We found that physicians’ choices for managing depression, especially for the patients with major depressive disorder (scenario A), were consistent with AHCPR guidelines. SSRIs were prescribed most often, usually accompanied by a recommendation for ongoing counseling with a psychologist or social worker. A definitive approach for effectively treating subthreshold depression was not established, but we found counseling, over-the-counter medications, SSRIs, and exercise were most often recommended. If depression is diagnosed, follow-up should occur sooner than we observed in our study visits.
We found that visit duration and charges were associated with severity of symptoms. Encounters for scenario A were approximately 16% longer and 7% more costly than for scenario B. First visits across all encounters were also approximately 45% longer and 25% more costly than second visits. Time is becoming an ever more important factor in managed care environments. We learned that those who pursued fewer than 2 depressive symptoms were much less likely to diagnose depression than those who pursued 2 or more symptoms. Time pressures can certainly influence physicians’ decisions about pursuit of symptoms.
The strengths of this study include the kind of data possible to collect. Our methods allowed for a level of detail that cannot be obtained using medical record review or surveys on the basis of patient recall or physician self-report. Our results on agreement between the standardized patients and the project staff indicate that the quality of the data at this level of detail is excellent. Second, since physicians would normally adjust their approaches or management on the basis of the characteristics of the patient, our study allowed us to obtain information on what many doctors do when faced with the same patient during a subsequent visit. Obtaining this kind of information is virtually impossible using actual patients. Another strength of our study is that we included 3 regions of the country, which assists in making our findings generalizable. Many studies using unannounced standardized patients suffer from single and first-visit bias. We were able to orchestrate 2 visits for all but 10 study participants, allowing the potential for a relationship to become established between the patient and the provider. This rapport could be a critical factor in the ability to identify and discuss depression, which continues to be associated with stigma that may affect patients’ receptivity.
Limitations
We acknowledge some important limitations to our study. Our sampling techniques were purposeful, since we wanted a blend of family physicians and internists with varying ages and sex, affecting generalizability. The physicians who agreed to our investigative methods may have been more courageous in exposing their practice styles to such scrutiny and may be more skilled or confident than nonvolunteers. Our evaluation method is performance-based rather than patient-based; thus, we cannot provide information that can be linked to actual patient outcomes.
Though we felt at least 2 visits were necessary to assess physicians’ approaches to and management of depression, 2 visits may not be enough to capture representative behaviors of primary care physicians who often have very well established relationships with patients that affect diagnosis and management strategies. Our design allowed us to evaluate how physicians approach a major depressive disorder with an obvious chief complaint versus minor depression with a less obvious presentation. However, to adequately evaluate how AHCPR classifications influence recognition and management of depression, a more discriminating factorial design and larger sample size would be required.
Conclusions
We found that the rate of recognition of major depression in primary care is very high, and many aspects of AHCPR guidelines are followed in primary care. Patient presentation influences recognition, diagnostic exploration, and management approaches. When major depression is recognized, more routine exploration of suicide ideation and more universal follow-up within 2 weeks are areas for improvement. For somatic presentations compatible with depression, more routine inquiry about mood and pleasurable activities may increase recognition.
Acknowledgment
This study was supported by the John D. and Catherine T. MacArthur Foundation.