The World Health Organization has defined primary care as “essential health care based on practical, scientifically sound, and socially acceptable methods and technology made universally accessible to individuals and their families”
(1). Although primary care physicians are only one group of health professionals who can deliver primary care, in North America they are often the first contact for patients with mental health problems
(2). In the United States, the majority of mentally ill patients who engage in treatment do so with a primary care physician
(3,
4). Once patients are seen by a primary care physician, referral to psychiatrists or other mental health professionals is low
(2,
3,
5). Thus, for reasons that include greater patient comfort with a primary care physician, the desire to avoid being labeled as mentally ill, and long waits for psychiatric consultation, primary care physicians play a pivotal role in the diagnosis and management of patients with mental illness
(2,
3,
5,
6).
However, it is frequently shown that primary care physicians do not always provide effective diagnosis, documentation, or treatment of mental illnesses
(6–
12). This is particularly problematic in rural and remote settings in which the closest psychiatrist may be hours away
(13–
15). In such settings it has been argued that providing continuing education for primary care physicians is more likely to improve the quality of mental health care than the recruitment of more psychiatrists
(5).
There is a long history of interest among psychiatrists in providing education to primary care physicians. The first efforts began during World War II as a means of meeting the shortage of psychiatric care for U.S. military personnel
(16). One of the earliest and most widely quoted writers was Balint
(17), who pioneered a system of continuing education for primary care physicians in 1950. He involved groups of 10 primary care physicians in weekly sessions that focused on difficulties in managing mentally ill patients with medical illness. “Balint groups” that emphasized the doctor-patient relationship and personal reactions to difficult patients met for months or even years
(17). Later, in the 1960s and 1970s, continuing education in psychiatry was propelled by the efforts of the American Psychiatric Association and funding from the National Institute of Mental Health
(16). By the end of the 20th century, continuing education for primary care physicians had become an important part of the mission of most medical schools, health care institutions, and national mental health organizations.
However, although desire among primary care physicians for psychiatric education has coincided with interest among psychiatrists to provide it, how to best do so has not been well articulated. Compared with undergraduate or postgraduate education, there is relatively little consensus regarding methods and few empirical studies of effectiveness. Recently, the number of publications concerning continuing education in psychiatry for primary care providers has risen and includes a few focused reviews; however, there has been no comprehensive review. The purpose of this article is to review all of the existing English language literature addressing psychiatric education for primary care physicians published in the last 50 years.
Four focused reviews published during the last decade have addressed mental health training for nonpsychiatrists. Two of these
(18,
19) were almost entirely limited to the education of primary care residents and generally described interventions that would not be feasible for primary care physicians, and the third
(20) reviewed educational strategies as a minor component of an epidemiological review of psychiatric disorders in primary care. The fourth article
(21) reviewed all existing experimental studies, but because of the small number of controlled trials reported, relatively few conclusions could be drawn. We felt that a focus on empirical studies alone eliminated multiple and varied educational studies described in the much wider literature published over 50 years.
For this review we cast the net widely and retraced the literature back to the 1950s, when Balint described the first education of primary care physicians. After reviewing more than 400 articles, we identified important educational themes, methods, and research strategies that we hope will improve educational and research design in this crucial and growing field.
Method
We searched the MEDLINE and PsycLIT databases from 1950 to 2000 using a wide range of subject headings, including physician/family doctor, psychiatry/mental health, and training/education. Relevant secondary references listed in the bibliographies of relevant articles were also retrieved. More than 400 English language abstracts and articles were reviewed. Articles on the training of psychiatrists were excluded.
Results
The most important or representative articles are cited in this review, which is divided into three categories: 1) needs and objectives, 2) methods, and 3) effectiveness.
Needs and Objectives
Many articles tried to define the knowledge, skills, or attitudes on which to focus continuing education for primary care physicians. Although learner needs can be defined either subjectively (as what learners perceive they need to know) or objectively (as what objective evidence indicates they need to know), most authors emphasized the latter. There were two approaches to selecting course content objectively: deficit-based objectives and epidemiological-based objectives. Deficit-based objectives were derived from information gathered about the actual knowledge, skills, or attitudes of primary care physicians. Educators then planned their programs to remedy the presumed deficits. Information was derived from many different sources; a test was a common source. For example, Cohen-Cole et al.
(22) asked 26 subjects to watch two simulated clinical encounters and respond to a series of standardized mental health questions. The responses were examined by experienced clinicians who used them to develop a 30-week curriculum
(23,
24). Other authors suggested that curricula should be based on observed deficits in specific domains, such as clinical evaluation of the suicidal patient
(25), interviewing skills
(26), or recognition and treatment of depression
(27,
28).
Epidemiological-based objectives were derived from epidemiological data. In fact, most articles on primary care psychiatry started by citing literature to illustrate gaps in mental health care delivery. Stoudemire
(29) published a set of epidemiological-based objectives after a comprehensive review of the literature regarding the incidence of mental health problems in primary care. He suggested a core curriculum for primary care physicians based on the skills necessary to diagnose and treat these specific conditions.
It is of interest that whether they used deficit-based or epidemiological-based methods, almost all the authors who published objectives for psychiatric education of primary care physicians were psychiatrists. By contrast, there was strong evidence that the topics most often selected by primary care physicians for continuing education were quite different from those selected by psychiatrists. Primary care physicians most often wanted to increase their knowledge regarding somatization, psychosexual problems, difficult patients, and stress management
(30–
33), whereas psychiatrists emphasized the diagnostic criteria of disorders such as schizophrenia, bipolar disorder, and depression. In the area of skills development, primary care physicians placed particular emphasis on crisis, family, individual, and marriage counseling and strategies to prevent their own burnout
(30). They were less likely to identify learning needs related to new pharmacological agents, a topic that figured prominently in many psychiatrists’ educational materials.
These differences have important educational implications. It is much easier for a psychiatrist educator to give a lecture on diagnosis or medication than it is to teach management of difficult or chronically ill patients or psychotherapy. Education that is focused on diagnosis and medication may neglect the very cornerstone of psychiatric primary care, which is learning to develop and maintain effective relationships with patients who have complex problems
(2).
Although it seems reasonable that psychiatrists should not be prescriptive in creating curricula for primary care physicians, there are few examples reported of collaboration between psychiatrists and family physicians in defining curricula. To be effective, objectives for the psychiatric education of primary care physicians must be derived from a needs assessment of learners
(1,
34), and the curriculum should be oriented toward the perspective of primary care physicians and the problems of specific patients. An obvious but generally ignored implication is that primary care is not conducive to a 50-minute psychiatric assessment. This makes primary care physician education difficult for psychiatrists who must try to adapt methods they use in 30–50-minute sessions to the context of family practice, which rarely allows for more than a 15-minute interaction
(31–
35).
No new knowledge or skills will be learned by primary care physicians without an open and engaged attitude toward the treatment of patients with mental illness
(36). Despite the obvious importance of attitude, few educators have addressed attitudinal shifts that, nevertheless, take time and require an ongoing relationship between teacher and learner
(36–
38). In one study
(38), a case-based course held weekly for 6 weeks had no impact on the attitudes of its 24 physician participants. Similarly, among physicians who learned psychiatry during consecutive 12-week seminar blocks, only those who participated for 3 or more years demonstrated changed attitudes and a broadened scope of practice in the management of patients with mental illness
(16).
There can be a gap between the perceptions of primary care physicians and the perceptions of psychiatrists that affects the teaching of attitudes. Some psychiatrist authors
(16,
36,
38) described a need for primary care physicians to value patients’ psychosocial difficulties to a greater degree. By contrast, primary care research showed that primary care physicians already felt strongly that they should care for the emotional needs of their patients
(39). It has been argued that the nature of psychiatry itself is responsible for this schism
(1). From the point of view of primary care physicians, psychiatry in past decades has employed a wide range of mysterious and even dangerous treatments without definitive proof of their effectiveness. The resulting perception of psychiatry as vague and imprecise has only recently begun to change as psychiatrists embrace an evidence-based approach
(1). It has been suggested that although psychiatrists tend to see primary care physicians as unskilled and “in over their heads,” primary care physicians may see psychiatrists as inaccessible, nonmedical, and uncommunicative after assessment
(40). A failure to address these important attitudinal topics during the education of primary care physicians is an omission that may impede a successful outcome.
Finally, some important systems issues that affect primary care physicians must be considered in designing continuing education. Most objectives for psychiatric education are predicated on the expectation that primary care physicians will provide more mental health care than they currently do, managing more psychiatric patients and using more sophisticated treatments
(33). At the same time, primary care physicians are often too busy to give intensive psychiatric care
(40). Primary care authors have argued that the implicit message to do more may be hard to hear from psychiatrists, who may be less receptive to their own need for more education in general medicine than are primary care physicians for education in psychiatry
(41). Furthermore, because of the historical separation of mental health treatment institutions from general medical facilities, psychiatrists may not have a good appreciation of the vicissitudes of general medical practice
(42,
43). Primary care education programs run the risk of failure unless the psychiatrists involved make a major commitment to learn about the needs of primary care physicians and understand the day-to-day aspects of their job
(43).
Gallagher and Chapman
(43) suggested that teachers of primary care psychiatry should develop objectives for their own learning that include a commitment to learn about the primary care context firsthand. They suggested exploring the actual efficiency and relevance of the mental health services provided to the primary care physicians taking the course. These authors also suggested provocatively that psychiatrists are unlikely to appreciate the context of primary care until postgraduate psychiatry training programs provide rotations in primary care settings
(43).
In emphasizing the perspective of primary care physicians, we do not want to suggest that psychiatrist educators cannot introduce topics they feel are underdeveloped. In some cases, primary care physicians may have unperceived needs that lead to suboptimal diagnosis and treatment of mental health problems. On the other hand, the imposition of curricula by psychiatrists without collaboration with primary care physician learners themselves is education that is neither contextually relevant nor likely to result in a significant change in knowledge, skills, or attitudes. Rather, educators should aim for mutual collaboration and ongoing dialogue between primary care physicians and psychiatrists in the design of all psychiatric education programs
(2).
Educational Methods
The heterogeneity of the educational methods reported used for the continuing education of primary care physicians makes it difficult to isolate the interventions that are the most effective
(21). However, important variables do emerge and are confirmed by educational research in other disciplines of medicine. Three important variables are the duration of the intervention, the degree of active participation of the learners, and the degree of integration of new learning into the learners’ clinical context. We shall look at each of these in turn.
Duration
The length of continuing education events ranges from a few hours to several years; however, the most common format remains the short lecture or conference held away from the doctor’s work location
(43–
45). Such an approach has the advantage of providing relaxation but has been strongly criticized for the overuse of a didactic format in a location removed from the primary care milieu. As discussed, programs of longer duration are necessary to effect meaningful change in attitudes and growth in skills
(43). Short courses and lectures may serve to introduce topics that learners can pursue in the future, but maximum value is obtained only if such sessions are followed by the more permanent teacher-learner relationships that provide the opportunity to solve clinical problems
(43). It is apparent from research in other medical disciplines that significant changes in clinical practice and patient outcome are observed only as the result of ongoing, longitudinal educational programs
(44).
Active participation
One of the most important variables in effective continuing education is the active participation of the learners. It has been shown for more than three decades that a high degree of involvement by the physicians learning psychiatry is necessary for change
(43). One-day conferences that feature long lectures with little opportunity to interact or practice reenforcing strategies have little or no impact on practice patterns or patient outcome
(45). On the other hand, interactive, longitudinal programs that provide opportunities to practice knowledge and skills can influence clinical practice
(44). Such courses require the use of simulation, role playing, standardized patients (actors trained to portray patients with psychiatric problems), case-based learning, and videotaping with feedback
(36,
46,
47). A broad spectrum of activities that allows for individual learning styles, career stages, and motivations as well as frequent opportunities to apply new knowledge and skills in actual practice are important
(2,
46,
48).
Evaluation is also essential, not only because it reinforces learning, but also because it acts as a source of data for assessing program effectiveness. Multiple evaluation techniques are needed to capture the spectrum of knowledge, skills, and attitudes learned and must vary according to the teaching method used
(44–
51). Domains of educational evaluation include the following:
1. General satisfaction feedback gathered through questionnaires or interviews.
2. Changes in knowledge measured by self-evaluation, instructors’ opinions, or objective tests.
3. Changes in skills measured by videotaped interactions, unannounced standardized patients, and objective structured clinical examinations.
4. Changes in attitudes measured by questionnaires or interviews.
5. Changes in patient outcome measured by frequency and length of outpatient visits and hospitalizations, medications prescribed, laboratory tests ordered, patient satisfaction, and compliance and clinical improvement.
6. Changes in knowledge, skills, or attitudes measured after the end of the educational program.
Clinical (contextual) relevance
Continuing education programs can occur in locations that are totally removed the primary care clinical setting or immediately on site, in a primary care office. In between, various settings integrate education with a real clinical context to a greater or lesser degree. The role of psychiatrist educators varies as well, from pure consulting, to providing education related to specific cases, to clinical teaching in the classroom, to clinical supervision in longitudinal preceptorships
(52,
53).
The most effective models are those in which the learner is as close as possible to the actual site of practice and the psychiatrist teacher is seen functioning in a real clinical role
(33,
54,
55). Learners benefit most from knowledge that is continually presented in terms of actual patients being treated in the primary care setting
(54,
56). To do this, psychiatrist teachers require an understanding of and close working relationship with primary care physicians
(54). One highly effective way of doing this is to combine education with clinical service right in the primary care setting
(54). Models of “shared care” locate both teaching and patient care in the primary care office, bringing knowledge and skills directly to the context in which primary care physicians will apply them. Similarly, “academic detailing,” in which a faculty member meets one on one with a primary care physician, allows personal, contextually relevant interactive dialogue
(55).
Finally, at a time when most physicians have a computer in their offices, we were surprised to find few articles that focused on the computer as an educational tool. A few articles described the potential value of programs that allowed primary care physicians to arrive at DSM diagnoses by answering a series of computer-generated questions
(57–
64). Theoretically, learning was enhanced by displaying information alongside each diagnostic question and by providing an explanation with the final diagnosis. One study
(58) demonstrated that primary care physicians made diagnoses more accurately and faster, whereas another
(59) showed that recording and comparing provisional diagnoses and choices of drug therapy was helpful in assessing gaps in knowledge. However, these were rudimentary reports, making it impossible to evaluate the learning benefits of computer-based continuing education. Undoubtedly, this will be a major area of growth and investigation in the coming years.
Effectiveness
Finally, in recent years, a growing number of articles have evaluated the effectiveness of educational interventions. There are many examples of interventions that have increased the knowledge of primary care physicians on written postcourse tests. For example, significant growth of knowledge was shown regarding medication use
(60), AIDS
(61), and symptoms of depression and anxiety
(62,
63), among other topics.
At a more comprehensive level, research has demonstrated improvement of global psychiatric knowledge and skills at the end of courses ranging from a few sessions to 6 months by use of multiple-choice tests, video reviews, and objective standardized clinical examinations
(64–
70). Improvement of interviewing skills has been shown in several studies that made use of videotaping with feedback, practice with simulated patients, and problem-based learning
(71–
73). In a rare case in which primary care physicians themselves designed objectives, Kaaya and colleagues
(74) demonstrated improved skills in the management of somatoform disorders using role playing and videotaping over an 8-week course. Finally, using a shared-care model, Kates et al.
(75) brought psychiatrists into the offices of 88 primary care physicians in Canada. After more than 1,000 consultations in 2 years, the authors found improved communication between primary care physicians and psychiatrists and enhanced continuity of care.
However, there were a number of studies that had negative findings. Adeyemi and Jegede
(76) attempted to improve interviewing skills by providing psychiatrist feedback after patient interviews over 3 months. Of seven participants, only three primary care physicians became better at detecting psychiatric disorders, two showed no noticeable improvement, and two showed some deterioration. Improvement was more related to personality and disposition than experience
(76). Jones and colleagues
(77,
78) and Gaskins et al.
(79) also failed to demonstrate a positive outcome of a program that included seminars and ongoing supervision of patients with chronic psychiatric disorders for 6 months. And Carr et al.
(80) found no evidence that availability of community consultation improved clinical confidence, referral patterns, or psychiatric knowledge of primary care physicians. Furthermore, this group reported that the best predictors of psychiatric knowledge were young age and female gender. Finally, Callahan and colleagues
(81) attempted to improve the primary care of depression in later life. Primary care physicians recruited elderly depressed subjects and scheduled them for three additional visits to address their depression. The primary care physicians were given depression rating scale scores for the patients, as well as a list of medications used to treat depression and an educational flyer to share with the patients. A control group of physicians saw their patients as they normally would have, without any additional appointments or materials. Although the physicians with intervention were more likely to diagnose depression and to prescribe antidepressants, neither the intervention nor the control group of patients showed significant improvement in depression scale scores or sickness impact profiles.
Similarly, two studies failed to show the effectiveness of academic detailing. The first
(82), which emphasized implementing practice guidelines for depression, showed an increased rate of treatment of mental disorders but no improvement of the patients’ symptoms or functional status. The second
(83), which combined case consultation and role playing about depression in an office setting, failed to demonstrate any lasting effect of antidepressant selection, adequacy of dose, or patient satisfaction outcomes.
Particularly troubling was the paucity of studies that went beyond short-term, uncontrolled measures of learning. There was only one study in the literature
(84) that described a rigorous long-term evaluation of a continuing education program. Rutz and colleagues
(84) wanted to educate general practitioners on the Swedish island of Gotland about depressive disorders. They developed a 20-hour program administered over 3 days that focused on the symptoms, causes, diagnosis, prevention, and treatment of depression for 90% of the primary care physicians in the community of 56,000. After the program, the number of referrals by the participants to psychiatrists decreased significantly, as did the amount of hospitalization for depression. There was a significant increase in antidepressant use and slightly decreased suicide rates on the island
(85). Had the authors concluded that the intervention was successful, however, they would have missed an important phenomenon. In a follow-up study conducted 3 years later, they found that the suicide rate had increased to its previous baseline level, hospitalization had again increased, and the prescription rate had stabilized
(86). The early effect of their educational program was temporary, and the course had to be repeated.
The study by Rutz and colleagues
(84) was unique in the literature in including a long-term follow-up, raising serious concerns about the positive findings reported immediately at the end of other educational interventions. The study by Rutz et al. was one of only four in the literature that demonstrated clinical practice outcomes
(21). Such measures are far more useful than satisfaction surveys or written tests that are frequently the only measures used to draw conclusions about the effectiveness of educational interventions.
In 1999, Kroenke and colleagues
(21) attempted to perform a meta-analysis of 48 experimental studies in psychiatric education for primary care physicians published between 1966 and 1998. Unfortunately, the considerable heterogeneity of the studies precluded such an analysis. Nevertheless, their findings were interesting. They found improvement in the diagnosis of mental disorders by learners in 18 of 23 (78%) of the studies and reported improvements in treatment in 14 of 20 (70%) of the studies. The authors concluded that a variety of interventions appeared to be effective but that the literature was insufficiently advanced to allow the identification of specifically effective educational methods
(21). In addition to calling for further research, they emphasized that effective interventions were multifaceted, conducted on site, individually tailored, and ongoing. They also noted that the effects of managed care had been insufficiently considered in studies and that the limited time, medical comorbidity, somatization, and stigmatization of mental disorders in the primary care context needed greater emphasis
(21).
Discussion
Ongoing, interactive, contextually relevant continuing education can improve the knowledge, skills, and attitudes of primary care physicians who provide care for people with mental illness. And for the psychiatrists and primary care physicians interested in continuing education, there is a growing literature to help guide the selection of objectives, educational methods, and forms of evaluation. However, there remain few methodologically sophisticated outcome studies and a surprising number of negative findings. Certainly, much more high-quality research is needed.
Most psychiatric educators have provided short courses focusing on diagnosis and pharmacological treatment that bear little relationship to the perceived needs of primary care physicians. Skills acquisition and application in clinical settings have been underemphasized by courses delivered didactically in locations removed from the context of real clinical practice. In order to improve the quality of continuing education provided for primary care physicians, we synthesized the following recommendations from the literature:
1. Before setting learning objectives, psychiatric educators must conduct a needs assessment of the physicians who will participate. It is not sufficient to derive objectives from a different group of learners or to rely on personal assumptions.
2. Course planners should involve leaders in primary care who can help define competencies in mental health care that are important in the primary care context.
3. Psychiatrists teaching primary care physicians should familiarize themselves with the context of primary care. One of the best ways of doing so is to actually work in a primary care setting, using a model such as shared care.
4. Objectives should be created for each of the domains of knowledge, skills, and attitudes. Objectives should be worded in behavioral terms that describe what the primary care physician will actually be able to do after the educational program. The articulation of clear, behaviorally specific objectives makes it much easier to assess outcome.
5. Particular attention should be paid to attitudinal issues. The program should include opportunities to discuss the primary care physicians’ perceptions of psychiatry, mental health care systems, and mental illness itself.
6. Learning should be tied to real clinical practice. Course planners should leave time for participants to raise questions and discuss problems that arise in their own clinical practices. The more the course material is connected to real problems, the greater will be the application of new learning.
7. Learning methods should be interactive and involve practice of new skills under observation. Videotape review and feedback, role playing, and use of standardized patients are all excellent methods.
8. The degree to which learning has occurred should be evaluated. Assessment measures should go beyond learner satisfaction and include multiple measures that assess knowledge, skills, and attitudes.
9. The best measures of outcome are changes in actual clinical practice, such as changes in the rates of prescription or referral, illness detection, and hospitalization. Demonstrating improved patient outcomes is even better.
10. Programs should be ongoing rather than single sessions and should be supplemented with educational refresher courses and periodic reassessments over time. Educators should watch for erosion of knowledge, skills, or attitudes over time by creating programs that are “tenacious,” reengaging learners repeatedly over time.
We believe that attention to all of these issues will strengthen educational programs for primary care physicians and also improve research aimed at demonstrating the effectiveness of educational programs in psychiatry.