for MDD in pediatric patients had large placebo response rates (approximately 60%), which resulted in smaller between-group differences, and estimates of an NNT closer to 12,¹³ which has muddied the waters in meta-analyses that include all trials.²⁰ Improvement in depressive symptoms also appears to be bolstered by concomitant CBT in MDD,¹⁹ but not as robustly as in GAD and OCD. While the full benefit of SSRIs for depression may take as long as 8 weeks, a meta-analysis of depression studies of pediatric patients suggests that significant benefits from placebo are observed as early as 2 weeks, and that further treatment gains are minimal after 4 weeks.¹⁵ Thus, we recommend at least a 4- to 6-week trial at therapeutic dosing before deeming a medication a treatment failure.

Continue to: Posttraumatic stress disorder...

Posttraumatic stress disorder is a fourth disorder in which SSRIs are a first-line treatment in adults. The data for using SSRIs to treat pediatric patients with PTSD is scant, with only a few RCTs, and no large NIMH-funded trials. Randomized controlled trials have not demonstrated significant differences between SSRIs and placebo^21,22 and thus the current first-line recommendation in pediatric PTSD remains trauma-focused therapy, with good evidence for trauma-focused CBT.²³ Practically speaking, there can be considerable overlap of PTSD, depression, and anxiety symptoms in children,²³ and children with a history of trauma who also have comorbid MDD may benefit from medication if their symptoms persist despite an adequate trial of psychotherapy.

Taken together, the current evidence suggests that SSRIs are often effective in pediatric GAD, OCD, and MDD, with low NNTs (ranging from 3 to 5 based on NIMH-funded trials) for all of these disorders; there is not yet sufficient evidence of efficacy in pediatric patients with PTSD.

Fluoxetine has been studied more intensively than other SSRIs (for example, it was the antidepressant used in the TADS trial), and thus has the largest evidence base. For this reason, fluoxetine is often considered the first of the first-line options. Additionally, fluoxetine has a longer half-life than other antidepressants, which may make it more effective in situations where patients are likely to miss doses, and results in a lower risk of withdrawal symptoms when stopped due to “self-tapering.”

SNRIs and atypical antidepressants. Other antidepressants commonly used in pediatric patients but with far less evidence of efficacy include serotonin-norepinephrine reuptake inhibitors (SNRIs) and the atypical antidepressants bupropion and mirtazapine. The SNRI duloxetine is FDA-approved for treating GAD in children age 7 to 17, but there are no other pediatric indications for duloxetine, or for the other SNRIs.

In general, adverse effect profiles are worse for SNRIs compared to SSRIs, further limiting their utility. While there are no pediatric studies demonstrating SNRI efficacy for neuropathic pain, good data exists in adults.²⁴ Thus, an SNRI could be a reasonable option if a pediatric patient has failed prior adequate SSRI trials and also has comorbid neuropathic pain.

Continue to: Neither bupropion nor mirtazapine...

Neither bupropion nor mirtazapine have undergone rigorous testing in pediatric patients, and therefore these agents should generally be considered only once other first-line treatments have failed. Bupropion has been evaluated for attention-deficit/hyperactivity disorder (ADHD)²⁵ and for adolescent smoking cessation.²⁶ However, the evidence is weak, and bupropion is not considered a first-line option for children and adolescents.

Tricyclic antidepressants. Randomized controlled trials have demonstrated that tricyclic antidepressants (TCAs) are efficacious for treating several pediatric conditions; however, their significant side effect profile, their monitoring requirements, as well as their lethality in overdose has left them replaced by SSRIs in most cases. That said, they can be appropriate in refractory ADHD (desipramine^27,28) and refractory OCD (clomipramine is FDA-approved for this indication²⁹); they are considered a third-line treatment for enuresis.³⁰

Why did my patient stop the medication?

Common adverse effects. Although the greatest benefit of antidepressant medications compared with placebo is achieved relatively early on in treatment, it generally takes time for these benefits to accrue and become clinically apparent.^15,31 By contrast, most adverse effects of antidepressants present and are at their most severe early in treatment. The combination of early adverse effects and delayed efficacy leads many patients, families, and clinicians to discontinue medications before they have an adequate chance to work. Thus, it is imperative to provide psychoeducation before starting a medication about the typical time-course of improvement and adverse effects (Table 2).

Adverse effects of SSRIs often appear or worsen transiently during initiation of a medication, during a dose increase,³² or, theoretically, with the addition of a medication that interferes with SSRI metabolism (eg, cimetidine inhibition of cytochrome P450 2D6).³³ If families are prepared for this phenomenon and the therapeutic alliance is adequate, adverse effects can be tolerated to allow for a full medication trial. Common adverse effects of SSRIs include sleep problems (insomnia/sedation), gastrointestinal upset, sexual dysfunction, dry mouth, and hyperhidrosis. Although SSRIs differ somewhat in the frequency of these effects, as a class, they are more similar than different. Adequate psychoeducation is especially imperative in the treatment of OCD and anxiety disorders, where there is limited evidence of efficacy for any non-serotonergic antidepressants.

Serotonin-norepinephrine reuptake inhibitors are not considered first-line medications because of the reduced evidence base compared to SSRIs and their enhanced adverse effect profiles. Because SNRIs partially share a mechanism of action with SSRIs, they also share portions of the adverse effects profile. However, SNRIs have the additional adverse effect of hypertension, which is related to their noradrenergic activity. Thus, it is reasonable to obtain a baseline blood pressure before initiating an SNRI, as well as periodically after initiation and during dose increases, particularly if the patient has other risk factors for hypertension.³⁴

Continue to: Although TCAs have efficacy...

Although TCAs have efficacy in some pediatric disorders,^27-29,35 their adverse effect profile limits their use. Tricyclic antidepressants are highly anticholinergic (causing dizziness secondary to orthostatic hypotension, dry mouth, and urinary retention) and antihistaminergic (causing sedation and weight gain). Additionally, TCAs lower the seizure threshold and have adverse cardiac effects relating to their anti-alpha-1 adrenergic activity, resulting in dose-dependent increases in the QTc and cardiac toxicity in overdose that could lead to arrhythmia and death. These medications have their place, but their use requires careful informed consent, clear treatment goals, and baseline and periodic cardiac monitoring (via electrocardiogram).

Serious adverse effects. Clinicians may be hesitant to prescribe antidepressants for pediatric patients because of the potential for more serious adverse effects, including severe behavioral activation syndromes, serotonin syndrome, and emergent suicidality. However, current FDA-approved antidepressants arguably have one of the most positive risk/benefit profiles of any orally-administered medication approved for pediatric patients. Having a strong understanding of the evidence is critical to evaluating when it is appropriate to prescribe an antidepressant, how to properly monitor the patient, and how to obtain accurate informed consent.

Pediatric behavioral activation syndrome. Many clinicians report that children receiving antidepressants experience a pediatric behavioral activation syndrome, which exists along a spectrum from mild activation, increased energy, insomnia, or irritability up through more severe presentations of agitation, hyperactivity, or possibly mania. A recent meta-analysis suggested a positive association between antidepressant use and activation events on the milder end of this spectrum in pediatric patients with non-OCD anxiety disorders,¹⁶ and it is thought that compared with adolescents, younger children are more susceptible to activation adverse effects.³⁶ The likelihood of activation events has been associated with higher antidepressant plasma levels,³⁷ suggesting that dose or individual differences in metabolism may play a role. At the severe end of the spectrum, the risk of induction of mania in pediatric patients with depression or anxiety is relatively rare (<2%) and not statistically different from placebo in RCTs of pediatric participants.³⁸ Meta-analyses of larger randomized, placebo-controlled trials of adults do not support the idea that SSRIs and other second-generation antidepressants carry an increased risk of mania compared with placebo.^39,40 Children or adolescents with bona fide bipolar disorder (ie, patients who have had observed mania that meets all DSM-5 criteria) should be treated with a mood-stabilizing agent or antipsychotic if prescribed an antidepressant.⁴¹ These clear-cut cases are, however, relatively rare, and more often clinicians are confronted with ambiguous cases that include a family history of bipolar disorder along with “softer” symptoms of irritability, intrusiveness, or aggression. In these children, SSRIs may be appropriate for depressive, OCD, or anxiety symptoms, and should be strongly considered before prescribing antipsychotics or mood stabilizers, as long as initiated with proper monitoring.

Serotonin syndrome is a life-threatening condition caused by excess synaptic serotonin. It is characterized by confusion, sweating, diarrhea, hypertension, hyperthermia, and tachycardia. At its most severe, serotonin syndrome can result in seizures, arrhythmias, and death. The risk of serotonin syndrome is very low when using an SSRI as monotherapy. Risk increases with polypharmacy, particularly unexamined polypharmacy when multiple serotonergic agents are inadvertently on board. Commonly used serotonergic agents include other antidepressants, migraine medications (eg, triptans), some pain medications, and the cough suppressant dextromethorphan.

The easiest way to mitigate the risk of serotonin syndrome is to use an interaction index computer program, which can help ensure that the interacting agents are not prescribed without first discussing the risks and benefits. It is important to teach adolescents that certain recreational drugs are highly serotonergic and can cause serious interactions with antidepressants. For example, recreational use of dextrometh­orphan or 3,4-methylenedioxymethamphetamine (MDMA; commonly known as “ecstasy”) has been associated with serotonin syndrome in adolescents taking antidepressant medications.^42,43

Continue to: Suicidality

Suicidality. The black-box warning regarding a risk of emergent suicidality when starting antidepressant treatment in children is controversial.⁴⁴ The prospect that a medication intended to ameliorate depression might instead risk increasing suicidal thinking is alarming to parents and clinicians alike. To appropriately weigh and discuss the risks and benefits with families, it is important to understand the data upon which the warning is based.

In 2004, the FDA commissioned a review of 23 antidepressant trials, both published and unpublished, pooling studies across multiple indications (MDD, OCD, anxiety, and ADHD) and multiple antidepressant classes. This meta-analysis, which included nearly 4,400 pediatric patients, found a small but statistically significant increase in spontaneously-reported suicidal thoughts or actions, with a risk difference of 1% (95% confidence interval [CI], 1% to 2%).⁴⁵ These data suggest that if one treats 100 pediatric patients, 1 to 2 of them may experience short-term increases in suicidal thinking or behavior.⁴⁵ There were no differences in suicidal thinking when assessed systematically (ie, when all subjects reported symptoms of suicidal ideation on structured rating scales), and there were no completed suicides.⁴⁵ A subsequent analysis that included 27 pediatric trials suggested an even lower, although still significant, risk difference (<1%), yielding a number needed to harm (NNH) of 143.⁴⁶ Thus, with low NNT for efficacy (3 to 6) and relatively high NNH for emergent suicidal thoughts or behaviors (100 to 143), for many patients the benefits will outweigh the risks.

Figure 1, Figure 2, and Figure 3 are Cates plots that depict the absolute benefits of antidepressants compared with the risk of suicidality for pediatric patients with MDD, OCD, and anxiety disorders. Recent meta-analyses have suggested that the increased risk of suicidality in antidepressant trials is specific to studies that included children and adolescents, and is not observed in adult studies. A meta-analysis of 70 trials involving 18,526 participants suggested that the odds ratio of suicidality in trials of children and adolescents was 2.39 (95% CI, 1.31 to 4.33) compared with 0.81 (95% CI, 0.51 to 1.28) in adults.⁴⁷ Additionally, a network meta-analysis exclusively focusing on pediatric antidepressant trials in MDD reported significantly higher suicidality-related adverse events in venlafaxine trials compared with placebo, duloxetine, and several SSRIs (fluoxetine, paroxetine, and escitalopram).²⁰ These data should be interpreted with caution as differences in suicidality detected between agents is quite possibly related to differences in the method of assessment between trials, as opposed to actual differences in risk between agents.

Epidemiologic data further support the use of antidepressants in pediatric patients, showing that antidepressant use is associated with decreased teen suicide attempts and completions,⁴⁸ and the decline in prescriptions that occurred following the black-box warning was accompanied by a 14% increase in teen suicides.⁴⁹ Multiple hypotheses have been proposed to explain the pediatric clinical trial findings. One idea is that potential adverse effects of activation, or the intended effects of restoring the motivation, energy, and social engagement that is often impaired in depression, increases the likelihood of thinking about suicide or acting on thoughts. Another theory is that reporting of suicidality may be increased, rather than increased de novo suicidality itself. Antidepressants are effective for treating pediatric anxiety disorders, including social anxiety disorder,¹⁶ which could result in more willingness to report. Also, the manner in which adverse effects are generally ascertained in trials might have led to increased spontaneous reporting. In many trials, investigators ask whether participants have any adverse effects in general, and inquire about specific adverse effects only if the family answers affirmatively. Thus, the increased rate of other adverse effects associated with antidepressants (sleep problems, gastrointestinal upset, dry mouth, etc.) might trigger a specific question regarding suicidal ideation, which the child or family then may be more likely to report. Alternatively, any type of psychiatric treatment could increase an individual’s propensity to report; in adolescent psychotherapy trials, non-medicated participants have reported emergent suicidality at similar frequencies as those described in drug trials.⁵⁰ Regardless of the mechanism, the possibility of treatment-emergent suicidality is a low-frequency but serious event that necessitates careful monitoring when starting medication. Current guidelines suggest seeing children weekly for the first month after medication initiation, every 2 weeks for the following month, and monthly thereafter.⁵¹

Continue to: How long should the antidepressant be continued?

How long should the antidepressant be continued?

Many patients are concerned about how long they may be taking medication, and whether they will be taking an antidepressant “forever.” A treatment course can be broken into an acute phase, wherein remission is achieved and maintained for 6 to 8 weeks. This is followed by a continuation phase, with the goal of relapse prevention, lasting 16 to 20 weeks. The length of the last phase—the maintenance phase—depends both on the child’s history, the underlying therapeutic indication, the adverse effect burden experienced, and the family’s preferences/values. In general, for a first depressive episode, after treating for 1 year, a trial of discontinuation can be attempted with close monitoring. For a second depressive episode, we recommend 2 years of remission on antidepressant therapy before attempting discontinuation. In patients who have had 3 depressive episodes, or have had episodes of high severity, we recommend continuing antidepressant treatment indefinitely. Although much less well studied, the risk of relapse following SSRI discontinuation appears much more significant in OCD, whereas anxiety disorders and MDD have a relatively comparable risk.⁵²

In general, stopping an antidepressant should be done carefully and slowly. The speed with which a specific antidepressant can be stopped is largely related to its half-life. Agents with very long half-lives, such as fluoxetine (half-life of 5 days for the parent compound and 9 days for active metabolite), can often be stopped altogether, being “auto-tapered” by the long half-life. One might still consider a taper if the patient were taking high doses. Medications with shorter half-lives must be more carefully tapered to avoid discontinuation syndromes. Discontinuation syndromes are characterized by flu-like symptoms (nausea, myalgias, fatigue, dizziness) and worsening mood. Medications with short half-lives (eg, paroxetine and venlafaxine) have the highest potential for this syndrome in children,⁵³ and thus are used less frequently.

What to do when first-line treatments fail

When a child does not experience sufficient improvement from first-line treatments, it is crucial to determine whether they have experienced an adequate dosing, duration, and quality of medication and psychotherapy.

Adequate psychotherapy? To determine whether children are receiving adequate CBT, ask:

1. if the child receives homework from psychotherapy
2. if the parents are included in treatment
3. if therapy has involved identifying thought patterns that may be contributing to the child’s illness, and
4. if the therapist has ever exposed the child to a challenge likely to produce anxiety or distress in a supervised environment and has developed an exposure hierarchy (for conditions with primarily exposure-based therapies, such as OCD or anxiety disorders).

If a family is not receiving most of these elements in psychotherapy, this is a good indicator that they may not be receiving evidence-based CBT.

Continue to: Adequate pharmacotherapy?

Adequate pharmacotherapy? Similarly, when determining the adequacy of previous pharmacotherapy, it is critical to determine whether the child received an adequate dose of medications (at least the FDA-recommended minimum dose) for an adequate duration of time at therapeutic dosing (at least 6 weeks for MDD, 8 weeks for anxiety disorders, and 8 to 12 weeks for pediatric patients with OCD), and that the child actually took the medication regularly during that period. Patient compliance can typically be tracked through checking refill requests or intervals through the patient’s pharmacy. Ensuring proper delivery of first-line treatments is imperative because (1) the adverse effects associated with second-line treatments are often more substantial; (2) the cost in terms of time and money is considerably higher with second-line treatments, and; (3) the evidence regarding the benefits of these treatments is much less certain.

Inadequate dosing is a common reason for non-response in pediatric patients. Therapeutic dose ranges for common antidepressants are displayed in Table 1. Many clinicians underdose antidepressants for pediatric patients initially (and often throughout treatment) because they fear that the typical dose titration used in clinical trials will increase the risk of adverse effects compared with more conservative dosing. There is limited evidence to suggest that this underdosing strategy is likely to be successful; adverse effects attributable to these medications are modest, and most that are experienced early in treatment (eg, headache, increased anxiety or irritability, sleep problems, gastrointestinal upset) are self-limiting and may be coincidental rather than medication-induced. Furthermore, there is no evidence for efficacy of subtherapeutic dosing in children in the acute phase of treatment or for preventing relapse.¹⁴ Thus, from an efficacy standpoint, a medication trial has not officially begun until the therapeutic dose range is reached.

Once dosing is within the therapeutic range, however, pediatric data differs from the adult literature. In most adult psychi­atric conditions, higher doses of SSRIs within the therapeutic range are associated with an increased response rate.^14,54 In pediatrics, there are few fixed dose trials, and once within the recommended therapeutic range, minimal data supports an association between higher dosing and higher efficacy.¹⁴ In general, pediatric guidelines are silent regarding optimal dosing of SSRIs within the recommended dose range, and higher antidepressant doses often result in a more significant adverse effect burden for children. One exception is pediatric OCD, where, similar to adults, the guidelines suggest that higher dosing of SSRIs often is required to induce a therapeutic response as compared to MDD and GAD.^31,55

If a child does not respond to adequate first-line treatment (or has a treatment history that cannot be fully verified), repeating these first-line interventions carries little risk and can be quite effective. For example, 60% of adolescents with MDD who did not initially respond to an SSRI demonstrated a significant response when prescribed a second SSRI or venlafaxine (with or without CBT).⁵⁶

When pediatric patients continue to experience significantly distressing and/or debilitating symptoms (particularly in MDD) despite multiple trials of antidepressants and psychotherapy, practitioners should consider a careful referral to interventional psychiatry services, which can include the more intensive treatments of electroconvulsive therapy, repetitive transcranial magnetic stimulation, or ketamine (see Box 1). Given the substantial morbidity and mortality associated with adolescent depression, interventional psychiatry treatments are under-researched and under-utilized clinically in pediatric populations.

Continue to: Antidepressants in general...

Antidepressants in general, and SSRIs in particular, are the first-line pharmacotherapy for pediatric anxiety, OCD, and MDD. For PTSD, although they are a first-line treatment in adults, their efficacy has not been demonstrated in children and adolescents. Antidepressants are generally safe, well-tolerated, and effective, with low NNTs (3 to 5 for anxiety and OCD; 4 to 12 in MDD, depending on whether industry trials are included). It is important that clinicians and families be educated about possible adverse effects and their time course in order to anticipate difficulties, ensure adequate informed consent, and monitor appropriately. The black-box warning regarding treatment-emergent suicidal thoughts or behaviors must be discussed (for suggested talking points, see Box 2). The NNH is large (100 to 143), and for many patients, the benefits will outweigh the risks. For pediatric patients who fail to respond to multiple adequate trials of antidepressants and psychotherapy, referrals for interventional psychiatry consultation should be considered.

Bottom Line

Although the evidence base for prescribing antidepressants for children and adolescents is smaller compared to the adult literature, properly understanding and prescribing these agents, and explaining their risks and benefits to families, can make a major difference in patient compliance, satisfaction, and outcomes. Antidepressants (particularly selective serotonin reuptake inhibitors) are the firstline pharmacologic intervention for pediatric patients with anxiety disorders, obsessive-compulsive disorder, or major depressive disorder.

Related Resource

• Bonin L, Moreland CS. Overview of prevention and treatment for pediatric depression. https://www.uptodate.com/contents/overview-of-prevention-and-treatment-for-pediatric-depression. Last reviewed July 2019.

Drug Brand Names

Bupropion • Wellbutrin, Zyban
Cimetidine • Tagamet
Citalopram • Celexa
Clomipramine • Anafranil
Desipramine • Norpramin
Desvenlafaxine • Pristiq
Duloxetine • Cymbalta
Escitalopram • Lexapro
Fluoxetine • Prozac
Fluvoxamine • Luvox
Imipramine • Tofranil
Mirtazapine • Remeron
Nortriptyline • Pamelor
Paroxetine • Paxil
Sertraline • Zoloft
Venlafaxine • Effexor
Vilazodone • Viibryd
Vortioxetine • Trintellix

Box 1

Box 2

PARIS – Lower levels of physical activity over time were linked to a significantly increased risk of both cardiovascular and all-cause mortality, according to results from a prospective cohort study that followed more than 23,000 adults for over 2 decades.

“Individuals who remained physically inactive, or who decreased their physical activity over 22 years, had substantially increased risk of dying from all causes, and from cardiovascular disease,” Trine Moholdt, PhD, said at a press conference at the annual congress of the European Society of Cardiology.

“The bad news is that sustained inactivity was associated with a 99% increase in all-cause mortality and 168% increase in cardiovascular deaths” compared with sustained physical activity, she said.

The news was similarly bad for individuals who had been highly active and then became inactive; this cohort had an increase in all-cause mortality of 116% and sustained a 173% increase in cardiovascular deaths. “Previous activity levels – they don’t count. You have to keep it up,” said Dr. Moholdt, speaking in a video interview.

However, the risk associated with inactivity was attenuated for those patients who became more active over the course of the study period; their risk was still higher than that of those who had sustained activity, but lower than those who remained inactive. “It’s never too late to start being physically active,” said Dr. Moholdt.

All groups were compared with individuals who remained highly active over the study period; this reference group had the lowest risk of both cardiovascular and all-cause death.

Dr. Moholdt noted a limitation of most previous research into how physical activity relates to cardiovascular and all-cause mortality: Data are obtained just at baseline, and then associated with a downstream outcome. “But people change!” she said, so it’s important to track how activity levels change over time.

To that end, the Nord-Trøndelag Health Study (HUNT study) measured activity levels for 23,156 participants at two points, and then assessed cardiovascular and all-cause mortality, over a period of 22 years.

[email protected]

SOURCE: Moholdt T. et al. ESC Congress 2019, Abstract P627.

PARIS – Lower levels of physical activity over time were linked to a significantly increased risk of both cardiovascular and all-cause mortality, according to results from a prospective cohort study that followed more than 23,000 adults for over 2 decades.

“Individuals who remained physically inactive, or who decreased their physical activity over 22 years, had substantially increased risk of dying from all causes, and from cardiovascular disease,” Trine Moholdt, PhD, said at a press conference at the annual congress of the European Society of Cardiology.

“The bad news is that sustained inactivity was associated with a 99% increase in all-cause mortality and 168% increase in cardiovascular deaths” compared with sustained physical activity, she said.

The news was similarly bad for individuals who had been highly active and then became inactive; this cohort had an increase in all-cause mortality of 116% and sustained a 173% increase in cardiovascular deaths. “Previous activity levels – they don’t count. You have to keep it up,” said Dr. Moholdt, speaking in a video interview.

However, the risk associated with inactivity was attenuated for those patients who became more active over the course of the study period; their risk was still higher than that of those who had sustained activity, but lower than those who remained inactive. “It’s never too late to start being physically active,” said Dr. Moholdt.

All groups were compared with individuals who remained highly active over the study period; this reference group had the lowest risk of both cardiovascular and all-cause death.

Dr. Moholdt noted a limitation of most previous research into how physical activity relates to cardiovascular and all-cause mortality: Data are obtained just at baseline, and then associated with a downstream outcome. “But people change!” she said, so it’s important to track how activity levels change over time.

To that end, the Nord-Trøndelag Health Study (HUNT study) measured activity levels for 23,156 participants at two points, and then assessed cardiovascular and all-cause mortality, over a period of 22 years.

[email protected]

SOURCE: Moholdt T. et al. ESC Congress 2019, Abstract P627.

PARIS – Lower levels of physical activity over time were linked to a significantly increased risk of both cardiovascular and all-cause mortality, according to results from a prospective cohort study that followed more than 23,000 adults for over 2 decades.

“Individuals who remained physically inactive, or who decreased their physical activity over 22 years, had substantially increased risk of dying from all causes, and from cardiovascular disease,” Trine Moholdt, PhD, said at a press conference at the annual congress of the European Society of Cardiology.

“The bad news is that sustained inactivity was associated with a 99% increase in all-cause mortality and 168% increase in cardiovascular deaths” compared with sustained physical activity, she said.

The news was similarly bad for individuals who had been highly active and then became inactive; this cohort had an increase in all-cause mortality of 116% and sustained a 173% increase in cardiovascular deaths. “Previous activity levels – they don’t count. You have to keep it up,” said Dr. Moholdt, speaking in a video interview.

However, the risk associated with inactivity was attenuated for those patients who became more active over the course of the study period; their risk was still higher than that of those who had sustained activity, but lower than those who remained inactive. “It’s never too late to start being physically active,” said Dr. Moholdt.

All groups were compared with individuals who remained highly active over the study period; this reference group had the lowest risk of both cardiovascular and all-cause death.

Dr. Moholdt noted a limitation of most previous research into how physical activity relates to cardiovascular and all-cause mortality: Data are obtained just at baseline, and then associated with a downstream outcome. “But people change!” she said, so it’s important to track how activity levels change over time.

To that end, the Nord-Trøndelag Health Study (HUNT study) measured activity levels for 23,156 participants at two points, and then assessed cardiovascular and all-cause mortality, over a period of 22 years.

[email protected]

SOURCE: Moholdt T. et al. ESC Congress 2019, Abstract P627.

The number of possible cases of vaping-related pulmonary illness has risen to 215, reported from 25 states, as of Aug. 27, 2019, according to the Centers for Disease Control and Prevention, Atlanta. Additional reports of pulmonary illness are under investigation.

The CDC has released a standardized case definition that states are using to complete their own investigations and verifications of cases. It appears that all cases are linked to e-cigarette product use, but the cause of the respiratory illnesses is still unconfirmed.

In many cases, patients reported a gradual start of symptoms, including breathing difficulty, shortness of breath, and/or chest pain before hospitalization. Some cases reported mild to moderate gastrointestinal illness including vomiting and diarrhea, or other symptoms such as fevers or fatigue. In many cases, patients have also acknowledged recent use of tetrahydrocannabinol (THC)-containing e-cigarette products while speaking to health care personnel or in follow-up interviews by health department staff, according to a statement from the CDC and the Food and Drug Administration.

The agencies are working with state health departments to standardize information collection at the state level to help build a more comprehensive picture of these incidents, including the brand and types of e-cigarette products, whether any of them would fall within the FDA’s regulatory authority, where they were obtained, and whether there is a link to specific devices, ingredients, or contaminants in the devices or substances associated with e-cigarette product use.

CDC staff have been deployed to Illinois and Wisconsin to assist their state health departments. The agencies have released a Clinician Outreach and Communication Activity (COCA) Clinical Action Alert describing this investigation and asking providers to report possible cases to their state health departments. In addition to a standardized case definition, the agencies have issued a medical chart abstraction form and case interview questionnaire, are reviewing and providing feedback on data collection and health messaging tools for states, and are facilitating information sharing between states with possible cases.

More information on the cases and reporting are available from the CDC.

[email protected]

The number of possible cases of vaping-related pulmonary illness has risen to 215, reported from 25 states, as of Aug. 27, 2019, according to the Centers for Disease Control and Prevention, Atlanta. Additional reports of pulmonary illness are under investigation.

The CDC has released a standardized case definition that states are using to complete their own investigations and verifications of cases. It appears that all cases are linked to e-cigarette product use, but the cause of the respiratory illnesses is still unconfirmed.

In many cases, patients reported a gradual start of symptoms, including breathing difficulty, shortness of breath, and/or chest pain before hospitalization. Some cases reported mild to moderate gastrointestinal illness including vomiting and diarrhea, or other symptoms such as fevers or fatigue. In many cases, patients have also acknowledged recent use of tetrahydrocannabinol (THC)-containing e-cigarette products while speaking to health care personnel or in follow-up interviews by health department staff, according to a statement from the CDC and the Food and Drug Administration.

The agencies are working with state health departments to standardize information collection at the state level to help build a more comprehensive picture of these incidents, including the brand and types of e-cigarette products, whether any of them would fall within the FDA’s regulatory authority, where they were obtained, and whether there is a link to specific devices, ingredients, or contaminants in the devices or substances associated with e-cigarette product use.

CDC staff have been deployed to Illinois and Wisconsin to assist their state health departments. The agencies have released a Clinician Outreach and Communication Activity (COCA) Clinical Action Alert describing this investigation and asking providers to report possible cases to their state health departments. In addition to a standardized case definition, the agencies have issued a medical chart abstraction form and case interview questionnaire, are reviewing and providing feedback on data collection and health messaging tools for states, and are facilitating information sharing between states with possible cases.

More information on the cases and reporting are available from the CDC.

[email protected]

The number of possible cases of vaping-related pulmonary illness has risen to 215, reported from 25 states, as of Aug. 27, 2019, according to the Centers for Disease Control and Prevention, Atlanta. Additional reports of pulmonary illness are under investigation.

The CDC has released a standardized case definition that states are using to complete their own investigations and verifications of cases. It appears that all cases are linked to e-cigarette product use, but the cause of the respiratory illnesses is still unconfirmed.

In many cases, patients reported a gradual start of symptoms, including breathing difficulty, shortness of breath, and/or chest pain before hospitalization. Some cases reported mild to moderate gastrointestinal illness including vomiting and diarrhea, or other symptoms such as fevers or fatigue. In many cases, patients have also acknowledged recent use of tetrahydrocannabinol (THC)-containing e-cigarette products while speaking to health care personnel or in follow-up interviews by health department staff, according to a statement from the CDC and the Food and Drug Administration.

The agencies are working with state health departments to standardize information collection at the state level to help build a more comprehensive picture of these incidents, including the brand and types of e-cigarette products, whether any of them would fall within the FDA’s regulatory authority, where they were obtained, and whether there is a link to specific devices, ingredients, or contaminants in the devices or substances associated with e-cigarette product use.

CDC staff have been deployed to Illinois and Wisconsin to assist their state health departments. The agencies have released a Clinician Outreach and Communication Activity (COCA) Clinical Action Alert describing this investigation and asking providers to report possible cases to their state health departments. In addition to a standardized case definition, the agencies have issued a medical chart abstraction form and case interview questionnaire, are reviewing and providing feedback on data collection and health messaging tools for states, and are facilitating information sharing between states with possible cases.

More information on the cases and reporting are available from the CDC.

[email protected]

Physician groups are praising a new option by the American Board of Internal Medicine (ABIM) that will offer doctors a self-paced pathway for maintenance of certification (MOC) in place of the traditional long-form assessment route.

The new longitudinal assessment option, announced in late August, would enable physicians to acquire and demonstrate ongoing knowledge through shorter evaluations of specific content. The option, currently under development, also would provide doctors with immediate feedback about their answers and share links to educational material to address knowledge gaps, according to an announcement. While details are still being flushed out, a summary of the longitudinal assessment concept by the American Board of Medical Specialties explains that the approach draws on the principles of adult learning and modern technology “to promote learning, retention, and transfer of information.”

Developing a longitudinal assessment option is part of ABIM’s ongoing evolution, Marianne M. Green, MD, chair for ABIM’s board of directors and ABIM President Richard J. Baron, MD, wrote in a joint letter to internists posted on ABIM’s blog.

“We recognize that some physicians may prefer a more continuous process that easily integrates into their lives and allows them to engage seamlessly at their preferred pace, while being able to access the resources they use in practice,” the doctors wrote.

Douglas DeLong, MD, chair of the American College of Physician’s (ACP) board of regents said the option is a positive, first step that will support lifelong learning. He noted the new option is in line with recommendations by the American Board of Medical Specialties’ Continuing Board Certification: Vision for the Future Commission, which included ACP concerns.

“It’s pretty clear that some of the principles of adult learning – frequent information with quick feedback, repetition of material, and identifying gaps in knowledge – is really how people most effectively learn,” Dr. DeLong said in an interview. “Just cramming for an examination every decade hasn’t ever really been shown to affect long-term retention of knowledge or even patient care outcomes.”

[email protected]

Physician groups are praising a new option by the American Board of Internal Medicine (ABIM) that will offer doctors a self-paced pathway for maintenance of certification (MOC) in place of the traditional long-form assessment route.

The new longitudinal assessment option, announced in late August, would enable physicians to acquire and demonstrate ongoing knowledge through shorter evaluations of specific content. The option, currently under development, also would provide doctors with immediate feedback about their answers and share links to educational material to address knowledge gaps, according to an announcement. While details are still being flushed out, a summary of the longitudinal assessment concept by the American Board of Medical Specialties explains that the approach draws on the principles of adult learning and modern technology “to promote learning, retention, and transfer of information.”

Developing a longitudinal assessment option is part of ABIM’s ongoing evolution, Marianne M. Green, MD, chair for ABIM’s board of directors and ABIM President Richard J. Baron, MD, wrote in a joint letter to internists posted on ABIM’s blog.

“We recognize that some physicians may prefer a more continuous process that easily integrates into their lives and allows them to engage seamlessly at their preferred pace, while being able to access the resources they use in practice,” the doctors wrote.

Douglas DeLong, MD, chair of the American College of Physician’s (ACP) board of regents said the option is a positive, first step that will support lifelong learning. He noted the new option is in line with recommendations by the American Board of Medical Specialties’ Continuing Board Certification: Vision for the Future Commission, which included ACP concerns.

“It’s pretty clear that some of the principles of adult learning – frequent information with quick feedback, repetition of material, and identifying gaps in knowledge – is really how people most effectively learn,” Dr. DeLong said in an interview. “Just cramming for an examination every decade hasn’t ever really been shown to affect long-term retention of knowledge or even patient care outcomes.”

[email protected]

Physician groups are praising a new option by the American Board of Internal Medicine (ABIM) that will offer doctors a self-paced pathway for maintenance of certification (MOC) in place of the traditional long-form assessment route.

The new longitudinal assessment option, announced in late August, would enable physicians to acquire and demonstrate ongoing knowledge through shorter evaluations of specific content. The option, currently under development, also would provide doctors with immediate feedback about their answers and share links to educational material to address knowledge gaps, according to an announcement. While details are still being flushed out, a summary of the longitudinal assessment concept by the American Board of Medical Specialties explains that the approach draws on the principles of adult learning and modern technology “to promote learning, retention, and transfer of information.”

Developing a longitudinal assessment option is part of ABIM’s ongoing evolution, Marianne M. Green, MD, chair for ABIM’s board of directors and ABIM President Richard J. Baron, MD, wrote in a joint letter to internists posted on ABIM’s blog.

“We recognize that some physicians may prefer a more continuous process that easily integrates into their lives and allows them to engage seamlessly at their preferred pace, while being able to access the resources they use in practice,” the doctors wrote.

Douglas DeLong, MD, chair of the American College of Physician’s (ACP) board of regents said the option is a positive, first step that will support lifelong learning. He noted the new option is in line with recommendations by the American Board of Medical Specialties’ Continuing Board Certification: Vision for the Future Commission, which included ACP concerns.

“It’s pretty clear that some of the principles of adult learning – frequent information with quick feedback, repetition of material, and identifying gaps in knowledge – is really how people most effectively learn,” Dr. DeLong said in an interview. “Just cramming for an examination every decade hasn’t ever really been shown to affect long-term retention of knowledge or even patient care outcomes.”

[email protected]

Colorectal cancer (CRC) survivor rates are improving, which means people are living long enough after the cancer to have other chronic conditions. CRC is the third most commonly diagnosed cancer among users of the US Department of Veterans Affairs (VA) health care system, according to VA researchers, and there is a high prevalence of cardiovascular disease (CVD). The researchers also say emerging evidence suggests that survivors of CRC may be more likely to develop diabetes mellitus (DM) in the 5 years following their cancer diagnosis. But they add that there is a paucity of research about control of CVD-related chronic conditions among survivors of CRC.

In a retrospective study, the researchers compared 9,758 nonmetastatic patients with CRC with 29,066 people who had not had cancer. At baseline, 69% of the survivors of CRC and the matched controls were diagnosed with hypertension, 52% with hyperlipidemia, and 37% with DM.

But somewhat contrary to expectations, the researchers found no significant differences between the 2 groups for DM in the year following the baseline assessment. The researchers point to the VA’s “strong history” of DM risk reduction research and 2 national programs targeting DM, although they do not know whether the people in their study participated in those.

The survivors of CRC also had half the odds of being diagnosed with hyperlipidemia. However, they did have 57% higher odds of being diagnosed with hypertension.

Although the researchers acknowledge that hypertension is a transient adverse effect of certain chemotherapy regimens, they found only 7 survivors of CRC and 11 controls were treated with bevacizumab during their first year postanchor date.

The relationship between nonmetastatic CRC and CVD risk-related chronic conditions is complex, the researchers say. But they share risk factors, including obesity, physical inactivity, and diet.

The researchers call behavioral change interventions that improve survivors of CRC physical activity, dietary habits, and body mass index a “promising beginning” but call for other similar interventions, particularly those targeting blood pressure management and adherence to antihypertensive medications (which was significantly lower among the survivors).

While the magnitude of the effect regarding hypertension seems relatively small, the researchers say, they believe it is still an important difference when considered from a population health perspective—and one that should be addressed. The researchers also note that nonmetastatic survivors of CRC and controls had very similar rates of primary care visits in the 3 years postanchor date and as a result similar opportunities to receive a hypertension diagnosis.

Colorectal cancer (CRC) survivor rates are improving, which means people are living long enough after the cancer to have other chronic conditions. CRC is the third most commonly diagnosed cancer among users of the US Department of Veterans Affairs (VA) health care system, according to VA researchers, and there is a high prevalence of cardiovascular disease (CVD). The researchers also say emerging evidence suggests that survivors of CRC may be more likely to develop diabetes mellitus (DM) in the 5 years following their cancer diagnosis. But they add that there is a paucity of research about control of CVD-related chronic conditions among survivors of CRC.

In a retrospective study, the researchers compared 9,758 nonmetastatic patients with CRC with 29,066 people who had not had cancer. At baseline, 69% of the survivors of CRC and the matched controls were diagnosed with hypertension, 52% with hyperlipidemia, and 37% with DM.

But somewhat contrary to expectations, the researchers found no significant differences between the 2 groups for DM in the year following the baseline assessment. The researchers point to the VA’s “strong history” of DM risk reduction research and 2 national programs targeting DM, although they do not know whether the people in their study participated in those.

The survivors of CRC also had half the odds of being diagnosed with hyperlipidemia. However, they did have 57% higher odds of being diagnosed with hypertension.

Although the researchers acknowledge that hypertension is a transient adverse effect of certain chemotherapy regimens, they found only 7 survivors of CRC and 11 controls were treated with bevacizumab during their first year postanchor date.

The relationship between nonmetastatic CRC and CVD risk-related chronic conditions is complex, the researchers say. But they share risk factors, including obesity, physical inactivity, and diet.

The researchers call behavioral change interventions that improve survivors of CRC physical activity, dietary habits, and body mass index a “promising beginning” but call for other similar interventions, particularly those targeting blood pressure management and adherence to antihypertensive medications (which was significantly lower among the survivors).

While the magnitude of the effect regarding hypertension seems relatively small, the researchers say, they believe it is still an important difference when considered from a population health perspective—and one that should be addressed. The researchers also note that nonmetastatic survivors of CRC and controls had very similar rates of primary care visits in the 3 years postanchor date and as a result similar opportunities to receive a hypertension diagnosis.

Colorectal cancer (CRC) survivor rates are improving, which means people are living long enough after the cancer to have other chronic conditions. CRC is the third most commonly diagnosed cancer among users of the US Department of Veterans Affairs (VA) health care system, according to VA researchers, and there is a high prevalence of cardiovascular disease (CVD). The researchers also say emerging evidence suggests that survivors of CRC may be more likely to develop diabetes mellitus (DM) in the 5 years following their cancer diagnosis. But they add that there is a paucity of research about control of CVD-related chronic conditions among survivors of CRC.

In a retrospective study, the researchers compared 9,758 nonmetastatic patients with CRC with 29,066 people who had not had cancer. At baseline, 69% of the survivors of CRC and the matched controls were diagnosed with hypertension, 52% with hyperlipidemia, and 37% with DM.

But somewhat contrary to expectations, the researchers found no significant differences between the 2 groups for DM in the year following the baseline assessment. The researchers point to the VA’s “strong history” of DM risk reduction research and 2 national programs targeting DM, although they do not know whether the people in their study participated in those.

The survivors of CRC also had half the odds of being diagnosed with hyperlipidemia. However, they did have 57% higher odds of being diagnosed with hypertension.

Although the researchers acknowledge that hypertension is a transient adverse effect of certain chemotherapy regimens, they found only 7 survivors of CRC and 11 controls were treated with bevacizumab during their first year postanchor date.

The relationship between nonmetastatic CRC and CVD risk-related chronic conditions is complex, the researchers say. But they share risk factors, including obesity, physical inactivity, and diet.

The researchers call behavioral change interventions that improve survivors of CRC physical activity, dietary habits, and body mass index a “promising beginning” but call for other similar interventions, particularly those targeting blood pressure management and adherence to antihypertensive medications (which was significantly lower among the survivors).

While the magnitude of the effect regarding hypertension seems relatively small, the researchers say, they believe it is still an important difference when considered from a population health perspective—and one that should be addressed. The researchers also note that nonmetastatic survivors of CRC and controls had very similar rates of primary care visits in the 3 years postanchor date and as a result similar opportunities to receive a hypertension diagnosis.

Three years after gabapentin received US Food and Drug Administration (FDA) approval in 1990 for epilepsy, case reports and animal studies emerged announcing its potential in the treatment of pain syndromes through then-novel analgesic mechanisms.¹ Fast forward 20 years to 2016: gabapentin and its close cousin, pregabalin, are internationally considered first-line agents for the treatment of neuropathic pain in guidelines from the Centers for Disease Control and Prevention, the Canadian Pain Society, and the National Institute for Health and Care Excellence. Gabapentin is the 10th most prescribed drug in the United States, and brand-name pregabalin sales were $4.4 billion USD, ranking 8th in invoice drug spending.²

The ascendancy of gabapentinoids as drugs of choice for pain, though, is fraught with controversy; yet, they were shepherded to commercial success. In 2004, the patent owner of gabapentin, Warner-Lambert (now owned by Pfizer), admitted guilt to charges that it violated federal regulations in its promotion: they encouraged off-label prescribing through paid physician-to-physician communications, publication of positive outcomes, and suppression of negative ones.³ Pfizer paid another settlement in 2009 for false claims about off-label indications for brand-name pregabalin.⁴

Mindful of historical biases, recent trials and meta-analyses have found less favorable outcomes for gabapentinoids in the treatment of off-label pain conditions and greater risks than previously reported. Cochrane reviews for gabapentin demonstrate efficacy only in postherpetic neuralgia (for which it has FDA approval) and diabetic peripheral neuropathy (for which it does not); pregabalin has efficacy in both these conditions as well as posttraumatic neuropathic pain and fibromyalgia (and FDA approval for all four). For other types of neuropathic pain, the evidence is of lower quality. Even for approved indications, the risk–benefit ratio is questionable, as the numbers needed to harm for dizziness and somnolence are similar to the numbers needed to treat for pain.^5,6 Further, case–control studies have found increased odds of opioid-related death when gabapentinoids were coprescribed with opioids,^7,8 prompting gabapentinoids to be reclassified as class C controlled substances in the UK as of April 2019.⁹

On this backdrop, Gingras and colleagues publish their retrospective cohort study on high-risk prescribing of these popular drugs in Montreal, Canada in this issue of Journal of Hospital Medicine.¹⁰ In their retrospective cohort study of 4,103 patients admitted to a clinical teaching unit, more than one in eight patients (13%) were being prescribed a gabapentinoid as an outpatient; chart review of the admission notes indicated that only 17% of them had an FDA-approved indication and 28% had no clear indication. Gabapentinoid users were more likely to be coprescribed an opioid than nonusers (28% vs 12%). There was no significant difference in length of stay or inpatient death between users and nonusers.

Gingras et al. thereby conclude that there is an opportunity to deprescribe on the basis of few gabapentinoid users having a documented indication and the recent research showing potentials for harm and abuse.¹¹ We agree. Messaging around gabapentinoids should be similar to that for opioids: these are medications with limited evidence supporting their use in the treatment of chronic pain, and prescribing them for unapproved indications risks doing greater harm than good. We offer two recommendations on how hospitalists can proceed with deprescribing them safely.

First, the urgency of deprescribing in inpatient settings should be titrated to the degree of risk. When the reason for hospitalization is potentially an adverse drug effect, culprit medications posing a substantial and near-term risk of harm should be stopped, such as when patients on gabapentinoids present with major alteration of their mental status.

In less urgent circumstances, hospitalists should speak first with outpatient prescribers because they may have important contextual information (eg, indication, patient preference, failure of alternative therapies, etc.) about previous care that the inpatient clinician lacks. For gabapentinoids, it is easy to imagine how treated pain syndromes without objective markers of disease may escape notice by a hospitalist and remain undocumented, which may encourage erroneous deprescribing. If the shared decision between the patient and providers is to deprescribe, patients on high doses warrant a tapering schedule.¹¹ Pharmacist consultation can help with this.

Second, before discharge, hospitalists should communicate their rationale for deprescribing medications to both patients and outpatient prescribers, especially if a prolonged tapering schedule is required. This type of communication occurred infrequently in this study: the reason for deprescribing a gabapentinoid was missing from the discharge summary 55% of the time. Without this, outpatient prescribers may simply reinitiate the medication after the patient is discharged.

To counter the overuse of gabapentinoids, hospitalists should look for opportunities to deprescribe them where there is concern about adverse events and when evidence-based indications do not exist. Successful deprescribing of these popular drugs will require deliberate collaboration and communication with the outpatient circle of care, as ongoing deprescribing ultimately depends on patients and outpatient prescribers agreeing to the change.

Disclosures

Dr. Steinman served as an unpaid expert witness in United States of America ex. Rel. David Franklin vs. Parke-Davis, Division of Warner-Lambert Company and Pfizer, Inc, litigation which alleged that the named pharmaceutical companies improperly marketed gabapentin for non-FDA-approved uses. Drs. Lam and Rochon have no conflicts of interest to declare.

Funding

Dr. Rochon is supported by the Retired Teachers of Ontario (RTO/ERO) Chair in Geriatric Medicine at the University of Toronto. Dr. Steinman is supported by the National Institute on Aging, US (K24AG049057 and P30AG044281).

Three years after gabapentin received US Food and Drug Administration (FDA) approval in 1990 for epilepsy, case reports and animal studies emerged announcing its potential in the treatment of pain syndromes through then-novel analgesic mechanisms.¹ Fast forward 20 years to 2016: gabapentin and its close cousin, pregabalin, are internationally considered first-line agents for the treatment of neuropathic pain in guidelines from the Centers for Disease Control and Prevention, the Canadian Pain Society, and the National Institute for Health and Care Excellence. Gabapentin is the 10th most prescribed drug in the United States, and brand-name pregabalin sales were $4.4 billion USD, ranking 8th in invoice drug spending.²

The ascendancy of gabapentinoids as drugs of choice for pain, though, is fraught with controversy; yet, they were shepherded to commercial success. In 2004, the patent owner of gabapentin, Warner-Lambert (now owned by Pfizer), admitted guilt to charges that it violated federal regulations in its promotion: they encouraged off-label prescribing through paid physician-to-physician communications, publication of positive outcomes, and suppression of negative ones.³ Pfizer paid another settlement in 2009 for false claims about off-label indications for brand-name pregabalin.⁴

Mindful of historical biases, recent trials and meta-analyses have found less favorable outcomes for gabapentinoids in the treatment of off-label pain conditions and greater risks than previously reported. Cochrane reviews for gabapentin demonstrate efficacy only in postherpetic neuralgia (for which it has FDA approval) and diabetic peripheral neuropathy (for which it does not); pregabalin has efficacy in both these conditions as well as posttraumatic neuropathic pain and fibromyalgia (and FDA approval for all four). For other types of neuropathic pain, the evidence is of lower quality. Even for approved indications, the risk–benefit ratio is questionable, as the numbers needed to harm for dizziness and somnolence are similar to the numbers needed to treat for pain.^5,6 Further, case–control studies have found increased odds of opioid-related death when gabapentinoids were coprescribed with opioids,^7,8 prompting gabapentinoids to be reclassified as class C controlled substances in the UK as of April 2019.⁹

On this backdrop, Gingras and colleagues publish their retrospective cohort study on high-risk prescribing of these popular drugs in Montreal, Canada in this issue of Journal of Hospital Medicine.¹⁰ In their retrospective cohort study of 4,103 patients admitted to a clinical teaching unit, more than one in eight patients (13%) were being prescribed a gabapentinoid as an outpatient; chart review of the admission notes indicated that only 17% of them had an FDA-approved indication and 28% had no clear indication. Gabapentinoid users were more likely to be coprescribed an opioid than nonusers (28% vs 12%). There was no significant difference in length of stay or inpatient death between users and nonusers.

Gingras et al. thereby conclude that there is an opportunity to deprescribe on the basis of few gabapentinoid users having a documented indication and the recent research showing potentials for harm and abuse.¹¹ We agree. Messaging around gabapentinoids should be similar to that for opioids: these are medications with limited evidence supporting their use in the treatment of chronic pain, and prescribing them for unapproved indications risks doing greater harm than good. We offer two recommendations on how hospitalists can proceed with deprescribing them safely.

First, the urgency of deprescribing in inpatient settings should be titrated to the degree of risk. When the reason for hospitalization is potentially an adverse drug effect, culprit medications posing a substantial and near-term risk of harm should be stopped, such as when patients on gabapentinoids present with major alteration of their mental status.

In less urgent circumstances, hospitalists should speak first with outpatient prescribers because they may have important contextual information (eg, indication, patient preference, failure of alternative therapies, etc.) about previous care that the inpatient clinician lacks. For gabapentinoids, it is easy to imagine how treated pain syndromes without objective markers of disease may escape notice by a hospitalist and remain undocumented, which may encourage erroneous deprescribing. If the shared decision between the patient and providers is to deprescribe, patients on high doses warrant a tapering schedule.¹¹ Pharmacist consultation can help with this.

Second, before discharge, hospitalists should communicate their rationale for deprescribing medications to both patients and outpatient prescribers, especially if a prolonged tapering schedule is required. This type of communication occurred infrequently in this study: the reason for deprescribing a gabapentinoid was missing from the discharge summary 55% of the time. Without this, outpatient prescribers may simply reinitiate the medication after the patient is discharged.

To counter the overuse of gabapentinoids, hospitalists should look for opportunities to deprescribe them where there is concern about adverse events and when evidence-based indications do not exist. Successful deprescribing of these popular drugs will require deliberate collaboration and communication with the outpatient circle of care, as ongoing deprescribing ultimately depends on patients and outpatient prescribers agreeing to the change.

Disclosures

Dr. Steinman served as an unpaid expert witness in United States of America ex. Rel. David Franklin vs. Parke-Davis, Division of Warner-Lambert Company and Pfizer, Inc, litigation which alleged that the named pharmaceutical companies improperly marketed gabapentin for non-FDA-approved uses. Drs. Lam and Rochon have no conflicts of interest to declare.

Funding

Dr. Rochon is supported by the Retired Teachers of Ontario (RTO/ERO) Chair in Geriatric Medicine at the University of Toronto. Dr. Steinman is supported by the National Institute on Aging, US (K24AG049057 and P30AG044281).

Three years after gabapentin received US Food and Drug Administration (FDA) approval in 1990 for epilepsy, case reports and animal studies emerged announcing its potential in the treatment of pain syndromes through then-novel analgesic mechanisms.¹ Fast forward 20 years to 2016: gabapentin and its close cousin, pregabalin, are internationally considered first-line agents for the treatment of neuropathic pain in guidelines from the Centers for Disease Control and Prevention, the Canadian Pain Society, and the National Institute for Health and Care Excellence. Gabapentin is the 10th most prescribed drug in the United States, and brand-name pregabalin sales were $4.4 billion USD, ranking 8th in invoice drug spending.²

The ascendancy of gabapentinoids as drugs of choice for pain, though, is fraught with controversy; yet, they were shepherded to commercial success. In 2004, the patent owner of gabapentin, Warner-Lambert (now owned by Pfizer), admitted guilt to charges that it violated federal regulations in its promotion: they encouraged off-label prescribing through paid physician-to-physician communications, publication of positive outcomes, and suppression of negative ones.³ Pfizer paid another settlement in 2009 for false claims about off-label indications for brand-name pregabalin.⁴

Mindful of historical biases, recent trials and meta-analyses have found less favorable outcomes for gabapentinoids in the treatment of off-label pain conditions and greater risks than previously reported. Cochrane reviews for gabapentin demonstrate efficacy only in postherpetic neuralgia (for which it has FDA approval) and diabetic peripheral neuropathy (for which it does not); pregabalin has efficacy in both these conditions as well as posttraumatic neuropathic pain and fibromyalgia (and FDA approval for all four). For other types of neuropathic pain, the evidence is of lower quality. Even for approved indications, the risk–benefit ratio is questionable, as the numbers needed to harm for dizziness and somnolence are similar to the numbers needed to treat for pain.^5,6 Further, case–control studies have found increased odds of opioid-related death when gabapentinoids were coprescribed with opioids,^7,8 prompting gabapentinoids to be reclassified as class C controlled substances in the UK as of April 2019.⁹

On this backdrop, Gingras and colleagues publish their retrospective cohort study on high-risk prescribing of these popular drugs in Montreal, Canada in this issue of Journal of Hospital Medicine.¹⁰ In their retrospective cohort study of 4,103 patients admitted to a clinical teaching unit, more than one in eight patients (13%) were being prescribed a gabapentinoid as an outpatient; chart review of the admission notes indicated that only 17% of them had an FDA-approved indication and 28% had no clear indication. Gabapentinoid users were more likely to be coprescribed an opioid than nonusers (28% vs 12%). There was no significant difference in length of stay or inpatient death between users and nonusers.

Gingras et al. thereby conclude that there is an opportunity to deprescribe on the basis of few gabapentinoid users having a documented indication and the recent research showing potentials for harm and abuse.¹¹ We agree. Messaging around gabapentinoids should be similar to that for opioids: these are medications with limited evidence supporting their use in the treatment of chronic pain, and prescribing them for unapproved indications risks doing greater harm than good. We offer two recommendations on how hospitalists can proceed with deprescribing them safely.

First, the urgency of deprescribing in inpatient settings should be titrated to the degree of risk. When the reason for hospitalization is potentially an adverse drug effect, culprit medications posing a substantial and near-term risk of harm should be stopped, such as when patients on gabapentinoids present with major alteration of their mental status.

In less urgent circumstances, hospitalists should speak first with outpatient prescribers because they may have important contextual information (eg, indication, patient preference, failure of alternative therapies, etc.) about previous care that the inpatient clinician lacks. For gabapentinoids, it is easy to imagine how treated pain syndromes without objective markers of disease may escape notice by a hospitalist and remain undocumented, which may encourage erroneous deprescribing. If the shared decision between the patient and providers is to deprescribe, patients on high doses warrant a tapering schedule.¹¹ Pharmacist consultation can help with this.

Second, before discharge, hospitalists should communicate their rationale for deprescribing medications to both patients and outpatient prescribers, especially if a prolonged tapering schedule is required. This type of communication occurred infrequently in this study: the reason for deprescribing a gabapentinoid was missing from the discharge summary 55% of the time. Without this, outpatient prescribers may simply reinitiate the medication after the patient is discharged.

To counter the overuse of gabapentinoids, hospitalists should look for opportunities to deprescribe them where there is concern about adverse events and when evidence-based indications do not exist. Successful deprescribing of these popular drugs will require deliberate collaboration and communication with the outpatient circle of care, as ongoing deprescribing ultimately depends on patients and outpatient prescribers agreeing to the change.

Disclosures

Dr. Steinman served as an unpaid expert witness in United States of America ex. Rel. David Franklin vs. Parke-Davis, Division of Warner-Lambert Company and Pfizer, Inc, litigation which alleged that the named pharmaceutical companies improperly marketed gabapentin for non-FDA-approved uses. Drs. Lam and Rochon have no conflicts of interest to declare.

Funding

Dr. Rochon is supported by the Retired Teachers of Ontario (RTO/ERO) Chair in Geriatric Medicine at the University of Toronto. Dr. Steinman is supported by the National Institute on Aging, US (K24AG049057 and P30AG044281).

Rapid dissemination and adoption of evidence-based guidelines remains a challenge despite studies showing that key evidence-based care processes improve outcomes in sepsis and heart failure.¹ Hospital medicine was virtually founded on the premise that hospitalists would be champions of delivering high-quality care. Hospitalists are now dealing with a new challenge—unprecedented growth of healthcare systems because of mergers and acquisitions. The year 2018 was a banner time for healthcare mergers and acquisitions, with a total of 1,182, up 14% from 2017.² These are in response to the belief that healthcare systems may better navigate the mixed reimbursement models of fee-for-service and fee-for-value by achieving a larger patient base and economies of scale. Hospitalists must now achieve consistent, evidence-based standards of care across larger networks by educating their colleagues (often separated by large geographic areas) to manifest durable changes in their group practice with demonstrable improvement in patient outcomes and cost savings.

The study by Yurso et al. focused on implementing an education program, which included standardized learning through Clinical Performance and Value (CPV) vignettes with process measurement and feedback for sepsis and heart failure.³ Sepsis and heart failure have been a focus for treatment standardization because of the associated morbidity, mortality, and high cost of care. The study by Yurso et al. is a prospective quasi-controlled cohort of hospitalists in eight hospitals who were matched with comparator hospitalists in six nonparticipating hospitals across the AdventHealth system. Measurement and feedback were provided using CPV vignettes. Over two years, hospitalists who participated improved CPV scores by 8%, compliance with the utilization of the three-hour sepsis bundle from 46.0% to 57.5%, and orders of essential medical treatment elements for heart failure from 58.2% to 72.1%. In year one, the average length of stay (LOS) observed/expected (O/E) rates dropped by 8% for participating hospitalists compared with 2.5% in the comparator group. By year two, cost O/E rates improved slightly resulting in cost savings. The authors concluded that CPV case simulation-based measurement and feedback helped drive improvements in evidence-based care, which was associated with lower costs and shorter LOS.

While studies using traditional didactic CME struggle to demonstrate changes in practice leading to improved patient outcomes,⁴ the study by Yurso et al. gives a glimpse into how simulation can be used to help improve clinical performance and measure adherence to best practice. A remarkably similar study used CPV for simulated patients with serial performance measurement and feedback for heart failure and pneumonia. The study showed reduced practice variation between hospitalists at 11 hospitals across four states and decreased LOS and readmissions. However, the sole clinical outcome was no change in in-house mortality.⁵ Another study using CPV training in breast cancer treatment demonstrated increased adherence to evidence-based practice standards and decreased variation in care between providers across four states.⁶ Of note, this study did not include clinical outcomes. These studies collectively imply that simulation training with interactive learning, educational feedback, repetitive practice, and curriculum integration has shown modest success in creating practice change and improving adherence to best practice standards. However, they have minimal measures of patient outcomes and fairly simple analyses for cost savings. Because the education is computer-based and feedback can be performed remotely, it can be deployed across large and diverse growing healthcare systems. To really move the needle, future research in the field of simulation should identify optimal simulation methods and be designed with more rigor to include patient and cost outcomes.

At Intermountain Healthcare, hospitalist expansion occurred through a strategic realignment from the different geographic regions into the One Intermountain model. This model is built on the commitment that our patients will receive the same high-quality, high-value care wherever they walk through our doors. We have found four substantive changes have been particularly powerful in spurring a group practice mentality toward standardizing best practice. One, hospitalists are now aligned across the system under a single operational leadership structure that encourages combined efforts to share best practices and develop and deploy strategic initiatives around them. Two, hospitalists continue to build on a culture of quality and measure what matters to patients. While Intermountain Healthcare has a long history of using quality improvement to achieve better patient outcomes and lower costs,⁷ the new structure is allowing our group to test novel methods including redesigned education to see what actually improves adherence to best practice. Three, the group knows where the system’s reimbursement is coming from; Intermountain Healthcare has transitioned to a larger percentage of capitation,⁸ currently about 40%, with a strong commitment to partner with services geared to transition patients home quickly and keep them at home. Four, the organization has created a structure of accountability and reporting; an executive-sponsored systemwide operating model has been designed to cut through system barriers being identified by the frontline, allowing them to be rapidly surfaced and then solved at the executive level through daily huddles.⁹

Innovative educational programs such as the one described in the study by Yurso et al. that help the busy hospitalist achieve improved adherence to best practice are likely to be an important component leading to improved outcomes, but only after a group has been structured for success. As hospitalist groups continue to act as a single effector arm for high-value care, this will help meet the expectations of our patients and deliver on the promise of our field.

Disclosures: Dr. Srivastava is a physician founder of the I- PASS Patient Safety Institute. His employer, Intermountain Healthcare owns his equity in the I-PASS Patient Safety Institute. Dr. Srivastava is supported in part by the Children’s Hospital Association for his work as an executive council member of the Pediatric Research in Inpatient Settings (PRIS) network. Dr. Srivastava has received monetary awards, honorariums, and travel reimbursement from multiple academic and professional organizations for talks about pediatric hospitalist research networks and quality of care. All other authors have nothing to disclose. No funding was provided for this editorial.

Disclosures

The authors have no disclosures of financial conflicts of interest.

Funding

Dr. Walke was supported an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

Rapid dissemination and adoption of evidence-based guidelines remains a challenge despite studies showing that key evidence-based care processes improve outcomes in sepsis and heart failure.¹ Hospital medicine was virtually founded on the premise that hospitalists would be champions of delivering high-quality care. Hospitalists are now dealing with a new challenge—unprecedented growth of healthcare systems because of mergers and acquisitions. The year 2018 was a banner time for healthcare mergers and acquisitions, with a total of 1,182, up 14% from 2017.² These are in response to the belief that healthcare systems may better navigate the mixed reimbursement models of fee-for-service and fee-for-value by achieving a larger patient base and economies of scale. Hospitalists must now achieve consistent, evidence-based standards of care across larger networks by educating their colleagues (often separated by large geographic areas) to manifest durable changes in their group practice with demonstrable improvement in patient outcomes and cost savings.

The study by Yurso et al. focused on implementing an education program, which included standardized learning through Clinical Performance and Value (CPV) vignettes with process measurement and feedback for sepsis and heart failure.³ Sepsis and heart failure have been a focus for treatment standardization because of the associated morbidity, mortality, and high cost of care. The study by Yurso et al. is a prospective quasi-controlled cohort of hospitalists in eight hospitals who were matched with comparator hospitalists in six nonparticipating hospitals across the AdventHealth system. Measurement and feedback were provided using CPV vignettes. Over two years, hospitalists who participated improved CPV scores by 8%, compliance with the utilization of the three-hour sepsis bundle from 46.0% to 57.5%, and orders of essential medical treatment elements for heart failure from 58.2% to 72.1%. In year one, the average length of stay (LOS) observed/expected (O/E) rates dropped by 8% for participating hospitalists compared with 2.5% in the comparator group. By year two, cost O/E rates improved slightly resulting in cost savings. The authors concluded that CPV case simulation-based measurement and feedback helped drive improvements in evidence-based care, which was associated with lower costs and shorter LOS.

While studies using traditional didactic CME struggle to demonstrate changes in practice leading to improved patient outcomes,⁴ the study by Yurso et al. gives a glimpse into how simulation can be used to help improve clinical performance and measure adherence to best practice. A remarkably similar study used CPV for simulated patients with serial performance measurement and feedback for heart failure and pneumonia. The study showed reduced practice variation between hospitalists at 11 hospitals across four states and decreased LOS and readmissions. However, the sole clinical outcome was no change in in-house mortality.⁵ Another study using CPV training in breast cancer treatment demonstrated increased adherence to evidence-based practice standards and decreased variation in care between providers across four states.⁶ Of note, this study did not include clinical outcomes. These studies collectively imply that simulation training with interactive learning, educational feedback, repetitive practice, and curriculum integration has shown modest success in creating practice change and improving adherence to best practice standards. However, they have minimal measures of patient outcomes and fairly simple analyses for cost savings. Because the education is computer-based and feedback can be performed remotely, it can be deployed across large and diverse growing healthcare systems. To really move the needle, future research in the field of simulation should identify optimal simulation methods and be designed with more rigor to include patient and cost outcomes.

At Intermountain Healthcare, hospitalist expansion occurred through a strategic realignment from the different geographic regions into the One Intermountain model. This model is built on the commitment that our patients will receive the same high-quality, high-value care wherever they walk through our doors. We have found four substantive changes have been particularly powerful in spurring a group practice mentality toward standardizing best practice. One, hospitalists are now aligned across the system under a single operational leadership structure that encourages combined efforts to share best practices and develop and deploy strategic initiatives around them. Two, hospitalists continue to build on a culture of quality and measure what matters to patients. While Intermountain Healthcare has a long history of using quality improvement to achieve better patient outcomes and lower costs,⁷ the new structure is allowing our group to test novel methods including redesigned education to see what actually improves adherence to best practice. Three, the group knows where the system’s reimbursement is coming from; Intermountain Healthcare has transitioned to a larger percentage of capitation,⁸ currently about 40%, with a strong commitment to partner with services geared to transition patients home quickly and keep them at home. Four, the organization has created a structure of accountability and reporting; an executive-sponsored systemwide operating model has been designed to cut through system barriers being identified by the frontline, allowing them to be rapidly surfaced and then solved at the executive level through daily huddles.⁹

Innovative educational programs such as the one described in the study by Yurso et al. that help the busy hospitalist achieve improved adherence to best practice are likely to be an important component leading to improved outcomes, but only after a group has been structured for success. As hospitalist groups continue to act as a single effector arm for high-value care, this will help meet the expectations of our patients and deliver on the promise of our field.

Disclosures: Dr. Srivastava is a physician founder of the I- PASS Patient Safety Institute. His employer, Intermountain Healthcare owns his equity in the I-PASS Patient Safety Institute. Dr. Srivastava is supported in part by the Children’s Hospital Association for his work as an executive council member of the Pediatric Research in Inpatient Settings (PRIS) network. Dr. Srivastava has received monetary awards, honorariums, and travel reimbursement from multiple academic and professional organizations for talks about pediatric hospitalist research networks and quality of care. All other authors have nothing to disclose. No funding was provided for this editorial.

Disclosures

The authors have no disclosures of financial conflicts of interest.

Funding

Dr. Walke was supported an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

Rapid dissemination and adoption of evidence-based guidelines remains a challenge despite studies showing that key evidence-based care processes improve outcomes in sepsis and heart failure.¹ Hospital medicine was virtually founded on the premise that hospitalists would be champions of delivering high-quality care. Hospitalists are now dealing with a new challenge—unprecedented growth of healthcare systems because of mergers and acquisitions. The year 2018 was a banner time for healthcare mergers and acquisitions, with a total of 1,182, up 14% from 2017.² These are in response to the belief that healthcare systems may better navigate the mixed reimbursement models of fee-for-service and fee-for-value by achieving a larger patient base and economies of scale. Hospitalists must now achieve consistent, evidence-based standards of care across larger networks by educating their colleagues (often separated by large geographic areas) to manifest durable changes in their group practice with demonstrable improvement in patient outcomes and cost savings.

The study by Yurso et al. focused on implementing an education program, which included standardized learning through Clinical Performance and Value (CPV) vignettes with process measurement and feedback for sepsis and heart failure.³ Sepsis and heart failure have been a focus for treatment standardization because of the associated morbidity, mortality, and high cost of care. The study by Yurso et al. is a prospective quasi-controlled cohort of hospitalists in eight hospitals who were matched with comparator hospitalists in six nonparticipating hospitals across the AdventHealth system. Measurement and feedback were provided using CPV vignettes. Over two years, hospitalists who participated improved CPV scores by 8%, compliance with the utilization of the three-hour sepsis bundle from 46.0% to 57.5%, and orders of essential medical treatment elements for heart failure from 58.2% to 72.1%. In year one, the average length of stay (LOS) observed/expected (O/E) rates dropped by 8% for participating hospitalists compared with 2.5% in the comparator group. By year two, cost O/E rates improved slightly resulting in cost savings. The authors concluded that CPV case simulation-based measurement and feedback helped drive improvements in evidence-based care, which was associated with lower costs and shorter LOS.

While studies using traditional didactic CME struggle to demonstrate changes in practice leading to improved patient outcomes,⁴ the study by Yurso et al. gives a glimpse into how simulation can be used to help improve clinical performance and measure adherence to best practice. A remarkably similar study used CPV for simulated patients with serial performance measurement and feedback for heart failure and pneumonia. The study showed reduced practice variation between hospitalists at 11 hospitals across four states and decreased LOS and readmissions. However, the sole clinical outcome was no change in in-house mortality.⁵ Another study using CPV training in breast cancer treatment demonstrated increased adherence to evidence-based practice standards and decreased variation in care between providers across four states.⁶ Of note, this study did not include clinical outcomes. These studies collectively imply that simulation training with interactive learning, educational feedback, repetitive practice, and curriculum integration has shown modest success in creating practice change and improving adherence to best practice standards. However, they have minimal measures of patient outcomes and fairly simple analyses for cost savings. Because the education is computer-based and feedback can be performed remotely, it can be deployed across large and diverse growing healthcare systems. To really move the needle, future research in the field of simulation should identify optimal simulation methods and be designed with more rigor to include patient and cost outcomes.

At Intermountain Healthcare, hospitalist expansion occurred through a strategic realignment from the different geographic regions into the One Intermountain model. This model is built on the commitment that our patients will receive the same high-quality, high-value care wherever they walk through our doors. We have found four substantive changes have been particularly powerful in spurring a group practice mentality toward standardizing best practice. One, hospitalists are now aligned across the system under a single operational leadership structure that encourages combined efforts to share best practices and develop and deploy strategic initiatives around them. Two, hospitalists continue to build on a culture of quality and measure what matters to patients. While Intermountain Healthcare has a long history of using quality improvement to achieve better patient outcomes and lower costs,⁷ the new structure is allowing our group to test novel methods including redesigned education to see what actually improves adherence to best practice. Three, the group knows where the system’s reimbursement is coming from; Intermountain Healthcare has transitioned to a larger percentage of capitation,⁸ currently about 40%, with a strong commitment to partner with services geared to transition patients home quickly and keep them at home. Four, the organization has created a structure of accountability and reporting; an executive-sponsored systemwide operating model has been designed to cut through system barriers being identified by the frontline, allowing them to be rapidly surfaced and then solved at the executive level through daily huddles.⁹

Innovative educational programs such as the one described in the study by Yurso et al. that help the busy hospitalist achieve improved adherence to best practice are likely to be an important component leading to improved outcomes, but only after a group has been structured for success. As hospitalist groups continue to act as a single effector arm for high-value care, this will help meet the expectations of our patients and deliver on the promise of our field.

Disclosures: Dr. Srivastava is a physician founder of the I- PASS Patient Safety Institute. His employer, Intermountain Healthcare owns his equity in the I-PASS Patient Safety Institute. Dr. Srivastava is supported in part by the Children’s Hospital Association for his work as an executive council member of the Pediatric Research in Inpatient Settings (PRIS) network. Dr. Srivastava has received monetary awards, honorariums, and travel reimbursement from multiple academic and professional organizations for talks about pediatric hospitalist research networks and quality of care. All other authors have nothing to disclose. No funding was provided for this editorial.

Disclosures

The authors have no disclosures of financial conflicts of interest.

Funding

Dr. Walke was supported an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

Like the rest of the world, the United States is experiencing an aging boom. The number of adults aged 65 years or older is expected to grow from 49 million in 2016 to 82 million in 2040, indicating an increase of 67%. Even more impressively, the population of individuals aged 85 years or older is expected to increase by 129% to 14.6 million within this same time period.¹ Considering that one in five Medicare Fee for Service beneficiaries are hospitalized at least once a year,² hospitals can expect the number of adults over the age of 65 requiring acute care will substantially increase over the next 20 years. These demographic changes have important implications for the overall healthcare costs in the US. Of persons with the highest annual healthcare expenditures, 40% are 65 years of age or older. ³ Thus, optimizing the care of hospitalized older adults will remain a critical component in the management of healthcare costs in the next 20 years.

As such, the Acute Care for the Elderly (ACE) unit, an interprofessional model of care that has been shown to provide high-quality care to hospitalized older adults without increasing costs,⁴ will become an increasingly important component of acute care as the older adult population grows. In this edition of the Journal of Hospital Medicine, Brennan et al.⁵ describe a quality improvement initiative in which an interprofessional team that included a geriatric clinician, nurses, pharmacist, and chaplain developed a daily plan of care for ACE unit patients aged 70 years or older. The daily care plan, which focused on symptom management and advance care planning, was the nidus for collaboration between the hospital medicine attending and geriatrics team. Their results demonstrate that ACE unit patients had lower hospital costs and shorter lengths of stay (LOS) as compared with age-matched, usual-care patients despite having higher comorbidity scores. In addition, the greatest benefits were seen among persons in the highest quartile of the comorbidity score.

These results add to the small but consistent body of literature that demonstrates quality and cost benefits to the ACE unit care. Importantly, however, in contrast with the prior ACE unit studies in which persons with moderate risk were the ones to demonstrate the greatest benefits, Brennan et al.⁵ were able to demonstrate the greatest effect for the highest-need, highest-cost population. Reasons for this impressive effect may be attributed to this intervention’s specific emphasis on symptom management and estimated life expectancy. In an era when Medicare and other payers are looking to increase the value proposition in population health-based approaches by reducing high costs while preserving high quality, these findings represent an important example that merits a broader dissemination.

Of course, ACE units are not the only hospital-based programs that have shown to improve outcomes for older adults. The Hospital Elder Life Program (HELP) is an evidence-based delirium prevention intervention that has been shown to not only prevent delirium but also prevent cognitive and functional decline while decreasing hospital LOS, hospital falls, and sitter use.⁶ Moreover, similar to ACE units, HELP has been shown to reduce inhospital patient costs. Geriatrics surgery comanagement programs are another hospital-based intervention that has shown to improve outcomes for older surgical patients. Reductions in LOS, improved mobility, and higher discharge to home have been demonstrated in patients who have undergone spinal surgery.⁷ Decreased LOS and lower hospital costs have also been demonstrated among patients with hip fracture undergoing repair.⁸ Programs such as ACE units, HELP, and geriatric surgery comanagement are well aligned with the growing emphasis on value-based healthcare and will be especially needed by hospitals that strive to be high-reliability organizations as the number of adults aged 65 and older continues to grow. To date, few studies have explored the potential synergistic effects (or redundancies) of these programs and how to maximize the impact of these evidence-based interventions across healthcare systems with multiple hospitals that care for older adults from various socioeconomic and cultural backgrounds.

Looking toward the future, the implementation of ACE units and other innovative geriatric programs will equip hospitals to develop into Age-Friendly Health Systems (AFHS). AFHS is an initiative being led by the Institute for Healthcare Improvement, The John A. Hartford Foundation, the American Hospital Association, and the Catholic Health Association of the United States in partnership with several other leading healthcare organizations to provide high-value care to every older adult.⁹ AFHS provide care focused on the 4M framework—What Matters, Medications, Mobility, and Mentation. The goal is for 20% of hospitals and medical practices to join the AFHS initiative by 2020; to date, over 70 organizations nationwide have done so. Clearly, to reach this goal, and beyond, a greater collaboration between aging-focused interprofessional teams including geriatricians and hospitalists will be essential.

Given the aging demographic and rising healthcare costs, Brennan et al.’s work⁵ suggests that each hospital should have an ACE unit by 2040. Consistently, hospital care delivery has appropriately developed in response to the needs of the patient population served. Intensive care units (ICUs), dialysis units, and emergency rooms are just a few innovations that were adopted by hospitals to provide specialty care to individuals with complex acute illnesses. While technology within the ICU certainly plays a role in the care delivered in that setting, it could be argued that what makes the ICUs most effective is the cohorting of interprofessional expertise. Since the implementation of ICUs, the survival rate for critically ill patients has substantially improved and additional specialty units with an interprofessional team model, eg, cardiac care units, dialysis units, emergency rooms, etc., have followed suit. Specialty units have become a part of the fabric of acute care, so much so that it would be hard to imagine a modern hospital without an ICU, dialysis unit, or emergency room. The same should be true for ACE units. Even hospitals without geriatricians on site can use teleconferencing to successfully implement an ACE unit.¹⁰ We owe it to our older patients to transform our institutions into AFHS; implementing models of care proven to improve outcomes, such as the ACE unit, is one of the critical first steps.

Disclosures

The authors have no disclosures or financial conflicts of interest.

Funding

Dr. Walke was supported by an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

Like the rest of the world, the United States is experiencing an aging boom. The number of adults aged 65 years or older is expected to grow from 49 million in 2016 to 82 million in 2040, indicating an increase of 67%. Even more impressively, the population of individuals aged 85 years or older is expected to increase by 129% to 14.6 million within this same time period.¹ Considering that one in five Medicare Fee for Service beneficiaries are hospitalized at least once a year,² hospitals can expect the number of adults over the age of 65 requiring acute care will substantially increase over the next 20 years. These demographic changes have important implications for the overall healthcare costs in the US. Of persons with the highest annual healthcare expenditures, 40% are 65 years of age or older. ³ Thus, optimizing the care of hospitalized older adults will remain a critical component in the management of healthcare costs in the next 20 years.

As such, the Acute Care for the Elderly (ACE) unit, an interprofessional model of care that has been shown to provide high-quality care to hospitalized older adults without increasing costs,⁴ will become an increasingly important component of acute care as the older adult population grows. In this edition of the Journal of Hospital Medicine, Brennan et al.⁵ describe a quality improvement initiative in which an interprofessional team that included a geriatric clinician, nurses, pharmacist, and chaplain developed a daily plan of care for ACE unit patients aged 70 years or older. The daily care plan, which focused on symptom management and advance care planning, was the nidus for collaboration between the hospital medicine attending and geriatrics team. Their results demonstrate that ACE unit patients had lower hospital costs and shorter lengths of stay (LOS) as compared with age-matched, usual-care patients despite having higher comorbidity scores. In addition, the greatest benefits were seen among persons in the highest quartile of the comorbidity score.

These results add to the small but consistent body of literature that demonstrates quality and cost benefits to the ACE unit care. Importantly, however, in contrast with the prior ACE unit studies in which persons with moderate risk were the ones to demonstrate the greatest benefits, Brennan et al.⁵ were able to demonstrate the greatest effect for the highest-need, highest-cost population. Reasons for this impressive effect may be attributed to this intervention’s specific emphasis on symptom management and estimated life expectancy. In an era when Medicare and other payers are looking to increase the value proposition in population health-based approaches by reducing high costs while preserving high quality, these findings represent an important example that merits a broader dissemination.

Of course, ACE units are not the only hospital-based programs that have shown to improve outcomes for older adults. The Hospital Elder Life Program (HELP) is an evidence-based delirium prevention intervention that has been shown to not only prevent delirium but also prevent cognitive and functional decline while decreasing hospital LOS, hospital falls, and sitter use.⁶ Moreover, similar to ACE units, HELP has been shown to reduce inhospital patient costs. Geriatrics surgery comanagement programs are another hospital-based intervention that has shown to improve outcomes for older surgical patients. Reductions in LOS, improved mobility, and higher discharge to home have been demonstrated in patients who have undergone spinal surgery.⁷ Decreased LOS and lower hospital costs have also been demonstrated among patients with hip fracture undergoing repair.⁸ Programs such as ACE units, HELP, and geriatric surgery comanagement are well aligned with the growing emphasis on value-based healthcare and will be especially needed by hospitals that strive to be high-reliability organizations as the number of adults aged 65 and older continues to grow. To date, few studies have explored the potential synergistic effects (or redundancies) of these programs and how to maximize the impact of these evidence-based interventions across healthcare systems with multiple hospitals that care for older adults from various socioeconomic and cultural backgrounds.

Looking toward the future, the implementation of ACE units and other innovative geriatric programs will equip hospitals to develop into Age-Friendly Health Systems (AFHS). AFHS is an initiative being led by the Institute for Healthcare Improvement, The John A. Hartford Foundation, the American Hospital Association, and the Catholic Health Association of the United States in partnership with several other leading healthcare organizations to provide high-value care to every older adult.⁹ AFHS provide care focused on the 4M framework—What Matters, Medications, Mobility, and Mentation. The goal is for 20% of hospitals and medical practices to join the AFHS initiative by 2020; to date, over 70 organizations nationwide have done so. Clearly, to reach this goal, and beyond, a greater collaboration between aging-focused interprofessional teams including geriatricians and hospitalists will be essential.

Given the aging demographic and rising healthcare costs, Brennan et al.’s work⁵ suggests that each hospital should have an ACE unit by 2040. Consistently, hospital care delivery has appropriately developed in response to the needs of the patient population served. Intensive care units (ICUs), dialysis units, and emergency rooms are just a few innovations that were adopted by hospitals to provide specialty care to individuals with complex acute illnesses. While technology within the ICU certainly plays a role in the care delivered in that setting, it could be argued that what makes the ICUs most effective is the cohorting of interprofessional expertise. Since the implementation of ICUs, the survival rate for critically ill patients has substantially improved and additional specialty units with an interprofessional team model, eg, cardiac care units, dialysis units, emergency rooms, etc., have followed suit. Specialty units have become a part of the fabric of acute care, so much so that it would be hard to imagine a modern hospital without an ICU, dialysis unit, or emergency room. The same should be true for ACE units. Even hospitals without geriatricians on site can use teleconferencing to successfully implement an ACE unit.¹⁰ We owe it to our older patients to transform our institutions into AFHS; implementing models of care proven to improve outcomes, such as the ACE unit, is one of the critical first steps.

Disclosures

The authors have no disclosures or financial conflicts of interest.

Funding

Dr. Walke was supported by an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

Like the rest of the world, the United States is experiencing an aging boom. The number of adults aged 65 years or older is expected to grow from 49 million in 2016 to 82 million in 2040, indicating an increase of 67%. Even more impressively, the population of individuals aged 85 years or older is expected to increase by 129% to 14.6 million within this same time period.¹ Considering that one in five Medicare Fee for Service beneficiaries are hospitalized at least once a year,² hospitals can expect the number of adults over the age of 65 requiring acute care will substantially increase over the next 20 years. These demographic changes have important implications for the overall healthcare costs in the US. Of persons with the highest annual healthcare expenditures, 40% are 65 years of age or older. ³ Thus, optimizing the care of hospitalized older adults will remain a critical component in the management of healthcare costs in the next 20 years.

As such, the Acute Care for the Elderly (ACE) unit, an interprofessional model of care that has been shown to provide high-quality care to hospitalized older adults without increasing costs,⁴ will become an increasingly important component of acute care as the older adult population grows. In this edition of the Journal of Hospital Medicine, Brennan et al.⁵ describe a quality improvement initiative in which an interprofessional team that included a geriatric clinician, nurses, pharmacist, and chaplain developed a daily plan of care for ACE unit patients aged 70 years or older. The daily care plan, which focused on symptom management and advance care planning, was the nidus for collaboration between the hospital medicine attending and geriatrics team. Their results demonstrate that ACE unit patients had lower hospital costs and shorter lengths of stay (LOS) as compared with age-matched, usual-care patients despite having higher comorbidity scores. In addition, the greatest benefits were seen among persons in the highest quartile of the comorbidity score.

These results add to the small but consistent body of literature that demonstrates quality and cost benefits to the ACE unit care. Importantly, however, in contrast with the prior ACE unit studies in which persons with moderate risk were the ones to demonstrate the greatest benefits, Brennan et al.⁵ were able to demonstrate the greatest effect for the highest-need, highest-cost population. Reasons for this impressive effect may be attributed to this intervention’s specific emphasis on symptom management and estimated life expectancy. In an era when Medicare and other payers are looking to increase the value proposition in population health-based approaches by reducing high costs while preserving high quality, these findings represent an important example that merits a broader dissemination.

Of course, ACE units are not the only hospital-based programs that have shown to improve outcomes for older adults. The Hospital Elder Life Program (HELP) is an evidence-based delirium prevention intervention that has been shown to not only prevent delirium but also prevent cognitive and functional decline while decreasing hospital LOS, hospital falls, and sitter use.⁶ Moreover, similar to ACE units, HELP has been shown to reduce inhospital patient costs. Geriatrics surgery comanagement programs are another hospital-based intervention that has shown to improve outcomes for older surgical patients. Reductions in LOS, improved mobility, and higher discharge to home have been demonstrated in patients who have undergone spinal surgery.⁷ Decreased LOS and lower hospital costs have also been demonstrated among patients with hip fracture undergoing repair.⁸ Programs such as ACE units, HELP, and geriatric surgery comanagement are well aligned with the growing emphasis on value-based healthcare and will be especially needed by hospitals that strive to be high-reliability organizations as the number of adults aged 65 and older continues to grow. To date, few studies have explored the potential synergistic effects (or redundancies) of these programs and how to maximize the impact of these evidence-based interventions across healthcare systems with multiple hospitals that care for older adults from various socioeconomic and cultural backgrounds.

Looking toward the future, the implementation of ACE units and other innovative geriatric programs will equip hospitals to develop into Age-Friendly Health Systems (AFHS). AFHS is an initiative being led by the Institute for Healthcare Improvement, The John A. Hartford Foundation, the American Hospital Association, and the Catholic Health Association of the United States in partnership with several other leading healthcare organizations to provide high-value care to every older adult.⁹ AFHS provide care focused on the 4M framework—What Matters, Medications, Mobility, and Mentation. The goal is for 20% of hospitals and medical practices to join the AFHS initiative by 2020; to date, over 70 organizations nationwide have done so. Clearly, to reach this goal, and beyond, a greater collaboration between aging-focused interprofessional teams including geriatricians and hospitalists will be essential.

Given the aging demographic and rising healthcare costs, Brennan et al.’s work⁵ suggests that each hospital should have an ACE unit by 2040. Consistently, hospital care delivery has appropriately developed in response to the needs of the patient population served. Intensive care units (ICUs), dialysis units, and emergency rooms are just a few innovations that were adopted by hospitals to provide specialty care to individuals with complex acute illnesses. While technology within the ICU certainly plays a role in the care delivered in that setting, it could be argued that what makes the ICUs most effective is the cohorting of interprofessional expertise. Since the implementation of ICUs, the survival rate for critically ill patients has substantially improved and additional specialty units with an interprofessional team model, eg, cardiac care units, dialysis units, emergency rooms, etc., have followed suit. Specialty units have become a part of the fabric of acute care, so much so that it would be hard to imagine a modern hospital without an ICU, dialysis unit, or emergency room. The same should be true for ACE units. Even hospitals without geriatricians on site can use teleconferencing to successfully implement an ACE unit.¹⁰ We owe it to our older patients to transform our institutions into AFHS; implementing models of care proven to improve outcomes, such as the ACE unit, is one of the critical first steps.

Disclosures

The authors have no disclosures or financial conflicts of interest.

Funding

Dr. Walke was supported by an award from the Health Resources and Services Administration Geriatric Workforce Enhancement Program to the University of Pennsylvania (U1QHP28720).

In the context of rapidly rising healthcare costs and increasing disparities in health outcomes in the United States, there has been increasing interest in identifying and addressing the needs of our country’s most frequently admitted patients. These patients account for a disproportionate percentage of healthcare expenditures^1-3; they also represent a vulnerable and high-risk population. Finding solutions to address the needs of these patients is important for the patients themselves and for the systems in which they receive care. The last 10-15 years have seen a proliferation of programs working to address the needs and contain the costs of frequently admitted patients,^2,4-6 as well as increased interest in understanding the risk factors and drivers that lead to high utilization.

In this edition of the Journal of Hospital Medicine, O’Leary et al. report on their study of patients enrolled in the CHAMP program at Northwestern University, in which the authors elicit patients’ perceptions of factors contributing to the onset and continuation of high hospital use.⁷ The authors identify several themes, including the important role of psychological, social, and economic factors in course fluctuation, the perception of acute illness as uncontrollable and unpredictable, and a strong desire to avoid hospitalization. As a group, the themes suggest multiple strategies that may be of use in developing individualized plans for patients.

Several of the most commonly cited risk factors for high utilization—including mental health issues, housing insecurity or homelessness, and substance use^2,3,8,9—did not emerge as themes identified by patients in this study as contributing to high hospital utilization. Although identified themes such as social support and psychological stress could certainly be related to these underlying risk factors, the risk factors themselves did not emerge. This is particularly notable in a population whose utilization is in line with other studies (participants had at least two unplanned 30-day inpatient readmissions within 12 months, and one readmission in the last six months, a referral, or at least three observation visits). In contrast to prior qualitative work with complex, high-needs patients,¹⁰ patients in this study did not identify difficult (or positive) relationships with care provider teams, or a history of early life trauma, as factors related to current utilization.

These findings raise several important questions. To what extent are frequently hospitalized patient populations comparable with each other? This is both a question about how populations are defined and a question about the inherent variability between populations (including geographic, social, socioeconomic, and other factors). It is not evident from the demographic information provided whether this population is fundamentally different from others that have been studied, or whether risk factors such as mental health issues, housing insecurity, substance abuse, and trauma history are present, but are just not identified by patients here as proximal contributors to their utilization. In either case, the findings raise important questions about the development of effective interventions for these patients. The discrepancies also highlight the utility of ascertaining and reporting the prevalence of these risk factors among study populations, ideally both among patients who opt in and those who opt out. Although obtaining this information adds an additional layer of complexity to data collection, this history, along with extended demographic data, would significantly improve our ability to assess the comparability of populations across studies. It would also help us understand whether perspectives of any specific groups of patients are not represented, due to frequent opting out of the study.

The fact that commonly identified risk factors for high utilization are not identified by patients in this study as contributing to their high hospital use highlights the importance of (1) including the patient perspective as an integral part of care plan and intervention development and (2) continuing local work aimed at understanding the risk factors and drivers of high utilization in specific populations. Many programs, including CHAMP at Northwestern and our own hospitalist-run program at Penn Medicine, work closely with patients to develop individualized care plans that aim to address the underlying drivers of high utilization. In our experience, a multidisciplinary committee reviewing patient cases has identified mental health conditions as likely drivers of frequent admissions in over 95% of program patients. In line with the findings here, however, patients themselves often do not see mental health as a significant contributor. If patients do not see factors such as mental health as important, this has significant implications for the development of interventions around these factors as part of a solution to high hospital use.

Patients are unlikely to respond to interventions targeting problems that they themselves do not identify as important. This is not to say that drivers such as mental health, housing instability, substance abuse, and behaviors rooted in childhood trauma cannot be addressed if they are not identified by a patient as problems. Rather, interventions must be sensitive to and developed within the context of the patient’s own perceptions and priorities. For any program aimed at addressing the underlying drivers of high utilization, therefore, it is critical to elicit individual patient perspectives and to incorporate them in the development of interventions tailored to a specific patient’s needs. This process not only informs the creation of an individualized care plan but also promotes engagement and builds trust.

In prior work,⁶ O’Leary et al. have joined others throughout the field in calling for standardized definitions of “high utilizers”; this is critical for our ability to compare study results across programs. However, standardizing definitions is just the first step. Individual site studies such as this are needed to help us understand which themes are universal, versus those that are population- and site-specific. They are also important for individual institutions in targeting, developing, and refining local interventions. As a whole, the results will help guide the development of best practices within the field and allow providers to better understand the needs of specific populations. This work is essential to our ability as providers, hospitals, and systems to develop effective interventions for individual patients in this heterogeneous, vulnerable, and high-risk population.

Disclosures

Dr. Knox and Dr. Greysen have nothing to disclose.

In the context of rapidly rising healthcare costs and increasing disparities in health outcomes in the United States, there has been increasing interest in identifying and addressing the needs of our country’s most frequently admitted patients. These patients account for a disproportionate percentage of healthcare expenditures^1-3; they also represent a vulnerable and high-risk population. Finding solutions to address the needs of these patients is important for the patients themselves and for the systems in which they receive care. The last 10-15 years have seen a proliferation of programs working to address the needs and contain the costs of frequently admitted patients,^2,4-6 as well as increased interest in understanding the risk factors and drivers that lead to high utilization.

In this edition of the Journal of Hospital Medicine, O’Leary et al. report on their study of patients enrolled in the CHAMP program at Northwestern University, in which the authors elicit patients’ perceptions of factors contributing to the onset and continuation of high hospital use.⁷ The authors identify several themes, including the important role of psychological, social, and economic factors in course fluctuation, the perception of acute illness as uncontrollable and unpredictable, and a strong desire to avoid hospitalization. As a group, the themes suggest multiple strategies that may be of use in developing individualized plans for patients.

Several of the most commonly cited risk factors for high utilization—including mental health issues, housing insecurity or homelessness, and substance use^2,3,8,9—did not emerge as themes identified by patients in this study as contributing to high hospital utilization. Although identified themes such as social support and psychological stress could certainly be related to these underlying risk factors, the risk factors themselves did not emerge. This is particularly notable in a population whose utilization is in line with other studies (participants had at least two unplanned 30-day inpatient readmissions within 12 months, and one readmission in the last six months, a referral, or at least three observation visits). In contrast to prior qualitative work with complex, high-needs patients,¹⁰ patients in this study did not identify difficult (or positive) relationships with care provider teams, or a history of early life trauma, as factors related to current utilization.

These findings raise several important questions. To what extent are frequently hospitalized patient populations comparable with each other? This is both a question about how populations are defined and a question about the inherent variability between populations (including geographic, social, socioeconomic, and other factors). It is not evident from the demographic information provided whether this population is fundamentally different from others that have been studied, or whether risk factors such as mental health issues, housing insecurity, substance abuse, and trauma history are present, but are just not identified by patients here as proximal contributors to their utilization. In either case, the findings raise important questions about the development of effective interventions for these patients. The discrepancies also highlight the utility of ascertaining and reporting the prevalence of these risk factors among study populations, ideally both among patients who opt in and those who opt out. Although obtaining this information adds an additional layer of complexity to data collection, this history, along with extended demographic data, would significantly improve our ability to assess the comparability of populations across studies. It would also help us understand whether perspectives of any specific groups of patients are not represented, due to frequent opting out of the study.

The fact that commonly identified risk factors for high utilization are not identified by patients in this study as contributing to their high hospital use highlights the importance of (1) including the patient perspective as an integral part of care plan and intervention development and (2) continuing local work aimed at understanding the risk factors and drivers of high utilization in specific populations. Many programs, including CHAMP at Northwestern and our own hospitalist-run program at Penn Medicine, work closely with patients to develop individualized care plans that aim to address the underlying drivers of high utilization. In our experience, a multidisciplinary committee reviewing patient cases has identified mental health conditions as likely drivers of frequent admissions in over 95% of program patients. In line with the findings here, however, patients themselves often do not see mental health as a significant contributor. If patients do not see factors such as mental health as important, this has significant implications for the development of interventions around these factors as part of a solution to high hospital use.

Patients are unlikely to respond to interventions targeting problems that they themselves do not identify as important. This is not to say that drivers such as mental health, housing instability, substance abuse, and behaviors rooted in childhood trauma cannot be addressed if they are not identified by a patient as problems. Rather, interventions must be sensitive to and developed within the context of the patient’s own perceptions and priorities. For any program aimed at addressing the underlying drivers of high utilization, therefore, it is critical to elicit individual patient perspectives and to incorporate them in the development of interventions tailored to a specific patient’s needs. This process not only informs the creation of an individualized care plan but also promotes engagement and builds trust.

In prior work,⁶ O’Leary et al. have joined others throughout the field in calling for standardized definitions of “high utilizers”; this is critical for our ability to compare study results across programs. However, standardizing definitions is just the first step. Individual site studies such as this are needed to help us understand which themes are universal, versus those that are population- and site-specific. They are also important for individual institutions in targeting, developing, and refining local interventions. As a whole, the results will help guide the development of best practices within the field and allow providers to better understand the needs of specific populations. This work is essential to our ability as providers, hospitals, and systems to develop effective interventions for individual patients in this heterogeneous, vulnerable, and high-risk population.

Disclosures

Dr. Knox and Dr. Greysen have nothing to disclose.

In the context of rapidly rising healthcare costs and increasing disparities in health outcomes in the United States, there has been increasing interest in identifying and addressing the needs of our country’s most frequently admitted patients. These patients account for a disproportionate percentage of healthcare expenditures^1-3; they also represent a vulnerable and high-risk population. Finding solutions to address the needs of these patients is important for the patients themselves and for the systems in which they receive care. The last 10-15 years have seen a proliferation of programs working to address the needs and contain the costs of frequently admitted patients,^2,4-6 as well as increased interest in understanding the risk factors and drivers that lead to high utilization.

In this edition of the Journal of Hospital Medicine, O’Leary et al. report on their study of patients enrolled in the CHAMP program at Northwestern University, in which the authors elicit patients’ perceptions of factors contributing to the onset and continuation of high hospital use.⁷ The authors identify several themes, including the important role of psychological, social, and economic factors in course fluctuation, the perception of acute illness as uncontrollable and unpredictable, and a strong desire to avoid hospitalization. As a group, the themes suggest multiple strategies that may be of use in developing individualized plans for patients.

Several of the most commonly cited risk factors for high utilization—including mental health issues, housing insecurity or homelessness, and substance use^2,3,8,9—did not emerge as themes identified by patients in this study as contributing to high hospital utilization. Although identified themes such as social support and psychological stress could certainly be related to these underlying risk factors, the risk factors themselves did not emerge. This is particularly notable in a population whose utilization is in line with other studies (participants had at least two unplanned 30-day inpatient readmissions within 12 months, and one readmission in the last six months, a referral, or at least three observation visits). In contrast to prior qualitative work with complex, high-needs patients,¹⁰ patients in this study did not identify difficult (or positive) relationships with care provider teams, or a history of early life trauma, as factors related to current utilization.

These findings raise several important questions. To what extent are frequently hospitalized patient populations comparable with each other? This is both a question about how populations are defined and a question about the inherent variability between populations (including geographic, social, socioeconomic, and other factors). It is not evident from the demographic information provided whether this population is fundamentally different from others that have been studied, or whether risk factors such as mental health issues, housing insecurity, substance abuse, and trauma history are present, but are just not identified by patients here as proximal contributors to their utilization. In either case, the findings raise important questions about the development of effective interventions for these patients. The discrepancies also highlight the utility of ascertaining and reporting the prevalence of these risk factors among study populations, ideally both among patients who opt in and those who opt out. Although obtaining this information adds an additional layer of complexity to data collection, this history, along with extended demographic data, would significantly improve our ability to assess the comparability of populations across studies. It would also help us understand whether perspectives of any specific groups of patients are not represented, due to frequent opting out of the study.

The fact that commonly identified risk factors for high utilization are not identified by patients in this study as contributing to their high hospital use highlights the importance of (1) including the patient perspective as an integral part of care plan and intervention development and (2) continuing local work aimed at understanding the risk factors and drivers of high utilization in specific populations. Many programs, including CHAMP at Northwestern and our own hospitalist-run program at Penn Medicine, work closely with patients to develop individualized care plans that aim to address the underlying drivers of high utilization. In our experience, a multidisciplinary committee reviewing patient cases has identified mental health conditions as likely drivers of frequent admissions in over 95% of program patients. In line with the findings here, however, patients themselves often do not see mental health as a significant contributor. If patients do not see factors such as mental health as important, this has significant implications for the development of interventions around these factors as part of a solution to high hospital use.

Patients are unlikely to respond to interventions targeting problems that they themselves do not identify as important. This is not to say that drivers such as mental health, housing instability, substance abuse, and behaviors rooted in childhood trauma cannot be addressed if they are not identified by a patient as problems. Rather, interventions must be sensitive to and developed within the context of the patient’s own perceptions and priorities. For any program aimed at addressing the underlying drivers of high utilization, therefore, it is critical to elicit individual patient perspectives and to incorporate them in the development of interventions tailored to a specific patient’s needs. This process not only informs the creation of an individualized care plan but also promotes engagement and builds trust.

In prior work,⁶ O’Leary et al. have joined others throughout the field in calling for standardized definitions of “high utilizers”; this is critical for our ability to compare study results across programs. However, standardizing definitions is just the first step. Individual site studies such as this are needed to help us understand which themes are universal, versus those that are population- and site-specific. They are also important for individual institutions in targeting, developing, and refining local interventions. As a whole, the results will help guide the development of best practices within the field and allow providers to better understand the needs of specific populations. This work is essential to our ability as providers, hospitals, and systems to develop effective interventions for individual patients in this heterogeneous, vulnerable, and high-risk population.

Disclosures

Dr. Knox and Dr. Greysen have nothing to disclose.