. Author manuscript; available in PMC: 2017 Dec 6.

Published in final edited form as: JAMA. 2016 Dec 6;316(21):2237–2252. doi: 10.1001/jama.2016.17573

Table 1.

Study comparison and outcome measures

Study	Design	Population	Comparison(s)	Sample Size(s)	Outcome Measures	Main Results	P Value
Pass/Fail Grading Systems
Bloodgood et al,¹⁴ 2009	Nonrandomized 2 group	First- and second- year medical students	Cohort with pass/fail grading system versus earlier cohort with 5-interval grading system (A/B/C/D/F)	n = 281 Pass/fail = 140 5-interval = 141	• Questions regarding satisfaction with school, satisfaction with personal life • Dupuy General Well-Being Schedule^a	Pass/fail versus graded after semester 1 (scores) • Anxiety: 18.14 (versus 15.98) • Depression: 17.62 (versus 15.89) • Well-being: 13.02 (versus 11.02) • Self-control: 15.51 (versus 14.12) • Vitality: 14.60 (versus 12.15) • General health: 12.56 (versus 11.48) Pass/fail versus graded after semester 2 (scores) • Anxiety: 19.01 (versus 17.65) • Depression: 17.61 (versus 16.65) • Well-being: 13.09 (versus 12.20) • Self-control: 15.10 (versus 14.45) • Vitality: 15.16 (versus 13.31) • General health: 11. 99 (versus 11.24) Pass/fail versus graded after semester 3 (scores) • Anxiety: 17.02 (versus 14.55) • Depression: 16.92 (versus 15.08) • Well-being: 12.37 (versus 10.74) • Self-control: 15.13 (versus 14.40) • Vitality: 14.10 (versus 11.95) • General health: 11.25 (versus 10.84) Pass/fail versus graded after semester 4 (scores) • Anxiety: 14.08 (versus 14.20) • Depression: 15.56 (versus 15.35) • Well-being: 10.59 (versus 10.40) • Self-control: 14.61 (versus 14.42) • Vitality: 12.88 (versus 12.06) • General health: 11.30 (versus 11.31)	.002 < .001 < .001 < .001 < .001 .02 .05 .05 .03 .13 .001 .15 .001 .001 < .001 .08 < .001 .47 .86 .71 .67 .63 .11 .99
Rohe et al,¹⁵ 2006	Nonrandomized 2 group	First- and second- year medical students	Cohort with pass/fail grading system versus earlier cohort with 5-interval grading system (A/B/C/D/F)	n = 81 Pass/fail = 40 5-interval = 41	• Perceived Stress Scale^a • Profile of Mood States^a • Perceived Cohesion Scale^a	Pass/fail versus graded at end of first year (score): • Perceived Stress Scale: 10.9 (SD 6.2) versus 13.8 (SD 6.4) • Profile of Mood States: 13.0 (SD 23.5) versus 32.0 (SD 39.0) • Perceived Cohesion Scale: 37.8 (SD 5.5) versus 32.9 (SD 8.4) Pass/fail versus graded at end of second year (score): • Perceived Stress Scale: 15.8 (SD 6.8) versus 20.5 (SD 7.8) • Profile of Mood States: 47.1 (SD 31.9) versus 64.6 (SD 40.5) • Perceived Cohesion Scale: 33.8 (SD 8.0) versus 29.0 (SD 9.9)	.02 .02 .01 .01 .07 .02
Reed et al,¹⁹ 2011	Nonrandomized ≥ 2 groups	First- and second- year medical students	Institutions with pass/fail grading systems versus institutions with 3+ interval grading systems (eg, honors/pass/fail)	n = 2,056 n = 1,192 (responded) Pass/fail = 701 3+ interval = 491	• Maslach Burnout Inventory^a • Perceived Stress Scale^a • Medical Outcomes Study Short Form^a	• Perceived stress score: β = 1.91; 95% CI, 1.05 to 2.78 • Mental quality of life: β = − 2.79; 95% CI, −4.09 to −1.5 • Burnout: OR 1.85; 95% CI, 1.24 to 2.01 • Seriously considered dropping out of medical school in the past year: OR 1.91; 95% CI, 1.30 to 2.80	< .001 < .001 < .001 .001
Mental Health Programs
Thompson et al,²⁰ 2010	Nonrandomized 2 group	Third-year medical students	Cohort with multi-pronged mental health program versus earlier cohort without the program	n = 120 Program cohort = 62 Earlier cohort = 58	• Center for Epidemiologic Studies Depression Scale^a question on suicidal ideation	Pre-/post-intervention (frequency) • Depressive symptoms: 26/44 (59.1%) versus 14/58 (24.1%); χ² = 12.84; df = 2 • Suicidal ideation: 13/43 (30.2%) versus 1/33 (3.0%); χ² = 13.05; df = 1	< .01 < .001
Seritan et al,²¹ 2013	Nonrandomized ≥ 2 group	All years	Cohort with mental health program versus earlier cohort without program and versus national average	No sample size provided for number of students referred to services Accreditation Council for Graduate Medical Education Graduation Survey, n = 525	• American Medical Colleges Graduation Questionnaire^a	Mental health service self-referral (percentage, no numbers provided) • Time 1: 50% • Time 2: 88% • Time 3: 91% Other referral • Time 1: 50% • Time 2: 12% • Time 3: 9% Satisfaction with program versus national average, pre- and most recent postintervention (score) • Personal counseling • 2009: 3.5 (3.7) • 2013: 4.4 (4.0) • Student mental health services • 2009: 3.5 (3.6) • 2013: 4.3 (4.0) • Stress-management programs (postintervention only) • 2009: 3.6 (3.8) • 2013: 4.3 (3.9)	---
Downs et al,²² 2014	Single group cross- sectional or post- test only	All years	---	n = 1008 (program) n = 343 (program and screen)	• Patient Health Questionnaire- 9^a	Among those screened, mental health service utilization (percentage, no numbers provided) • Year 1: 11.5% • Year 4: 15.0% • χ²: 1.27, df = 3 Among those screened, suicide risk • Year 1: 8.8% • Year 4: 6.2% • χ² = 0.45; df = 3	--- --- NS --- NS ---
Moutier et al,²³ 2012	Single group cross- sectional or post- test only	All years	---	n = 498 n = 132 (screened)	• Items from Patient Health Questionnaire-9^a measuring mental health service referral rate	Referred to mental health professional based in part on Patient Health Questionnaire -9, of those screened: 15/132 (11%)	---
Mind-Body Skills Education/Training Programs
Erogul et al,²⁴ 2014	Randomized clinical trial	First-year medical students	Mindfulness- based stress reduction intervention versus control (randomized)	n = 58 Intervention = 28 Control = 30	• Perceived Stress Scale^a • Self-Compassion Scale^a	Change in case from pre- to postintervention (change score) • Perceived Stress Scale: 3.63; 95% CI, 0.37 to 6.89 • Self-Compassion Scale: 0.58; 95% CI, 0.23 to 0.92 Change in case from pre- to 6- month follow-up (change score) • Perceived Stress Scale: 2.91; 95% CI, −0.37 to 6.19 • Self-Compassion Scale: 0.56; 95% CI, 0.25 to 0.87	.03 .002 .08 .001
Holtzworth- Munroe et al,²⁶ 1985	Randomized clinical trial	First- and second- year medical students	Mind-body program versus control (randomized)	n = 40 Intervention = 20 Control = 20	• Spielberger Trait Anxiety Inventory^a • Anxiety in test and social situation questionnaire • Tension and depression questionnaire • Self-esteem measure • Stress questionnaire	Intervention versus control at follow-up (score) • More aware of tension: F(5, 18) = 37.16 • Dealing better with school stress: F(5, 18) = 5.05 Anxiety before test (score): F(1, 22) = 10.42	< .001 < .04 < .005
Kraemer et al,²⁷ 2016	Nonrandomized 2 group	First- and second- year medical students	Mind-body program versus control (non- randomized)	n = 52 Intervention = 28 Control = 24	• Distress Tolerance Scale^a • Perceived Stress Scale-10^a • Positive Affect Negative Affect Schedule^a	Changes in distress tolerance (change score) • Mind-body: 0.53; t = − 2.81; 95% CI, 0.92 to 0.14 • Control: 0.25; t = −1.66; 95% CI, −0.06 to 0.55 •	.01 .11
Rosenzweig et al,²⁹ 2003	Nonrandomized 2 group	Second- year medical students	Mindfulness- based stress reduction program versus control (non- randomized)	n = 302 Intervention = 140 Control = 162	• Profile of Mood States^a	Profile of Mood States total mood disturbance for intervention versus control (score) • Intervention: 38.7 (SD 33.3) versus 31.8 (SD 33.8) • Control: 28.0 (SD 31.2) versus 38.6 (SD 32.8) • Interaction: d = −0.18	.05 <.001 < .001
Finkelstein et al,³¹ 2007	Nonrandomized 2 group	Second- year medical students	Mind-body elective versus control (nonrandomized)	n = 72 Intervention = 26 Control = 46	• Symptom Checklist-90 Anxiety Subscale^a • Profile of Mood States^a • Perceived Stress of Medical School Scale^a • The 2-item Depression Index^a	Time/group interaction for scores • Anxiety (Symptom Checklist-90): F(1,2) = 3.95 • The Profile of Mood States: F(1,2) = 3.77 • Perceived Stress of Medical School Scale: F(1,2) = .11	< .05 < .05 NS
Greeson et al,³² 2015	Single group pre- and post-test	All years	Before versus after mind-body skills intervention	n = 44	• Cognitive and Affective Mindfulness Scale-Revised^a • Perceived Stress Scale^a • Open-ended feedback	Pre-/post-intervention (score) • Perceived Stress Scale: 29.73 (SD 9.61) versus 20.25 (SD 9.03) ◦ t (33) = 7.90; d = 1.38 • Mindfulness: 29.24 (SD 5.54) versus 33.88 (SD 6.13) ◦ t (33) = 5.27; d = 0.92	< .001 < .001
Bond et al,³³ 2013	Single group pre- and post-test	First- and second- year medical students	Before versus after mind-body course	n = 27	• Cohen’s Perceived Stress Scale • Self-regulation questionnaire^a • Self-Compassion Scale^a • Jefferson Scale of Physician Empathy^a	Pre-/post-intervention (change score): • Perceived stress: −.05 (SD 0.62); d = .14 • Self-regulation: 0.13 (SD 0.2); d = −0.41 • Self-compassion: 0.28 (SD 0.61); d = −0.55 • Empathy: 0.11 (SD 0.5); d = −0.13	.70 .003 .04 .30
Curriculum Structure
Reed et al,¹⁹ 2011	Nonrandomized; ≥ 2 groups	First- and second- year medical students	7 institutions’ curriculum structures	n = 2056 n = 1192 (responded)	• Maslach Burnout Inventory^a • Perceived Stress Scale^a • Medical Outcomes Study Short Form^a	Association between clinical experiences and the following scores • Perceived stress: β .02; 95% CI [−.10 to .13] • Burnout: OR 1.01; 95% CI, 0.98 to 1.05 • Mental quality of life: β .00; 95% CI, −0.16 to 0.16 • Serious thoughts of dropping out: OR 0.96; 95% CI, 0.93 to 1.00 Association between testing experiences and the following scores • Perceived stress: β .29; 95% CI, 0.10 to 0.84 • Burnout: OR 1.10; 95% CI, 0.89 to 1.23 • Mental quality of life: b–β − .63; 95% CI, −0.29 to 0.96 • Serious thoughts of dropping out: OR 1.19; 95% CI, 1.12 to 1.27 Association between number of tests and the following scores • Perceived stress: β −0.02; 95% CI, −0.6 to 0.03 • Burnout: OR 0.99; 95% CI, 0.97 to 1.01 • Mental quality of life: β 0.03; 95% CI, −0.05 to 0.04 • Serious thoughts of dropping out: OR 1.00; 95% CI, 0.97 to 1.02	.79 .42 .98 .03 .003 .09 < .001 < .001 .48 .19 .44 .82
Camp et al,³⁵ 1994	Nonrandomized 2 group	First- and second- year medical students	Problem-based learning versus lecture-based learning	n = 275 Problem-based learning = 60 Lecture-based learning = 215	• Zung Self-Rating Depression Scale^a	Depression problem-based learning versus lecture-based learning (score) • Overall OR 0.42; 95% CI, 0.14 to 1.21 • Adjustment for sex and self actualization OR 0.45; 95% CI, 0.14 to 1.42	.07 .14
Kornitzer et al,³⁶ 2005	Cross-sectional post-test only	All cohorts	---	n = 92	• Questions regarding program attendance factors, subjective medical school transition factors, program ratings and student perceptions, and academic benefits of program	Underrepresented in Medicine group (percentage, no numbers provided) • Gained confidence: 85.7% • Made the transition easier: 100% • Made friends:100% Humanities and Medicine group (percentage, no numbers provided) • Gained confidence: 97% • Made the transition easier: 97% • Made friends: 93.9%	--- --- --- --- --- ---
Multi-component Program Reform
Drolet and Rodgers,³⁷ 2010	Single group cross- sectional or post- test only	All years	---	n = 116	• Satisfaction survey	Student Wellness Committee satisfaction (percentage, no numbers provided) • Positive experience with Student Wellness Committee: 95%	---
Fleming et al,³⁸ 2013	Single group cross- sectional or post- test only (for the outcome measure relevant to this review)	All years	---	n = 245	• Vanderbilt University student affairs survey	Reported that colleges design contributed meaningfully or somewhat meaningfully to their Vanderbilt University experience (percentage, no numbers provided): 91%	---
Real et al,³⁹ 2015	Single group cross- sectional or post- test only	All years	---	n = 450	• Maslach Burnout Inventory^a • Primary Care Evaluation of Mental Disorders^a • Participation survey • Perception of burnout survey	Level of burnout within aspects of program (Score: 0 = more burnout, 100 = less burnout) • Faculty mentors: 70 • Annual retreats: 58.6 • Student-led programming committee: 64 • Overall wellness program: 69.2 Faculty mentors correlation with the following scores • Emotional exhaustion: r = −0.27 • Depersonalization: r = − 0.22 • Personal accomplishment: r = 0.19 Annual retreats correlation with the following scores • Emotional exhaustion: r = −0.32 • Depersonalization: r = − 0.32 • Personal accomplishment: r = 0.16 Student-led programming committee correlation with the following scores • Emotional exhaustion: r = −0.31 • Depersonalization: r = − 0.3 • Personal accomplishment: r = 0.23 Overall wellness program correlation with the following scores • Emotional exhaustion: r = −0.32 • Depersonalization: r = − 0.23 • Personal accomplishment: r = 0.1	--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---
Slavin et al,⁴⁰ 2014	Nonrandomized ≥ 2 group	First- and second- year medical students	Cohorts with different phases of multi- program reform implementation	n = 875–890	• Center for Epidemiological Studies Depression Scale^a • Spielberger State-Trait Anxiety Inventory^a • Perceived Stress Scale^a • Perceived Cohesion Scale^a • American Medical Colleges Graduation Questionnaire^a	Moderate or severe depression according to the Center for Epidemiological Studies Depression Scale (percentage, no numbers provided) • End year 1: Cramér V = 0.16 • Phase 1 versus control: 21% (versus 27%) • Phase 1/2 versus control: 18% (versus 27%) • Phase 1/2/3 versus control: 11% (versus 27%) • End year 2: Cramér V = 0.18 • Phase 1 versus control: 17% (versus 32%) • Phase 1/2 versus control: 18% (versus 32%) • Phase 1/2/3 versus control: 16% (versus 32%) Anxiety (percentage, no numbers provided) • End year 1: Cramér V = 0.23 • Phase 1 versus control: 45% (versus 55%) • Phase 1/2 versus control: 31% (versus 55%) • Phase 1/2/3 versus control: 31% (versus 55%) • End year 2: Cramér V = 0.18 • Phase 1 versus control: 61% (versus 60%) • Phase 1/2 versus control: 39% (versus 60%) • Phase 1/2/3 versus control: 46% (versus 60%) Stress according (score) • End year 1: η² = 0.06 • Phase 1 versus control: 14.9 (SD 6.7) versus 16.3 (SD 7.4) • Phase 1/2 versus control: 13 (SD 6.8) versus 16.3 (SD 7.4) • Phase 1/2/3 versus control: 12.1 (SD 6.1) versus 16.3 (SD 7.4) • End year 2: partial η² = 0.05 • Phase 1 versus control: 14.4 (SD 5.8) versus 16.9 (SD 7.3) • Phase 1/2 versus control: 13.9 (SD 6.4) versus 16.9 (SD 7.3) • Phase 1/2/3 versus control: 13.5 (SD 6.8) versus 16.9 (SD 7.3) Cohesion (score) • End year 1: partial η² = 0.03 • Phase 1 versus control: 8.1 (SD 1.7) versus 7.9 (SD 2.1) • Phase 1/2 versus control: 8.5 (SD 2.1) versus 7.9 (SD 2.1) • Phase 1/2/3 versus control: 8.8 (SD 1.8) versus 7.9 (SD 2.1) • End year 2: partial η² = 0.02 • Phase 1 versus control: 8.2 (SD 1.6) versus 7.7 (SD 2.0) • Phase 1/2 versus control: 8.5 (SD 2.2) versus 7.7 (SD 2.0) • Phase 1/2/3 versus control: 8.1 (SD 2.0) versus 7.7 (SD 2.0)	NS NS <.05 < .05 < .05 < .0 NS < .05 < .05 NS < .05 < .05 NS < .05 < .05 < .05 < .05 < .05 NS < .05 <0.5 < .05 < .05 < .05
Strayhorn,⁴¹ 1989	Nonrandomized 2 group	First-year medical students	Cohort with multi-component program reform versus earlier cohort University of North Carolina, Chapel Hill versus comparison school	Responders n = 478 (original sample size not reported)	• Learning Environment Questionnaire^a • Rand Health Insurance Questionnaires^a • Environment stresses questionnaire • Social support questionnaire	New versus old curriculum stress questionnaire • Overall fewer stresses; t(223) = −1.7 • Less perceived stress from social and recreational sources • No reduction in financial- related stress New versus old curriculum mental well-being • Greater overall well- being t(197) = −2.04 • Greater sense of positive well-being • Greater sense of vitality • Less depression • Less anxiety • Social well-being t (223) = −1.66 New versus old curriculum social support • No perceived difference in availability of social supports t (227) = −0.36 • Less class advisor support • Class advisors less willing to listen • Class advisors less willing to help with personal problems • Concerned about students welfare • Greater support from administrators • Could rely on administrators when things got tough • Perceived level of support from • Fellow students • Friends • Significant others Time - control (University of North Carolina, Chapel Hill versus comparison) learning environment • Fewer environmental stressors F(1467) = 6.41 • Greater mental well- being F(1460) = 9.32 • Greater social well-being F(1466) = 5.37 • No difference in social support F(1477) = 0.01	.09 .03 --- .04 < .001 < .001 < .001 < .001 .10 .721 .002 .003 < .001 .003 .05 .01 --- --- --- .01 .002 .02 .91
Miscellaneous Wellness Programs
Whitehouse et al,⁴² 1996	Randomized clinical trial	First-year medical students	Self-hypnosis intervention versus control (randomized)	n = 35 Intervention = 21 Control = 14	• Medical history • Profile of Mood States • Brief Symptom Inventory • University of California, Los Angeles Loneliness Scale	Time-group intervention analysis of score • Brief Symptom Inventory Anxiety: F[3, 96] = 2.96 At examination period self- hypnosis subjects significantly lower stressfulness scores: t (30) = 2.11	< .05 < .05
Goetzel et al,⁴⁴ 1984	Single group cross- sectional or post- test only	First-, second-, and third- year medical students	---	n = 26	• Group Environment Scale	Agreement with statement on scale (Likert 1–5): • “I am no longer as lonely; I feel more together with people”: 3.33 (of 5)	---
Lee and Graham,⁴⁵ 2001	Single group cross- sectional or post- test only	First- and second- year medical students	---	n = 66	• Questionnaire related to the wellness elective	Students appreciated that the Wellness Elective helped them realize the importance of personal well-being, gave permission for self-care and an opportunity to find collegiality, and provided various coping strategies (frequency) • 4/22 (18.2%) strongly agree • 17/22 (77.3%) agree Students felt that the Wellness Elective over-emphasized stress itself and devalued the worth of hard work; realistic expectations offered in this course seemed discouraging (frequency) •1/22 (4.5%) agree	---
Kushner et al,⁴⁶ 2011	Single group cross- sectional or post- test only	Second- year medical students	---	n = 343 (9 related to mental and emotional health)	• Form relating to goal and achievement	Self-reported achievement of mental/emotional health behavior change goals (frequency) • 6/9 (66.7%) agree	---
Group-based Faculty Advisor/Mentor Programs
Sastre et al,⁴⁷ 2010	Nonrandomized 2 group	First-, second-, and third- year medical students	Cohort with Advisory College Program versus earlier cohort with Faculty Advisory Program	n = 318 Cohort with program = 103 Earlier cohort = 215	• Questionnaires on perceived effectiveness of the system and role of advisor in promoting wellness and career counseling	Advisory College Program versus Faculty Advisory Program wellness advising (percentage, no numbers provided) • I feel comfortable discussing my personal stress with my advisor: 62% versus 24%; χ² = 40.9 • I feel comfortable discussing my mental health with my advisor: 51% versus 27%; χ² = 31.84 Satisfaction with how well advisors promoted wellness (percentage, no numbers provided: 27% versus 72%	< .001 < .001 <.001
Coates et al,⁴⁸ 2008	Nonrandomized 2 group	Fourth-year medical students	Cohort with mentoring program versus earlier cohort	n = 100 Cohort with program = 70 Earlier cohort = 30	• 25-item telephone survey	Cohort with mentoring program versus earlier cohort • Feels connected with faculty (frequency): 14/30 (47%) versus 49/70 (70%) • Feels connected with classmates (frequency): 11/30, (37%) versus 30/70 (43%)	--- ---
Ficklin et al,⁴⁹ 1983	Single group cross- sectional or post- test only	First-year medical students	---	n = 151	• Survey assessing 12 personal needs of first- year medical students	Program helpfulness (only descriptive summary of results provided) • Becoming better acquainted with peers • Becoming close to some classmates • Helping students with the anxieties of starting school	--- --- ---

Literature describing the development and validation of the various scales, scores, and questionnaires are as follows: Dupuy General Well-being Schedule,¹⁶ Perceived Stress Scale,^17,66 Profile of Mood States,³⁰ Perceived Cohesion Scale,¹⁸ Maslach Burnout Inventory,⁶⁷ Medical Outcomes Study Short Form,^68,69 Center for Epidemiologic Studies Depression Scale,⁷⁰ American Medical Colleges Graduation Questionnaire,⁷¹ Patient Health Questionnaire,^72,73 Self-Compassion Scale,²⁵ Spielberger Trait Anxiety Inventory,⁷⁴ Distress Tolerance Scale,²⁸ Positive Affect Negative Affect Schedule,⁷⁵ Symptom Checklist-90 Anxiety Subscale,⁷⁶ Perceived Stress of Medical School Scale,⁷⁷ 2-Item Depression Index,⁷⁸ Cognitive and Affective Mindfulness Scale- Revised,⁷⁹ Self-Regulation Questionnaire,³⁴ Jefferson Scale of Physician Empathy,⁸⁰ Zung Self-Rating Scale,⁸¹ Primary Care Evaluation of Mental Disorders,^72,82 Learning Environment Questionnaire,⁸³ Rand Health Insurance Questionnaires,⁸⁴ Brief Symptom Inventory,⁴³ University of California, Los Angeles Loneliness Scale,⁸⁵ Group Environmental Scale.^86–88

CI, confidence interval; OR, odds ratio; SD, standard deviation.