Shapiro, Shauna L. MA; Shapiro, Daniel E. PhD; Schwartz, Gary E. R. PhD
Medical education has deleterious consequences. Trainees (students, interns, and residents) suffer high levels of stress, which lead to alcohol and drug abuse,1 interpersonal relationship difficulties,2 depression and anxiety,3,4 and even suicide.5 Medical students have mean anxiety scores one standard deviation above those of non-patients, and their depression levels increase significantly throughout the first year of medical school.6 Stress may also harm trainees' professional effectiveness: it decreases attention,7 reduces concentration,8 impinges on decision-making skills,9,10 and reduces trainees' abilities to establish strong physician-patient relationships.11
To address these problems, programs have been changed in a variety of ways, including reducing the work week,12 instituting curricular reforms (e.g., smaller classes, less rote memorization), and providing psychological services such as couples counseling, child care services, social activities, support groups, and stress-reduction programs.13
A decade ago, after a comprehensive literature review of stress in medical education, Butterfield concluded that “the body of literature on effective interventions needs to be expanded.”14 Unfortunately, this recommendation has not been followed. Despite numerous articles that decry the negative consequences of stress and call for intervention and change, few have studied the specific effects of stress-management interventions in medical education, and even fewer have provided empirical data. Although there is a large literature on stress management in general, its specific application to medical education has been largely unexplored.
To fill this gap, in this article we review the literature on stress-management programs in medical education; specifically, programs providing trainees with coping techniques (such as meditation, hypnosis, imagery, and muscle relaxation), education regarding the psychological and physiological effects of stress, affiliation with peers and opportunities for emotional expression (support groups), and intensified relationships with faculty. We then discuss implications for the integration of stress management in medical training and make suggestions for future research.
While the literature is replete with papers discussing stress reduction or describing specific programs, we found only 24 studies that reported data. Table 1 provides a three-part matrix describing these 24 studies across eight categories: participants, randomization, control groups, structure of intervention, content of intervention, follow-up, outcome measures, and results. Part I contains the six rigorously designed studies, Part II has the one study that violated randomization, and the remaining 18 studies comprise Part III.
As can be seen in the table, the studies looked at a heterogeneous group of programs that blended a variety of interventions and means of delivery. Of the 24 studies, only seven used control groups or attempted to randomize participants. This diversity and lack of consistent method makes drawing firm conclusions difficult.
Below we discuss each column of the matrix, highlighting key points and issues and discussing implications for future directions. We begin our review of the results by addressing the question, “Are stress-management programs in medical education effective?” We then refine our analysis by examining three aspects of the studies: (1) participants and methods, (2) interventions, and (3) outcome measures.
The 24 studies—the earliest one published in 1969, the latest in 1998— showed that the stress-management programs were helpful psychologically and/or physiologically, and virtually all trainees who responded in all studies were in favor of the programs' being offered regularly or integrated into the curriculum. Authors reported that participating trainees demonstrated improved immunologic functioning15,16; decreases in depression and anxiety15,17; increases in spirituality and empathy17; enhanced knowledge of alternative therapies for future referrals18; improved knowledge about stress19; improved sensitivity toward themselves,20 their peers,20,21 and their patients20; reduced perceptions of isolation20; greater use of positive coping skills and less use of negative coping skills22; and resolution of professional role conflicts.21
Only four studies reviewed found no difference between experimental and control groups on standardized measures of psychological functioning, immune functioning, or health at post-assessment.15,19,23,24 However, of these four only one conducted any follow-up assessment, and this one did find a significant positive change in the intervention group compared to the control group. Trainee achievement, as assessed by examination performance, was consistently not correlated with participation in stress-reduction programs.
Although all of the studies reported results based on assessments immediately after the interventions, only four20,24–26 assessed trainees beyond the end of the intervention. Of these, three reassessed participants a few months later and one assessed subjects one year later. While follow-ups supported the effectiveness of the interventions, the durations of stress-management effects for the majority of programs remain unclear.
Fifteen of the studies focused on medical students, one included both residents and medical students, and the remaining eight focused on residents. Half of the studies allowed participants in different years of training to participate in the same programs. The other half focused on single years of training, most commonly the first year of medical school or residency. Because of the heterogeneity of the programs described, it is difficult to draw conclusions about how these inclusion decisions influenced outcomes.
Most of the studies based their conclusions on small samples. The majority of the studies had samples under 30; the smallest had six participants.26 Such small sample sizes made it difficult to find statistically meaningful results, make generalizations, and rule out type II errors (incorrectly concluding that an intervention is not effective).
Finally, studies varied in their recruitment of participants. Most stress-management programs recruited volunteers, but they may not have reached those trainees most in need. As noted by Reuben,27 severely impaired residents are least likely to use support systems such as groups, mental health counselors, or faculty mentors. Other stress-management programs made attendance mandatory.28 This was more common in programs for residents than in programs for medical students. A minority of participants in the mandatory programs resented the requirement; one study found that some residents felt that attending a support group indicated decreased competence.28 Future research should compare the results of mandatory and volunteer interventions. In addition, given Reuben's observation, it may be particularly important to screen and then target those trainees found to be most impaired, as they appear to be the least likely to volunteer.
Only seven of the studies15–17,22–24,28 randomly assigned participants, and only one of those28 used stratified sampling to ensure that equal numbers of different participants, in that case first-, second-, and third-year family practice residents, would be included in each group. One study19 described a procedure that actively violated randomization: students were recruited on a first-come-first-served basis, with those who called earlier being assigned to the experimental group and those who called later (who may have been less organized), to the control group. The rest of the studies did not describe how participants were assigned to groups.
With eight exceptions,15–17,19,22–24,29 the studies reported did not use control groups. Most studies used pre/post designs, comparing baseline and post-intervention ratings. Unfortunately, there is ample evidence that stress levels fluctuate considerably during training. Medical students report experiencing considerably more stress during examination periods, and interns and residents report varying levels of stress depending on assigned rotations, on-call schedules, or time off. Given these factors, it is likely that pre/post designs are vulnerable to these fluctuations. An example illustrates this problem. Shapiro and colleagues17 reported no change in state anxiety levels (a measure of current anxiety, as opposed to trait anxiety, which measures characteristics or general anxiety) for the experimental group from before to after the intervention. However, significant differences in state anxiety were found between groups (experimental and control) after the intervention. The post-intervention assessment coincided with the examination period, suggesting that the stress-management intervention had buffered students against the negative effects of examination stress. If the study had not included a control group, the intervention would have erroneously appeared ineffective.
In some cases not using control groups is understandable. Researchers have to balance the value of information gleaned from control groups against the possible negative consequences for medical trainees enrolled in wait-list controls, inert control groups, or comparison groups suspected to be less effective than the experimental groups. Of those who did use control groups, five used wait-list control groups,15–17,19,29 two compared interventions,22,23 and one had a no-treatment control group.24
Structure of the intervention. The majority of the programs used a group structure where trainees met with peers or with leaders. Interventions varied in the amounts of time required of participants, from two consecutive intensive days30 to weekly hour-long meetings throughout the academic year.13 Participants tended to meet from one to two hours.
The optimal duration for interventions, both in frequency and duration of meetings, is unknown. No study systematically examined what length or intensity of intervention was most effective. Considering the demands already made on trainees' time, many of the programs required only modest additional time commitments. However, this also meant that only a few programs provided long-term support. Controlled empirical research needs to determine what length of intervention is most beneficial, considering the trainees' rigorous time commitments. Paradoxically, stress-management programs may initially elicit greater levels of stress.22 Therefore, a cost-benefit analysis should be performed to determine what intensity is most effective both short- and long-term.
The facilitators of the interventions varied in terms of training, background, and experience level. In selecting group leaders, one should consider a few issues. On the one hand, group leaders with training experiences similar to those of the participants will be familiar with their specific stresses and struggles.32 On the other hand, effective group facilitation is not a simple task33 and should be conducted by those with professional training, such as social workers, psychologists, and psychiatrists.34 Another problem with using facilitators with similar experiences is that many have dual relationships with the participants. For example, in some cases the leaders of groups had grading responsibility for the participants,35,36 a clear violation of basic support-group theory. Further, some stress-management interventions are based on specialized training and the facilitator must be an expert in order to effectively teach the stress-management skills (e.g., hypnosis or meditation). As a result, researchers frequently had to choose between selecting facilitators who had training similar to that of the participants, those who had training in group leadership, or those who had specialized training.
Content of the intervention. No “gold standard” exists for the content of stress-reduction programs for medical trainees. As in stress-management programs offered in other workplaces,37 content varied considerably. Our review revealed that a wide variety of interventions were included under the umbrella of stress reduction; e.g., directed and non-directed support groups, relaxation training (including meditation and hypnosis), time-management and coping skills, mindfulness-based stress reduction, and mentoring programs. The groups also varied in the degrees to which they encouraged emotional expression, incorporated personal as well as professional issues, and focused on technique versus discussion.
The most common outcome measure used in the studies was the trainees' evaluations of the stress-management programs. The trainees almost universally found the programs helpful and in many cases urged researchers to integrate the programs into the curriculum. Most medical schools rely heavily on students' evaluations when making teaching assignments and rotation selections.
Although students' evaluations are vital as an outcome measure, more objective measures (e.g., behavioral and physiologic measures) are needed to provide more comprehensive assessment of outcomes. Unfortunately, systematic evaluation of other outcomes was scarce. Only seven studies used standardized measures. Most researchers relied on non-standardized inventories they had constructed themselves. The problems with non-standardized measures are well known to behavioral scientists and are not discussed here.
Future researchers might focus on a number of outcomes that have yet to be examined. For example, what is the influence of stress-management programs on physician-patient communication? How does stress management for physician trainees influence patient outcomes? Do programs with stress-management interventions have less trainee dropout?
We believe one unintended and unfortunate side effect of medical training is that it produces physicians who believe that self-denial is valuable and necessary and that living under stress is normal. Until physicians recognize the significant health impact stress has on them and the importance of modeling healthy behaviors (including relative stress-free living) for their patients, it is unlikely that the skills learned in a stress-management program will be incorporated into their professional and personal lives.
One reason that most of the studies did not use validated measures is that there is no “gold standard” for assessment of stress management. Researchers studying stress tend to rely on checklists that ask respondents to report the number of symptoms they are experiencing or the severity of events they have experienced. The most widely employed include the Unpleasant Events Schedule,38 the Hassles Scale,39 the Assessment of Daily Experiences Questionnaire,40 the Inventory of Small Life Events,41 and the Daily Stress Inventory.42 However, the checklist approach has many limitations. Clearly, all self-report instruments are open to response biases, social desirability, and unconscious (repressive) coping.43 Even more problematic, none of these instruments was designed to apply to medical trainees, who experience not only predictable and general pressures, such as interpersonal stressors, economic problems, fatigue, and confidence deficits,44 but also stressors specific to medicine, such as 24-hour schedules and issues of life and death. The particular demands of medical training merit measurement by a tool sensitive and specific to this population. One direction may be to develop a medical education stress inventory.
Although self-report measures of stress are important, examination of physiologic measures of stress should supplement them to validate the effectiveness of the stress-management techniques. According to Cambell and Fiske,45 multi-trait, multi-method assessment is the most sophisticated and accurate research design. The following are common physiologic measures cited in the literature46,47: electroencephalagram (EEG), electrocardiogram (EKG), blood pressure (baseline and return to baseline), cortisol levels, measures of immune functioning, finger-pulse transist time (FPTT), ear-pulse transit time (EPTT) (see Cacioppo and Tassinary47 for a more complete description of these measures).
One of the more common criticisms of these forms of assessment is that they may not generalize outside of the laboratory. This important question deserves attention, and measures have been or are being developed (e.g., ambulatory blood pressure monitor) to address it. Other potential limitations include cost, confounding variables, and time. Despite possible concerns, physiologic measures of stress give another relevant piece of the picture by providing an objective comparison with trainees' subjective self-reports. This is especially important in light of past studies of repression in which physiologic arousal was objectively measured even though participants did not self-report arousal.43 Further, these physiologic measures are often markers of future physiologic pathology.
Finally, although the negative consequences of stress include decreased attention and concentration, poor decision-making ability, alcohol and drug abuse, depression and anxiety, relationship difficulties, and even suicide, few of these variables have been assessed as outcome measures of stress-management programs for medical trainees. Future research must include outcome measures that will determine whether the stress-management program is able to buffer against these potential negative consequences of stress.
CONCLUSIONS AND IMPLICATIONS FOR FUTURE RESEARCH
The purpose of this article is to review the literature on stress management and medical education. Of the 24 studies reviewed, the vast majority supported the effectiveness of interventions designed to reduce the stress of medical education and training. In almost all cases (where measured) the participants found the programs useful. Unfortunately, a lack of careful control in most studies, few validated outcome measures, and heterogeneous interventions make drawing firm conclusions beyond this premature.
Despite continued calls for research on stress-management programs in medical education, there have been few carefully conducted trials. In our search, we found over 600 articles discussing the importance of addressing the stress of medical training. Common themes described by authors of these articles suggest that the reigning paradigm in medical education emphasizes performance under stress, competition, and self-denial. Unfortunately, only 24 of these articles reported empirically assessed intervention programs, and only six of these15–17,22–24 used rigorous scientific method.
This discrepancy illustrates that there is much work to be done. Although some may feel that the obvious conclusion is to implement stress-management programs immediately without future research, there are still many unanswered questions. It is unclear which types of stress-management programs are most effective (e.g., meditation versus support group). While it is encouraging that researchers are exploring many approaches, it is difficult to make comparisons among diverse treatments. Interventions have varied in treatment modality (e.g., individual versus group), in format (e.g., structured versus unstructured), and in therapeutic techniques (e.g., hypnosis versus meditation). Further research is needed to disentangle research designs and explore which components of a complex array of interventions are most effective.
It is also unclear what duration and frequency are necessary to produce results. Precise comparison of interventions of differing durations and frequencies must be made to determine the most efficient and effective stress-management programs for medical trainees. Further, although a variety of interventions have proven effective, sensitivity to medical trainees in general, and individual differences in particular, is necessary. Future research must accurately determine which interventions work best for whom by assessing moderator variables.
Based on our review, the following considerations should be incorporated into future research: (1) rigorous design, including randomization and control (comparison) groups; (2) precise study of varying durations and frequencies of interventions (e.g., two-day intervention versus eight-week intervention); (3) measurement of moderator variables to determine which interventions work best for whom; (4) specificity of outcome measures; and (5) follow-up assessment, including effectiveness of future patient care.
This review has described the widespread interest in stress-management programs, the promising start made by those already implemented, and the great unexplored territory that must be charted if these interventions are to efficiently and effectively succeed in the twin goals of benefiting physicians and their patients and establishing a sound scientific base for future research.
1. Johnson N, Michels P, Thomas J. Screening tests identify the prevalence of alcohol use among freshman medical students and among students' family of origin. J South Carolina Med Assoc. 1990;86:13–4.
2. Gallegos K, Bettinardi-Angres K, Talbott G. The effect of physician impairment on the family. Maryland Med J. 1990;39:1001–7.
3. Pitts FN, Winokur G, Stewart MA. Psychiatric syndromes, anxiety symptoms and responses to stress in medical students. Am J Psychiatry. 1961;118:333–40.
4. Salt P, Nadelson C, Notman M. Depression and anxiety among medical students. Paper presented at APA Annual Meeting, Los Angeles, CA, 1984.
5. Richings JC, Khara GS, McDowell. Suicide in young doctors. Br J Psychiatry. 1986;149:475–8.
6. Vitaliano P, Maiuro R, Russo J, Mitchell E. Medical student distress: a longitudinal study. J Nerv Ment Dis. 1989;177:70–6.
7. Smith A. Stress and information processing. In: Johnston M, Wallace L, et al (eds). Stress and Medical Procedures. Oxford Medical Publications. Oxford, England: Oxford University Press, 1990:184.
8. Askenasy J, Lewin I. The impact of missile warfare on self-reported sleep quality. Sleep. 1996;19:47–51.
9. Lehner P, Seyed-Solorforough M, O'Connor M, Sak S, Mullin T. Cognitive biases and time stress in team decision making. IEEE Trans Systems, Man & Cybernetics. 1997;27:698–703.
10. Klein G. The effect of acute stressors on decision making. In: Driskell J, Salas E (eds). Stress and Human Performance. Mahwah, NJ: Lawrence Erlbaum, 1996:48–88.
11. Pastore FR, Gambert SR, Plutchik A, Plutchik R. Empathy training for medical students. Unpublished manuscript, New York Medical College, 1995.
12. Kelly A, Marks F, Westhoff C, Rosen M. The effect of the New York State restrictions on resident work hours. Obstet Gynecol. 1991;78:468–73.
13. Kahn NB, Addison RB. Support services for family practice residents. J Fam Prac. 1992;34:78–180.
14. Butterfield PS. The stress of residency: a review of the literature. Arch Intern Med. 1988;148:1428–35.
15. Whitehouse WG, Dinges DF, Orne EC, et al. Psychosocial and immune effects of self-hypnosis training for stress management through the first semester of medical school. Psychosom Med. 1996;58:249–63.
16. Kiecolt-Glaser J, Glaser R, Strain E, et al. Modulation of cellular immunity in medical students. J Behav Med. 1986;9:5–21.
17. Shapiro SL, Schwartz GE, Bonner G. Effects of mindfulness-based stress reduction on medical and premedical students. J Behav Med. 1998;21:581–99.
18. Soskis DA. Teaching meditation to medical students. J Religion and Health. 1978;17:136–43.
19. Kelly JA, Bradlyn AS, Dubbert PM, St. Lawrence JS. Stress management training in medical school. J Med Educ. 1982;57:91–9.
20. Dashef SS, Espey WM, Lazarus JA. Time-limited sensitivity groups for medical students. Am J Psychiatry. 1974;131:287–92.
21. Hilberman E, Konanc J, Perez-Reyes M, Hunter R, Scagnelli J, Sanders S. Support group for women in medical school: a first year program. J Med Educ. 1975;50:867–75.
22. Palan BM, Chandwani S. Coping with examination stress through hypnosis: an experimental study. Am J Clin Hypnosis. 1989;31:173–80.
23. Nathan RG, Nixon FE, Robinson LA, Bairnsfather L, Allen JH, Hack M. Effects of a stress management course on grades and health of first-year students. J Med Educ. 1987;62:514–7.
24. Holtzworth-Munroe A, Munroe MS, Smith RE. Effects of a stress-management training program on first and second year medical students. J Med Educ. 1985;60:418–9.
25. Klamen DL. The stress management workshop for medical students. Acad Psychiatry. 1997;21:42–7.
26. Webster TG, Robinowitz CB. Becoming a physician: long-term student group. Gen Hosp Psychiatry. 1979;53–61.
27. Reuben DB, Novack DH, Wachtel TJ, Wartman SA. A comprehensive support system for reducing house staff distress. Psychosomatics. 1984;25:815–20.
28. Blitch JW, Bowman DO, Adams CE, Jett CR, Campbell DC. Promoting the “whole physician's” well-being: the experience of a support group for family practice residents. Fam Prac Res J. 1983;2:231–8.
29. Golden JS, Rosen AC. A group dynamics course for medical students. Int J Group Psychother. 1975;25:305–14.
30. Johnson AH. Resident self-awareness through group process. J Fam Pract. 1977;4:681–4.
31. Kabat-Zinn J, Chapman-Waldrop A. Compliance with an outpatient stress reduction program. J Behav Med. 1988;11:333–52.
32. Strahilevitz A, Yunker R, Picanick AM, Smith L, Richardson J. Initiating support groups for pediatric house officers. Clin Pediatr. 1982;21:529–31.
33. Yalom I. Theory and Practice of Group Psychotherapy. New York: Basic Books, 1985.
34. Brock CD, Stock RD. A survey of Balint group activities in U.S. family practice residency programs. Fam Med. 1990;22:33–7.
35. Franco KS, Tamburrino MB, Carroll BT, Somani A, Wagner SM. Reducing the stress of a medical education: two approaches. Ohio Med. 1987;October:691–2.
36. Matthews D, Classen D, Willms J, Cotton J. A program to help interns cope with stresses in an internal medicine residency. J Med Educ. 1988;63:539–47.
37. Pelletier KR, Lutz R. Healthy people—healthy business: a critical review of stress management programs in the workplace. Am J Health Promotion. 1988;winter:5–12, 19.
38. Lewinsohn P, Amenson C. Some relations between pleasant and unpleasant mood-related events and depression. J Abnorm Psychol. 1978;87:644–54.
39. Kanner AD, Coyne JC, Schaefer C, Lazarus RS. Comparison of two modes of stress management: daily hassles and uplifts versus major life events. J Behav Med. 1981.
40. Stone A, Neale J. Development of a methodology for assessing daily experiences. In: Baum A, Singer J (eds). Advances in Environmental Psychology: Environment and Health. Hillsdale, NJ: Lawrence Erlbaum, 1982:49–83.
41. Zautra A, Guarnaccia C. Measuring small events. Am J Commun Psychol. 1986;14:629–55.
42. Brantley PJ, Waggoner CD, Jones GN, Rappaport NB. A daily stress inventory: development, reliability, and validity. J Behav Med. 1987;10:61–74.
43. Schwartz G. Psychobiology of repression and health: a systems approach. In: Je S (ed). Repression and Dissociation: Implications for Personality Theory, Psychopathology and Health. Chicago, IL: University of Chicago Press, 1990:337–87.
44. Rudner H. Stress and Coping Mechanisms in a Group of Family Practice Residents. J Med Educ. 1985;60:565–6.
45. Campbell EG, Louis KS, Blumenthal D. Looking a gift horse in the mouth: corporate gifts supporting life sciences research. JAMA. 1998;279:995–9.
46. Tyson P. Task-related stress and EEG alpha biofeedback. Biofeedback & Self-Regulation. 1987;12:105–19.
47. Liang S, Jemerin JM, Tschann JM, Wara DW, Boyce W. Life events, frontal electroencephalogram laterality, and functional immune status after acute psychological stressors in adolescents. Psychosomat Med. 1997;59:178–86.
48. Weiner PS. A social-discussion group for first-year residents. J Med Educ. 1984;59:137–9.
49. Ziegler JL, Kanas N, Strull WM, Bennet N. A stress discussion group for medical interns. J Med Educ. 1984;59:205–7.
50. Siegel B, Donnelloy JC. Enriching personal and professional development: the experience of a support group for interns. J Med Educ. 1978;53:908–14.
51. Cadden JJ, Flach FF, Blakeslee S, Charlton R Jr. Growth in medical students through group process. Am J Psychiatry. 1969;126:862–8.