The medical care needs of patients with psychiatric disorders have received increasing attention over the past decade (¹—⁵). However, many psychiatric residency training programs continue to offer little training in primary medical care beyond the Accreditation Council for Graduate Medical Education (ACGME) required minimum of 4 months. This exposure often takes place during internship on inpatient medical units for severely ill patients. Residents in psychiatry may sense little connection between this type of medical training and the medical skills they need to counsel patients or to integrate medical care into their own practice.

During this time, the U.S. Department of Veterans Affairs (VA) reorganized its health care priorities from an inpatient-based model to an emphasis on outpatient primary care. In 1996 VA leadership in mental health began planning the Psychiatry Primary Care Education (PsyPCE) initiative to encourage psychiatric resident education at the interface of psychiatry and primary care. Veterans Affairs leadership did not determine the content of the education program, encouraging a variety of programs to be developed. The PsyPCE initiative therefore provided a national laboratory to explore models for enhancing primary care education in psychiatric residency training (⁶, ⁷).

In this article we review these programs to determine common elements that could be considered for a model program. We also report on resident knowledge, satisfaction, and practice behavior following implementation of the PsyPCE intervention in one representative program.

Six months after the VA selected 19 PsyPCE training sites (among 133 eligible facilities), we conducted a survey to determine the structure and educational interventions implemented at each site, using a structured interview of 21 items and an open-ended question about barriers to implementation (Table 1).

To investigate the experience of residents and outcomes, we focused on a representative training program and a relatively large residency. Three years after initiation of the PsyPCE program at the site, two settings in this university-based residency program completed a Residency Experience Questionnaire. Both settings, a VA mental hygiene outpatient clinic and a community mental health center, care for patients with severe mental illness. The settings differ in that the VA patients are more likely to be Caucasian, male, and older, with a broader range of socioeconomic backgrounds. In both settings, third-year residents spend 12 consecutive months treating outpatients using pharmacological and psychotherapeutic techniques. Residents do have input into whether they are assigned to the VA or community mental health center setting; however, no resident listed the PsyPCE program as a reason for choosing the VA.

Resident educational programs are similar at both sites, with the exception that residents at the VA have an enhanced education and clinical immersion in primary care psychiatry. The PsyPCE program at the VA included lectures on health promotion and disease prevention techniques, health screening, and management of common medical disorders in patients with severe psychiatric illness. Residents at the VA also had contact with a primary care medical team embedded in the psychiatry service and were encouraged to elect to provide primary care to a panel of patients. The Psychiatry Primary Care Clinic was designed to encourage communication between the primary care team (an internist, two nurse practitioners, and a registered nurse) and the outpatient clinicians (⁸). The resident clinic is on the same floor as the Psychiatry Primary Care Clinic, allowing enhanced personal communication between residents and medical providers. In addition, psychiatric and primary care clinicians write notes in the same electronic record and communicate by e-mail about patient concerns. Psychiatry Primary Care Clinic staff attend monthly psychiatry team meetings. The residents at the community mental health center did not have the lecture series, easy access to and liaison with a primary care team, or the ability to elect to provide primary care services.

The Resident Experience Questionnaire consisted of 15 questions assessing resident attitudes toward provision of primary medical care, knowledge about the extent of medical disorders in the seriously mentally ill patient population, satisfaction with provision of primary care services at their site, and questions assessing whether residents integrated primary care precepts into their own clinical practice (Table 2). Questionnaires were distributed to residents who had completed their third postgraduate year (PGY-3) in either setting during the past 2 years. Residents were asked to complete the questionnaires anonymously as part of a survey of primary care education in the department. The first 12 questions utilized Likert-scale responses, with ratings from 1 (strongly disagree) to 5 (strongly agree). Questions 13 to 15 asked for the frequency of specific primary care practice behaviors that residents had completed with their own patients, such as blood pressure checks, physical examinations, and developing treatment plans that integrated medical problems.

To ensure that the site chosen to analyze resident experience in the program was representative of the majority of sites, we used hierarchical cluster analysis (SPSS, 1998), including all 19 programs but only 12 of the 21 variables (Table 1). Variables were excluded if they were endorsed by more than 70% of the programs, because high levels of endorsement provide minimal variation for statistical cluster analyses. Using the dendrogram from these analyses, we determined the optimum number of program or variable clusters using scree analysis. Using these cluster scores, we computed a mean score for each of the 19 programs, which can range from a minimum of 0 to a maximum of 1.

To evaluate the Resident Experience Questionnaire, we compared PGY-3 resident responses at the VA and community mental health center sites. Because of the ordinal quality of our data, we utilized nonparametric statistics (Wilcoxon log rank test).

Of the 19 PsyPCE sites, 100% participated in the survey. Sites were relatively evenly distributed throughout the United States, representing 14 states. All had well-established psychiatry and primary care medicine residency programs. The number of residents influenced by the initiative varied based on the overall residency size, which was up to 84 residents in one program, as well as decisions about which residency years were involved in the program. Because of this variability, we decided to use the 19 programs themselves as the unit of analysis rather than the residents trained by the initiative.

The survey results are presented in Table 1. Nine variables were present in 70% or more of the primary care education programs. Two variables (weekly didactics and use of VA screening and treatment protocols) were included in each education program. The seven other common variables were treatment team concept (90%), internists or nurse practitioners as team members (90%), registered nurses as team members (80%), social workers as team members (80%), residents performing histories and physical exams (85%), incorporation of a disease prevention plan (95%), and involving PGY-3 residents in the training program (85%).

Twelve other items were a component of the primary care education program in less than 70% of the sites: mental health clinic as training setting (68%), ambulatory medicine clinic as training setting (42%), psychologist team members (42%), use of case conferences (37%), evaluations of patient satisfaction (68%) and resident satisfaction (47%), utilization review data feedback (42%), resident competency evaluations (47%), and which residency years, excluding PGY-3, were targeted for the training: PGY-1 (37%), PGY-2 (47%), and PGY-4 (42%).

To develop a typology that uniquely categorized the types of programs, we used cluster analysis for the 19 programs and the 12 items with less than 70% concordance. Using skree analysis, the 12 variables separated into three clusters, including one variable (resident testing), three variables (mental health clinic setting and patient and resident satisfaction), and the remaining seven variables. Using these three cluster scores, we computed a mean score for each of the 19 programs, ranging from 0.10 to 0.90, thereby providing excellent variance in this categorization.

Several barriers to implementation were reported in response to an open-ended question. The most frequently cited barrier was adequate clinic space, followed by obtaining sufficient medical supervision. Gaining acceptance of the PsyPCE program from psychiatry, nursing, and medical faculty was also problematic at several sites.

To examine resident experience and outcomes, we chose the program that had a median cluster score of 0.48, suggesting it was "representative" of the national programs, based on the above typology. Response rates to the questionnaire were 60% for residents in the VA PsyPCE program and 79% for residents at the community mental health center.

The results of the questionnaire are summarized in Table 2. Residents in PGY-3 receiving the VA PsyPCE intervention and those in the traditional program at the community mental health center believed that primary care education was important and that psychiatrists have the responsibility to ensure that their patients have quality medical care. Both groups also believed that providing primary care for patients was an important component in their degree of satisfaction with outpatient training, and both groups were relatively satisfied with their overall outpatient training. Compared with residents at the community mental health center, residents receiving the VA PsyPCE intervention were significantly more likely to believe that patients with severe mental illness are at high risk of having comorbid medical illness and that patients they treated had access to high-quality primary medical care and knew the practitioners providing that care. They reported meeting more frequently with primary care providers and being more likely to develop collaborative treatment plans integrating mental health and primary care concerns. However, the practice behavior of residents in the PsyPCE program, as indicated by the number of integrated treatment plans written or by the number of times residents examined their patients, did not significantly differ between the two groups. Few residents in either group wrote treatment plans that integrated medical issues, and few completed elements of a physical examination on their patients. No PGY-3 resident elected to provide primary care to a panel of patients, although several spent time in the clinic observing the provision of primary care.

The VA enhances collaboration with medical schools through sponsorship of residency training initiatives like the PsyPCE program. Although the VA patient population and training environment may at times limit generalization of results, these initiatives have the potential to significantly impact residency training by developing innovative programs that go on to become national models.

This study describes program elements of 19 VA-sponsored PsyPCE program interventions across the United States. Several elements were common to many of these programs and differ from the usual medical training in psychiatric residency. Traditionally, psychiatrists have learned medical skills during PGY-1 in settings divorced from mental health clinics. Psychiatric educators in this initiative strongly suggested that primary care education should be incorporated in the PGY-3 outpatient year in settings like outpatient mental health or substance abuse clinics. The majority of these initiatives conceptualized incorporating primary care providers as part of the treatment team rather than a separate component of care; the VA’s electronic medical record and easy e-mail communication enhanced this conceptualization. Lastly, primary care precepts like health promotion and disease prevention were incorporated as standard elements of the curriculum. Because these four elements were so commonly included in the PsyPCE programs, we believe they should be considered elements of any model curriculum in primary care education for psychiatric residents.

To better understand resident experience and outcomes, we compared PGY-3 residents in a representative PsyPCE program with residents at a community mental health center. The generalization of these results could be limited by our modest sample size and by the possibility that residents interested in primary care may have selected to go to the VA. We found that both groups overwhelmingly felt education about primary care was an important component of outpatient psychiatric education and that psychiatrists were responsible for ensuring that patients with psychiatric disorders receive primary care. These results suggest trainees have abandoned a previous model of care in which psychiatrists did not "own" the medical problems of their patients. Residents in the PsyPCE program were more knowledgeable about the prevalence of medical disorders in patients with psychiatric illness and more likely to collaborate closely with primary care providers than those trained in the community mental health setting.

In addition to educating residents about primary care precepts, the majority of the PsyPCE programs taught residents to provide primary medical care to their patients by educating patients about health promotion activities, integrating medical problems into their problem lists, and conducting physical exams. We did not assess resident confidence in educating patients about their medical concerns or encouraging health promotion activities; future research in primary care education should include this outcome. When compared with residents at the community mental health center, residents in the representative PsyPCE program did not report a higher frequency of incorporating primary care practice behaviors, such as writing integrated treatment plans or conducting elements of a physical examination into their own practice with patients. Although this was based on self-report, these results are similar to an evaluation of clinical practice after graduation from a PsyPCE program (⁹) and suggest that despite having improved knowledge, barriers to psychiatrists providing clinical elements of primary care are established early in training.

Why didn’t residents in the PsyPCE programs incorporate primary care precepts into their practice? The issue of role modeling may be critical in understanding the reluctance. Although residents were provided didactic instruction and had backup in the Psychiatry Primary Care Clinic, none of their supervisors in the clinic were providing primary care services, and some had expressed concern about being responsible for supervising provision of these services. Paradoxically, the ease of access to primary care providers who were embedded in the psychiatry outpatient clinic may have inhibited residents from actually providing primary care services themselves. Until further research delineates the etiology of these barriers, educators should understand that efforts to encourage direct provision of primary care services may not influence resident practice behavior.

Manuscripts authored by an editor of Academic Psychiatry or a member of its editorial or advisory board undergo the same editorial review process, including blinded peer review, applied to all manuscripts. Additionally, the editor is recused from any editorial decision making.

At the time of submission, the authors disclosed no competing interests.