The HIV/AIDS (human immunodeficiency virus/acquired immunodeficiency syndrome) epidemic has resulted in a tremendous health care burden with an increasing need for mental health services. Hence, psychiatrists have both a unique opportunity and a professional obligation to provide care to HIV-affected persons in a competent and compassionate manner (1). To provide professional guidance, the American Psychiatric Association (APA) has developed a Position Statement on AIDS that, among other things, calls attention to psychiatrists' responsibility to educate themselves, their colleagues, and patients about the medical, psychiatric, social, ethical, and legal aspects of HIV infection (2).
Targeted psychiatric education about HIV/AIDS is essential to ensure that psychiatrists have appropriate knowledge to understand the predominantly behavioral mechanisms underlying viral transmission, the myriad neuropsychiatric complications linked to HIV/AIDS, and the substantial psychosocial ramifications of living with HIV disease. However, 15 years into the HIV/AIDS epidemic, little research has examined the HIV-related educational needs of psychiatrists, particularly psychiatrists in training. This study, through the implementation of a national, randomized survey, sought to examine the current level of knowledge and attitudes among psychiatrists in training about HIV/AIDS.
Two studies offer limited information about psychiatrists and HIV/AIDS attitudes. The first of these, a 1988 survey of program directors of 207 accredited psychiatry residency programs, revealed that only 69% of programs offered formal HIV/AIDS instruction (3). Further, the median number of hours of instruction was minimal, with a median of 2 hours spent on HIV-related neuropsychiatric complications (offered in 64% of programs), a median of 1.5 hours spent on other biomedical aspects of HIV/AIDS (offered in 46% of programs), and 2 hours spent on psychosocial aspects of HIV/AIDS (offered in 56% of programs).
In 1991, Chaimowitz (4) administered a questionnaire to measure AIDS-related attitudes and homophobia among 72 Canadian psychiatry faculty, psychiatry residents, and family practice residents and found that 12%, 23%, and 26%, respectively, endorsed prejudice against homosexuals. When asked whether they thought homosexuals with AIDS "got what they deserved," no psychiatry residents responded affirmatively, whereas 5% of family practice residents and 7% of psychiatry faculty answered yes.
These limited studies suggest likely gaps in knowledge and attitudes that are apt to adversely influence the provision of HIV-related mental health care. These studies also highlight the lack of information about HIV/AIDS knowledge and attitudes among psychiatrists. As such, the study reported herein is intended to contribute further empirical data on HIV/AIDS knowledge and attitudes of psychiatrists, with a specific focus on psychiatrists in training.
A sample of 2,252 residents and fellows (postgraduate years [PGY] 1—6) were randomly selected and geographically stratified by APA district (Areas I—VII) from the 1993—1994 APA Census of Residents. The rationale for choosing 2,252 mailings was based upon our desire to obtain at least a 10% representation of the census (6,059 addresses), expecting an average 30%—50% return rate (5,6). Each potential participant was mailed a survey questionnaire during a 4-week period June—July 1994, and asked to complete and return it anonymously by mail. By October 1994, 825 surveys (37%) were completed and returned, comprising our study sample. This return rate is consistent with other reports using return mail questionnaires in behavioral science (5) and with other HIV/AIDS surveys of health professionals in training (6). Overall, 53% of the sample were women, 69% were within their first 4 years of training, over half had seen 10 or more HIV patients during training (57%), most (82%) were in urban training sites, and all APA districts were similarly represented. There were no significant differences among these demographic variables for respondents. T1 summarizes the demographic information for all respondents and compares available national APA Census information with our sample.
The survey instrument consisted of 15 knowledge items, 15 attitude items, demographic information, and questions regarding how HIV/AIDS affected residency selection and self-perceived needs for further HIV-related education. The instrument was adapted from an earlier, similar survey of medical students (7) and HIV-related educational material specific to psychiatrists contained in the APA AIDS Education Project training curriculum (8). To establish face and content validity for the survey, an initial draft of the instrument was distributed to an experienced, multidisciplinary team of mental health clinicians working in a large, public HIV clinic.
The HIV knowledge assessment consisted of 15 statements scored as True, False, or Don't Know (T2). An overall knowledge score was computed by giving 1 point for each correctly answered item, counting "Don't Know" as an incorrect response; the possible range of the knowledge score was 0 to 15. The 15 HIV attitude-assessment items were rated on a 5-point Likert scale, ranging from strongly agree [SA] to strongly disagree [SD] (T3). Items were scored such that a higher value represented a less desirable response.
In addition to demographic information, the respondents were asked to choose one of five responses about how working with HIV patients might have influenced their residency selection. This ranged from "Tried to avoid HIV patients as much as possible"  to "Wanted to see as many HIV patients as possible" . The respondents also were asked to endorse any of seven HIV/AIDS topic areas in which they perceived further training needs: cultural competency, death and dying, neuropsychiatric complications, risk prevention, sexual behavior, sexual orientation, and women's issues.
We used separate analyses of variance (ANOVAs) to examine differences in total knowledge scores for gender (m; f), training location (urban; suburban; rural), HIV experience (none; 1—10 patients; more than 10 patients), residency year (PGY-1 to PGY-5 or more), and APA district (I to VII). We conducted ANOVAs using these variables to examine differences in scores reflecting how HIV affected residency choice.
Finally, instead of examining a "total attitude score," which would obscure specific patterns of misconceptions about HIV/AIDS, we ran a descriptive frequency analysis of all responses (SA to SD) for each item. As our interest was focused on possible negative attitudes and how they relate to training, we conducted ANOVAs for each of the three items endorsed desirably by less than 75% of the respondents, for the five independent variables listed earlier. Post-hoc analyses were conducted using Schefee's test (P<0.01) to discern specific differences among the levels of these variables, if the overall F-value was significant. We also conducted χ2 analyses to look at residents' perceived training needs according to how they responded to corresponding knowledge and attitude items.
The mean knowledge score was 9.57 (SD = 2.3; n = 815), with a range from 2 to 15 and a median of 10. Internal consistency reliability was computed using Cronbach's alpha and Split-Half with Spearman Brown correction for unequal length forms; both coefficients were in acceptable ranges (r=0.75 and r=0.78, respectively). T2 summarizes the percentage of correct, incorrect, and uncertain responses for each knowledge item. Most items (70%) were answered correctly by at least half of the respondents.
Excluding "Don't Know" responses, two items were incorrectly endorsed by a majority of the respondents: "Dementia due to HIV disease is rare in asymptomatic HIV patients" (70% incorrect) and "The only bodily fluids known to transmit HIV are blood and semen" (53% incorrect). The most intriguing findings regarding the knowledge items were among five neuropsychiatry items. For each of these, although from 47% to 76% of the respondents answered correctly, those who did not do so tended to endorse "Don't Know" as opposed to the incorrect answer. These items dealt with neuropsychiatric complications and management issues (e.g., opportunistic infections, medication side effects; pharmacotherapy, encephalopathy). Finally, a sizable number of the respondents endorsed "Don't Know" for three other items (from 16.8% to 18.2%). These items dealt with issues concerning special populations (e.g., women, newborns, persons of color).
Significant differences were found for the total knowledge score for gender (F=5.55, df=1, 812, P<0.02), number of HIV clients seen (F=19.69, df=2, 812, P<0.0001), and year of residency (F=2.97, df=4, 810, P<0.02). The men tended to have higher mean scores than the women (9.74 vs. 9.36, respectively). Post-hoc analyses indicated that for HIV clinical experience, there were significant differences in the mean knowledge score: the residents who saw more than 10 HIV patients (9.99) were more knowledgeable than those who saw from 1 to 10 patients (9.05) or no patients (8.45). Post-hoc analyses revealed no differences among the five levels of training, but the range was from a mean score of 9.29 (PGY-2) to 9.96 (PGY-5 or greater).
T3 shows the percentage of the respondents who agreed (SA/A), were neutral (N), or disagreed (D/SD) with each attitude item. Generally, attitudes were not extreme; the range in mean scores was from 2 to 3 (Agree to Neutral) in most cases. However, some important attitudinal items deserve comment. For example, 9% of the respondents felt uncomfortable coming into social contact with HIV-positive persons, 8% would refer HIV patients if they did not risk professional recrimination, 7% would quarantine HIV patients to stop the spread of infection, 8% believed homosexuality to be a psychiatric illness, and 9% would choose to opt out of treating HIV patients in residency training.
T4 summarizes significant findings from the series of ANOVAs run for the three items endorsed in the desired direction by fewer than 75% of the respondents: "I feel comfortable managing neuropsychiatric complications of HIV infection" (49% SA or A), "I feel comfortable working with dying patients" (69% SA or A), and "I believe there is a broad range of normal sexual behavior and homosexuality falls within this range" (71% SA or A). Significant gender differences were found for each item: the women were more comfortable managing HIV-related neuropsychiatric complications (F=19.62, df=1, 818, P<0.0001) and were more likely to view homosexuality as normal (F=13.56, df=1, 814, P<0.0005). The men were more comfortable working with dying patients (F=8.23, df=1, 816, P<0.005).
Regarding amount of HIV experience, the residents differed significantly regarding comfort working with dying patients (F=4.52, df =2, 816, P<0.01), acceptance of homosexuality (F=3.75, df=2 ,814, P<0.03), and comfort with neuropsychiatric complications (F=40.92, df=2, 818, P<0.0001). However, post-hoc tests indicated significant differences between residents with different levels of experience only for neuropsychiatric complications; the residents who saw more than 10 HIV patients expressed significantly more comfort than those with less experience.
Similarly, regarding residency year, the residents differed significantly regarding comfort working with dying patients (F=5.52, df=4, 814, P<0.0005), acceptance of homosexuality (F=4.50, df=4, 812, P<0.002), and comfort working with neuropsychiatric complications (F=4.06, df=4, 816, P<0.003). Post-hoc tests indicated significant differences between residents from different residency years regarding comfort with neuropsychiatric complications and working with dying patients; those beyond PGY-4 felt more comfortable working with neuropsychiatric complications than the PGY-1 residents, whereas the residents beyond PGY-3 felt more comfortable than the PGY-1 residents working with dying patients.
There were no significant differences among respondents in urban, suburban, or rural training locations in working with dying patients. Significant differences were found for training location in managing neuropsychiatric complications (F=3.02, df=2, 807, P<0.05) and acceptance of homosexuality (F=3.54, df=2, 803, P<0.03), but there were no specific post-hoc findings among the levels of this variable. Significant differences were found among APA districts for comfort working with dying patients (F=4.59, df=6, 802, P<0.0001) and acceptance of homosexuality (F=3.85, df=6, 807, P<0.001), but no differences were found for comfort managing neuropsychiatric complications. The residents in New York rated themselves significantly less comfortable working with dying patients than those in the Midwest and the West; those in California were significantly more positive in viewing homosexuality as part of normal sexual behavior than those in the South.
T5 summarizes data regarding HIV-related training needs. Seventy-seven percent of residents stated they wanted more training on the neuropsychiatric aspects of HIV. For each of the eight neuropsychiatry knowledge items, a majority of residents (>70%) desired more training irrespective of whether they responded correctly, incorrectly, or endorsed "Don't Know." Similarly, the respondents expressed a desire for further training in this area regardless of their response to the attitude item reflecting comfort in managing neuropsychiatric complications.
The next most frequent training area identified dealt with the topic of death and dying, endorsed by 48% of the respondents. There was a significant relationship between the subjects' attitudes and their perceived need for training (χ2[1,2]=71.8, P<0.0001). Sixty-one percent of the subjects who expressed comfort working with dying patients did not perceive a need for further training, whereas 75% of those who endorsed discomfort did perceive the need for further training. This was also the case for the respondents' perception of training needs in the area of risk prevention (χ2[1,2]=26.3, P<0.0001). For all other areas of training need, the majority of respondents did not perceive a need for further training irrespective of how they endorsed any particular corresponding knowledge or attitude item.
Finally, there were no significant differences in the likelihood that working with HIV patients would affect one's residency selection. Seventy-eight percent of the respondents stated that this was "Not a Consideration" in their choice.
The results indicate that, for the most part, psychiatrists in training possess reasonable knowledge and generally positive attitudes about treating patients with HIV/AIDS. Most attitude scores were within the middle-to-positive ends of the rating scale, and the possibility of working with HIV patients did not influence residency choice. However, there also were significant deficiencies in some residents' knowledge and comfort level on the diagnosis and management of HIV-related neuropsychiatric complications, an area where psychiatrists are uniquely qualified to manage patients. Importantly, 77% of all respondents perceived a need for further education regarding neuropsychiatric aspects of HIV/AIDS, rating this area highest among HIV-related educational needs.
With regard to attitudes about comfort in managing neuropsychiatric complications, the women were more comfortable than the men, as were individuals in urban areas compared with those in rural and suburban areas. Those residents who had more experience with HIV patients expressed less discomfort in managing complications, and residents with more years of training were most comfortable in this area. These findings suggest that HIV-related education on neuropsychiatric issues should sensitively address possible discomfort on the part of male residents and residents in nonurban locales. Further, HIV/AIDS education and clinical experience early in the course of residency training may be important for increasing residents' comfort in managing neuropsychiatric complications.
Education targeting experience with terminally ill patients remains an important curriculum component for all residents, especially for HIV-related work. Almost half of the residents indicated they wanted more training during residency about death and dying, making it the second-most requested training topic. The majority of the residents who expressed discomfort working with dying patients also desired more training. Study findings reveal that residents' level of comfort about these issues increased with more exposure to HIV patients; however, even residents with significant exposure did not express high comfort levels.
Regarding attitudes about homosexuality, the men were less likely to believe that homosexuality reflects normal sexual behavior, a gender difference that has been mirrored in other studies (9). Urban residents were more likely to agree that homosexuality was within the normal range of sexuality. Numerous studies, including a recent report by the American Medical Association's Council on Scientific Affairs (10), have shown pervasive negative attitudes about gay men and lesbians throughout the health professions, including among mental health professionals (11). Moreover, a 1994 membership survey conducted by the Gay and Lesbian Medical Association revealed significant discrimination against gay and lesbian psychiatrists and patients (12). Our study, like others before it, suggests that educating residents about sexual orientation will be a continuing challenge, especially because only 20% of the respondents expressed interest in further training on this topic.
Most psychiatry residents did not consider amount of contact with HIV patients when choosing their residency, a finding that contrasts with prior studies of nonpsychiatric residents (13). While this study did not examine residency selection further, the lack of concern among psychiatry residents may be related to the fact that psychiatrists have less exposure to bodily fluids than other disciplines, or that psychiatry residents have less fear or prejudice toward HIV/AIDS patients. Future research could specifically address these issues.
Regarding this study's findings, although statistically significant differences were found among several demographic variables for attitude and knowledge scores, these differences may not be practically meaningful given the large sample size. For example, a significant difference between the mean knowledge scores of men (9.74) and women (9.36) may lack practical importance given the small actual difference. Furthermore, in any questionnaire study of this design, it is not possible to ascertain whether attitude responses truly reflect actual attitudes as opposed to respondents' attempting to answer correctly. From a methodological perspective, the possibility of sampling bias also arises. On the one hand, it is possible that individuals who did not respond to the survey may have had little interest in HIV/AIDS or may have been adverse to patients with HIV/AIDS and those who did respond were more interested and positively disposed, thus skewing the sample toward residents with more positive attitudes and/or more knowledge. Alternately, more positively interested residents may not have responded, whereas residents with more negative attitudes did.
These findings suggest that HIV/AIDS training in psychiatry residency programs should focus particularly on the wide spectrum of neuropsychiatric complications in HIV disease. The findings also draw attention to the need to address in current curricula and in future research the issues of sensitivity toward homosexuality, comfort in working with dying patients, and HIV-related issues of special populations, including persons of color, women, and children. We feel that it is critical that residents maintain up-to-date HIV knowledge and appropriate attitudes to enable them to discuss sensitively such key components of HIV-related mental health care as prevention methods and sexual history taking. Perhaps HIV/AIDS education can be more fully integrated into other curriculum topics, such as psychiatric care of the medically ill, consultation-liaison, substance abuse, sexuality training, and cultural diversity training. Further investigation of current educational efforts is needed to provide directions for implementing HIV/AIDS training and to assess the extent of negative biases. Future research should seek to establish the extent of current HIV-related training, perhaps through surveying directors of residency training programs. If, as earlier research has indicated, most psychiatry residency programs offer HIV/AIDS education (3), then training content and methods may need evaluation to further understand the deficits in knowledge and attitudinal barriers found in this study.
This study was supported by the Substance Abuse and Mental Health Services Administration, Center for Mental Health Services Grant (Grant No. 1T15MH19894) to Dr. McDaniel.
The authors thank Miles Crowder, M.D., and Nancy Thompson, Ph.D., M.P.H., for their assistance in survey methodology and manuscript preparation.