Mental Health Conditions Among Struggling GME Learners: Results From a Single Center Remediation Program Open Access

Surveys of program directors (PDs) and single center reviews reported that 3.5% to 12% of learners struggle with clinical performance during residency training.^1–4  A subset of these trainees have underlying mental health conditions (MHCs).^1–3,5,6  In a survey of internal medicine residency PDs, stress and depression were identified by 42% and 24% of PDs, respectively, as being present half the time or more among struggling learners.¹  A follow-up survey of internal medicine PDs reported that 32.6% of residents experiencing difficulty had contributing factors such as depression, anxiety, and personality disorders, and 6.6% had learning disabilities.²  Guerrasio et al identified mental well-being as a deficit in approximately 18% of residents referred to a single center remediation program.⁶  These studies offered little detail on how and by whom mental health diagnoses were made, which calls into question their accuracy and raises the possibility that some diagnoses were missed. This represents a gap in our understanding of the true prevalence of MHC among struggling learners.

Building on previously published remediation programs,^6,7  the University of Virginia (UVA) developed COACH (Committee for Achieving Competence through Help) in 2016 to assist graduate medical education (GME) learners who struggle with clinical performance. COACH is 1 of 2 programs at UVA that provide support to trainees and is the only program primarily focused on trainee performance. All learners suspected of having a substance use disorder are referred to the Faculty and Employee Assistance Program. Learners may be involved simultaneously with both programs. The primary aim of our study was to characterize the MHCs associated with unsatisfactory trainee performance and their relationship to specific clinical performance deficits (CPDs).

This study utilized a retrospective cohort design. Between 2016 and 2019, residents and fellows struggling with clinical performance at UVA were given the option to self-refer or be referred to COACH by their PD. Our institution has a formal remediation process (policy provided as online supplemental material). Referral to COACH may be recommended as part of the formal remediation plan or may occur independently. For PD-led referrals, whether or not the learner is on formal remediation as defined by the GME office, COACH communicates regularly with the PD and provides a written assessment and remediation plan. For self-referrals, there is no communication between COACH and the PD.

COACH includes 7 physicians and 1 clinical psychologist with experience in remediation. The chair of COACH, a physician remediation expert with formalized training in the recognition and management of mental well-being and substance use disorders among physicians, assessed each learner using a biopsychosocial approach.⁷  CPDs were categorized as follows: medical knowledge, clinical reasoning, organization/efficiency, professionalism, communication, and other. Following assessment and identification of the CPDs, the chair of COACH, with input of other COACH faculty, developed an individualized written remediation plan. At the conclusion of remediation, the learner was reassessed by a group independent of the remediation process, typically the departmental clinical competency committee. Participation in the COACH program was voluntary.

As part of the initial assessment, the remediation expert questioned the learner about current and previous history of depression, anxiety, psychosocial stressors, substance use, and cognitive impairment. If an MHC was suspected during this initial assessment, and the learner was not already engaged in treatment, the learner was given a choice to pursue mental health services through COACH or a local mental health professional, or to not pursue treatment. Information related to the diagnosis or treatment of MHC is confidential and not shared with the PD.

We utilized a licensed mental health provider to diagnose learners using a multimethod assessment. If learners chose to pursue treatment with the COACH psychologist, they underwent a psychiatric diagnostic interview, including psychological measures that assess mood (Beck Depression Inventory–II⁸ ), anxiety (Beck Anxiety Inventory⁹ ), personality (Personality Assessment Inventory¹⁰ ), attention deficit and hyperactivity (Conners' Adult ADHD Rating Scale–Self Report: Long Version¹¹ ), and social support (Social Support Scale¹² ). Mental health diagnoses were based on DSM-5 (Diagnostic and Statistical Manual of Mental Disorders–5) criteria, as well as assessment findings of the above psychological instruments. Recommendations were provided regarding the need for psychotherapy, medication management, and/or neuropsychological testing.

Learners were tracked by COACH for outcomes including termination, voluntarily departure from the program, and distinction of “in good standing” (ie, not on formal remediation) or “not in good standing” (ie, on formal remediation as defined by GME policy).

This project was deemed exempt by the University of Virginia Institutional Review Board for the Social and Behavioral Sciences.

Over a 3-year period, COACH assessed 61 GME learners struggling with clinical performance out of a total of 820 (7%). Residents and fellows were referred from 14 different departments. Sixty-four percent (39 of 61) of learners were referred by their PD, and 36% (22 of 61) were self-referrals. Thirty-nine percent (24 of 61) had more than 1 primary CPD. CPDs that were identified included professionalism (41%, 25 of 61), organization/efficiency (36%, 22 of 61), clinical reasoning (34%, 21 of 61), medical knowledge (25%, 15 of 61), communication (5%, 3 of 61), operative skill (3%, 2 of 61), and systems-based practice (2%, 1 of 61).

Thirty-eight percent (23 of 61) of learners were found to have an MHC associated with the CPD, including anxiety (48%, 11 of 23), depression (17%, 4 of 23), and cognitive dysfunction (17% [4 of 23] in all cases ADHD). None were diagnosed with a substance use disorder (figure 1). The distribution of CPDs among learners with MHCs is illustrated in figure 2.

Learners with MHCs were referred for (74%, 17 of 23) or already engaged in (26%, 6 of 23) therapeutic intervention for the MHC. All learners were offered a remediation program for the CPD. One learner was terminated prior to initiation of remediation. Three learners declined remediation. Of the remaining learners, 47% (9 of 19) have successfully finished remediation, 21% (4 of 19) were terminated or voluntarily left their program, and 32% (6 of 19) are still being remediated. Of learners still being remediated, 83% (5 of 6) are in good standing, and 17% (1 of 6) are not in good standing, a distinction assigned by the GME Office at UVA for learners on formal remediation.

Several studies have reported the coexistence of mental well-being issues and struggle with clinical performance among GME learners.^1–3,5,6  A survey from Yao and Wright¹  found that 24% of PDs identify depression as an underlying cause in 50% or more of struggling learners, but did not identify a prevalence of depression. At our program, 38% of learners struggling with clinical performance had an MHC, similar to survey findings from Dupras et al, in which 32.6% of residents experiencing difficulty had contributing factors such as depression, anxiety, and personality disorders, and 6.6% had learning disabilities,²  and higher than the identification of mental well-being as a deficit in approximately 18% of residents referred to a single center remediation program by Guerrasio et al.⁶  This program categorized mental well-being as a primary type of performance deficit, rather than as a contributing factor,⁶  highlighting the importance of mental well-being among struggling learners, but making it difficult to determine its relationship with other primary deficits. Our experience is that there is commonly overlap between MHC and other CPDs, and this overlap requires a specialized approach that takes into account their coexistence.

The 2 PD surveys relied on recollection of PDs and did not detail their methods of assessing mental well-being.^1,2  The University of Colorado incorporated a mental health professional on their remediation team but provides no details on the assessment of mental well-being.⁶  In contrast, our study offers a novel approach, using a multimethod assessment by a licensed mental health professional and separate categorization of mental well-being issues and CPDs. These findings underscore the increasing realization of the major role that mental well-being plays in the struggle with clinical performance and adds important detail that allows for a better understanding of its prevalence and relationship to specific CPDs.

Our study is limited by its small sample size and the fact that it is conducted at a single center. Our data only capture those who were referred to our program, which is new and growing rapidly, and may underestimate the true prevalence of struggling learners at UVA. While we have descriptive data on learner outcomes, we lack objective data such as milestone scores pre- and post-remediation. Future studies should examine statistical differences among learners with and without MHCs as well as the relationship between unprofessional behavior and MHCs. Finally, because there is another program at UVA to help trainees with substance use disorder, our study likely underreports the incidence of it among struggling learners.

In this study, we describe the prevalence of MHCs and CPDs in a struggling learner population within a single center and depict the distribution of CPDs among learners with MHCs. MHCs represent an important factor impacting the performance of struggling learners. At present, literature has not described specific diagnosed MHCs experienced by GME learners. Given the intersection of MHCs and CPDs, early identification of specific MHCs is a vital step to assist struggling learners.