Depression is a common and disabling condition. In the primary-care setting, only about half of depressed patients receive treatment, and treatment often falls short of practice guidelines (1). Physicians who are confident in their ability to treat depression are more likely to recognize depressed patients (2). Also, physicians’ attitudes affect willingness to inquire about symptoms and ability to diagnose and treat depression (2).

During medical training, physicians acquire attitudes and approaches to patient care. However, educational efforts aimed at improving recognition and treatment of depression yield disappointing results (3, 4). Underrecognition may be encouraged by a “hidden curriculum” that reinforces biomedical over biopsychosocial models of care (5). The “hidden curriculum” includes “behavior exemplified by members of the faculty” (5). Thus, one must examine the attitudes and behaviors of supervisors, or student perceptions of the same, to better understand the “hidden curriculum.”

A review of the literature failed to identify any research addressing this topic. Therefore, we surveyed internal-medicine residents to assess residents’ perceptions of supervision surrounding the care of depressed patients and examine the relationship between residents’ attitudes and behaviors and their perceptions of supervisors’ attitudes.

Study Population

Internal-medicine residents at a southern regional medical school participated in this study with approval of the institutional IRB. Surveys were disseminated at residents’ continuity clinics during preclinical conferences. All residents in attendance on the day of survey administration were eligible to participate.

Survey Instrument

Participants completed a questionnaire measuring demographic data, attitudes about depression, retrospective estimates of provider behavior and perceptions of supervisors’ attitudes and behaviors. The survey included items from a previously validated questionnaire, the Depression Attitude Questionnaire (DAQ) (6), and additional items developed by the researchers. Possible questionnaire order effects were controlled for by counter-balancing.

The DAQ is a 20-item visual-analog questionnaire designed to assess General Practitioners’ (GPs) attitudes toward depressive illness; it contains three components: 1) Negative Attitude to Depression and Its Treatment; 2) Professional Ease; and 3) Tendency to Defer to Specialists (6, 7). The Supervising Attending Depression Attitude Questionnaire (SADAQ) is a 10-item visual-analog questionnaire designed for use in this study to assess residents’ perceptions of supervisors’ attitudes about depression and training adequacy. Responses were grouped by content into three subscales 1) Attendings' Attitude; 2) Supervisors’ Impact; and 3) Training Quality. All visual-analog scales ranged from Strongly Disagree (0 mm) to Strongly Agree (100 mm). Researchers also developed a 4-item Provider Behavior Questionnaire (PBQ), assessing residents’ estimates of the prevalence of depression in their patient population and their screening, prescribing, and referring behaviors. A copy of all newly-developed questionnaires is available upon request.

Statistical Analysis

Descriptive analyses were conducted for demographic, attitudinal, and behavioral variables. DAQ component scores and SADAQ subscale scores were submitted to a Pearson's product-moment correlation analysis. SADAQ subscale scores and PBQ item scores were subjected to a Spearman rank-correlation analysis.

Participant Characteristics

Of the 139 residents in the internal medicine program, 54 (39%) attended preclinical conferences on the days of survey administration. Of the residents in attendance, 94% participated; 18 of the residents were female, 30 were male, and 3 did not indicate gender; 73% attended medical school in the United States. There were 9 PGY-1s, 16 PGY-2s, 17 PGY-3s, 8 PGY-4s, and 1 PGY-6.

Depression Attitudes

The overall mean for the Negative Attitude Toward Depression and Its Treatment component was 25.8 mm (SD: 11.6 mm), which indicated disagreement with negative, stereotypical views of depression and an optimistic view of its treatment and prognosis. The Professional Ease component mean was 55.5 mm (SD: 14.1 mm), signifying slight agreement with statements indicating professional ease in treating depressed patients. However, 49% of residents reported that they do not feel comfortable dealing with depressed patients’ needs. The overall mean for the Tendency to Defer to Specialists component was 62.1 mm (SD: 16.5 mm), which indicated agreement that psychotherapy is a beneficial treatment for depression and that depression treatment is best left to specialists.

Self-Reported Behaviors

Most residents (56%) reported screening ≤20% of patients they saw in the past 3 months for depression. Most (52%) reported prescribing antidepressants to ≤30% of depressed patients, and 53% referred ≤30% for additional mental health services.

Supervisors’ Attitudes

The internal consistency of the SADAQ was computed with Cronbach’s alpha (α). For the overall SADAQ, α=0.72. For the Attending Attitude, Supervisor Impact, and Training Quality subscales, α=0.79, 0.48, and 0.65, respectively.

About half (49%) of the residents stated that most attending physicians think every patient should be screened for depression. Thirty residents (59%) had observed a supervisor screen for depression. The majority of residents (71%) indicated that they are more likely to screen if it is a priority for their attending physician. About half (53%) of the respondents thought supervisor attitudes have affected their attitudes about treating depression. Twenty-four residents (48%) indicated that they received adequate training in managing depression. More positive perceptions of supervisor attitudes were correlated with improved perceived training quality (r=0.68; p <0.01).

Relationship Between Residents’ and Supervisors’ Attitudes

No significant correlations were observed between residents’ perceptions of supervisors’ attitudes and residents’ attitudes about managing depression (including all three DAQ components). There was a positive correlation between residents’ perception of training adequacy and their professional ease in treating depression (r=0.36; p=0.01). Also, residents who perceived training to be adequate were less likely to defer the care of depressed patients to specialists (r = −0.33; p=0.02).

Relationship Between Supervisors’ Attitudes and Residents’ Self-Reported Behavior

Residents who perceived supervisor attitudes to be positive reported screening patients for depression more frequently (ρ=0.41; p <0.01). Residents who felt their training to be adequate reported more regular screening (ρ=0.48; p <0.01). No other significant correlations between supervisors’ or training factors and residents’ behaviors were observed.

Postgraduate Training Year

Senior residents expressed less negative attitudes about depression and its treatment (r = −0.35; p=0.02). However, no significant correlations were observed between postgraduate training year and professional ease in managing depression (r=0.01; NS) or tendency to defer treatment to specialists (r=0.20; NS). Also, there were no significant correlations between training year and tendency to screen patients (ρ=0.12; NS), prescribe antidepressants (ρ=0.26; NS), or refer for mental health services (ρ=0.20; NS). Furthermore, postgraduate training year was not correlated with confidence in managing depression (r=0.04; NS).

Only about half of surveyed residents felt comfortable treating depressed patients. This is significant when one considers that GPs who feel comfortable managing depression are more likely to make accurate diagnoses (2). Fewer than half (48%) of the residents indicated that they received adequate training in managing depression. Similarly, 46% of surveyed primary-care training directors were satisfied with their institution’s psychiatric training (8). However, 74% of internal-medicine training directors surveyed considered depression-related teaching in their programs to be “adequate” or “thorough” (9). Postgraduate training year was not associated with improved confidence in managing depression or tendency to screen, prescribe antidepressants, or refer for mental health services. The finding that years in residency training alone are not sufficient to improve depression management underscores the importance of identifying aspects of training that promote residents’ ability to treat depression.

To our knowledge, this pilot study represents the first attempt to assess the impact of supervision on trainees’ management of depression. Over half of surveyed residents (58%) indicated that supervisors’ attitudes affect their own attitudes about managing depression, although the correlation was not statistically significant. Seventy-one percent of residents reported they are more likely to screen for depression if it is a priority for their attending physicians. Unfortunately, only half believed it was important to supervisors to screen all patients, and only 59% had ever observed an attending screen for depression. Perhaps these findings explain why most respondents reported screening fewer than 20% of patients.

The present study had some important limitations. First, study findings may have limited generalizability because they were based on a small sample from a single residency program, and only 39% of the residents were in attendance on the day of survey administration. However, as residents did not have foreknowledge of the data collection, it is unlikely that attitudes about depression or supervision affected attendance. Also, measures assessing perceptions of supervisor attitudes were developed for use in this pilot study because no validated measure existed. The internal consistency of the Supervisor Impact subscale of the SADAQ was below what is generally considered acceptable, which highlights the difficulty of assessing this construct. Future research in this area should work toward development of a more psychometrically sound measurement tool. Also, a social desirability response bias may have influenced resident responses in a positive direction. Last, the cross-sectional, self-report nature of the data limited our ability to draw causal conclusions regarding the impact of supervisor attitudes on resident trainees. Cohort studies involving objectively-measured data are warranted in this area.

Despite these limitations, our findings underscore the importance of adequate role-modeling to promote resident preparedness to manage depression in primary care. Preceptors must foster a sense of responsibility by encouraging trainees to inquire about depressive symptoms as a routine part of their examination. Both the formal and “hidden” mental health curriculum must be examined and altered to improve GPs’ ability to effectively treat depression. Because GPs care for most depressed patients seeking treatment, enhancing providers’ ability to manage depression will significantly improve public health.