Patient and provider gender have been cited and studied as potential variables in disparities in health care delivery, outcome, and diagnostic differences since the 1970s. Since the Tobin et al. (1) 1987 study reported a sixfold difference in the rate of coronary artery catheterization based on the gender of the patient, other studies have addressed this potential disparity (2–4). A review of management by race and sex in 2005 showed that the differences have not narrowed in recent years (5). New data analysis suggests that the pathophysiology underlying acute myocardial ischemia differs between men and women, and thus the presentation, diagnosis, and treatment may need to be gender-specific (6). Likewise, the diagnostic evaluation and treatment recommendations of deep venous thrombosis (7) and colon cancer screening practices (8) have been shown to vary by the gender of the patient as well as the physician.

Safran et al. (9) studied practitioner and patient gender, race, and age differences in post-operative activity restriction recommendations. They discovered that female patients’ activities were restricted three and a half times more often than those of their male counterparts after standardization for severity of illness. This activity restriction was imposed predominantly by the older physicians in the study. The female patients were perceived to be more frail and in need of special treatment (9).

Psychiatric disorders vary by gender in their prevalence and their recommended treatments. Community-based surveys, such as the Epidemiological Catchment Area studies (10), circumvented this problem by looking at local populations. They found that women are two times more likely than men to experience an episode of major depressive disorder in their lifetime. Prior to the widespread use of selective serotonin reuptake inhibitors (SSRIs), Hohmann (11) found primary care physicians prescribing antidepressants to women two times more often than to depressed men. More recently, in 2001, Olfson et al. (12) investigated gender bias in outpatient psychiatric treatment of patients with major depressive disorder and found no significant difference in treatment recommendations. However, they did find that female patients of female psychiatrists were much less likely than female patients of male psychiatrists to receive an assessment of their sexual functioning (12). Although Kales et al. found no overall difference in likelihood of primary care physicians (13) and psychiatrists (14) to diagnose depression using vignettes and photos of a male, female, Caucasian, or African-American geriatric patient, some ethnic and training characteristics of the physicians did have an impact on the approach and diagnosis made.

With the emergence of the “greying” of America (15, 16), increasing emphasis has been placed on geriatric undergraduate and graduate medical education. Geriatric depression and cognitive decline place a special burden on geriatric psychiatric education (17, 18). Late life depression remains underdiagnosed in the primary care sector, with additional differences related to race and clinician bias (13, 19). According to the 2000 U.S. Census Report (20), 12.4% of the population is 65 years of age or older. Of this group, 59% are women, and if only those older than 85 years are considered, 71% are women. The magnitude of these gender differences, as well as the “graying of America,” raises the concern about the influence of gender bias in the treatment of this patient population. There are many gender-related health, longevity, and socioeconomic variables that affect this group. Women are more likely than men to outlive their partners, live alone in poverty, and be institutionalized in a nursing home. With the “graying of America,” it is estimated that by 2020, 18% of the population will be older than 65 and by 2050 they will exceed 25%, thus emphasizing the importance of gender considerations (21).

Standardized patients are used in the Observed Structured Clinical Examination (OSCE), a routine clinical skills assessment instrument in undergraduate and graduate medical education. The impact of student and standardized patient gender on these examinations, student performance, and other medical education settings has been examined in a number of studies. According to Colliver et al. (22), male and female medical student examinees performed equally well with respect to interpersonal and communication skills, irrespective of the gender of the standardized patient. Furman et al. (23) found no significant interaction of student gender and standardized patient gender on history-taking, physical examination, or communication skills. By contrast, Wiskin et al. (24) found female students performed better overall than male students on their Valid Oral Interactive Contextualized Examinations (VOICEs), which are used in the United Kingdom to evaluate medical student clinical skills. Hatala and Case (25) examined the role of USMLE Step 2 performance based on the gender of the patients in case vignettes. They found that students did less well on vignettes that involved female patients compared to vignettes involving male patients, often making the wrong diagnosis or intervention.

We took the opportunity of our annual OSCE to examine whether gender would influence students’ formulations of the differential diagnoses in a geriatric clinical encounter. We designed a case of a 65-year-old recently widowed patient, living alone, with moderate cognitive decline, some sadness and dysphoria, and systemic complaints. We took advantage of the mixed and often comorbid conditions and shared clinical characteristics of depression and cognitive decline in the elderly (13, 26). We hypothesized that the female standardized patients would more likely be diagnosed as depressed than male standardized patients. This was related to the above cited health disparity and educational literature, as well as our own observations of students during their clinical rotations. We believed that in spite of our educational efforts, our medical students would still see a female patient as more vulnerable than a male patient.

During the summer semester of 2003, 196 fourth-year medical students were required to take a 10-station OSCE. Cases varied across all disciplines with one above-cited geriatric case in which we utilized both a male and a female standardized patient. Faculty exam case writers reviewed each case for clinical appropriateness and standardized patient portrayal. They tested and observed each case to optimize its validation and appropriateness for use. Both the male and female African-American standardized patients were identically trained together in portrayal, and they observed each other’s portrayals frequently over the course of the OSCE to enhance standardization (27, 28). Ten students were randomly assigned to each examination date. The female standardized patient worked 11 days and the male standardized patient worked 9 days. Students were given 20 minutes to interview and examine each standardized patient and 10 minutes to document their findings. Each report consisted of an initial history, physical and differential diagnoses, and treatment recommendations. Standardized patients also rated students’ professionalism, whether they asked appropriate questions and performed pertinent physical examinations. These checklist findings are not included in this study.

Researchers reviewed all students’ written reports. We focused on their choice of primary diagnosis in relationship to the standardized patient and student gender. Results by standardized patient and student gender were analyzed using a chi-square. The student’s race was not examined. In addition, we observed six live exams on site and reviewed videotapes of another eight student examinations to assess the performance of the standardized patients and examination conditions.

Eleven of the 196 examinations were excluded from analysis for the following reasons: the sex of the student could not be determined in five and that of the patient in two, and four were excluded due to insufficient written material. The remaining 185 students and standardized patient distribution by gender are shown in Table 1 (78 students examined male patients and 107 students examined female patients). Tables 2 and 3 show the two major diagnoses, cognitive disorder or major depressive disorder, made by all of the students.

Overall, the female standardized patients were more often diagnosed as depressed compared to the male standardized patients (p≤0.05), when the gender of the student was not considered. When the gender of the student was looked at, the female students were more likely to diagnose the women patients as depressed than their male patients (p≤0.025). There was no significant difference in the distribution of diagnosis made by the male students in either the male or female standardized patients. All diagnoses of cognitive disorders by either group of students and standardized patients were not significantly different among the groups (Table 3).

Concerns about gender bias in medical education and practice have received considerable attention in recent years. Our literature review outlined enduring gender differences in the treatment of chest pain, depression, health screening practices, and some aspects of student performance on standardized examinations.

We were very surprised to discover that our female students were much more likely to diagnose the female standardized patients with depression than their male counterparts. Just as in the study of the USMLE Step 2 performance (25) whose students followed known demographic differences, our students diagnosed the illness of depression more often in female patients than in male, which coincides with the higher prevalence of depression in women. However, the predominant effect was seen in the female students, a result that we cannot explain.

Cultural and social science issues may play a part in this gender bias. Armitage’s classic study of Anglo-Americans in Southern California (29) reported the existence of stereotypes that regard the male as typically “stoic” and the female as typically “hypochondriacal.” Likewise, Nathanson (30) suggested that women report more illness than men because it is culturally more acceptable for them to be ill. These stereotypes, though reported in the past, have been cited in different ways in the current literature. This, in part, may explain Safran et al.’s findings (9) that women are more likely to be treated as “frail” and are prescribed additional post-operative activity restrictions.

The ambiguity of the written case vignette, although it reflects the frequently ambiguous clinical presentation of the elderly, can be a confusing one for trainees. Fourth-year students may have difficulty making the challenging clinical distinction between geriatric depression and cognitive decline because these share many characteristics. Their clerkship training has been discipline-specific, and this case required crossing the boundaries of primary care, internal medicine, neurology, and psychiatry—a task often difficult for even experienced practicing clinicians.

Although we anticipated that female standardized patients were going to be diagnosed with depression more often than male standardized patients, the finding that the female students were the main contributors to this difference was unexpected. Perhaps the female medical students identified more with their female patients. They may have projected their belief that they would be depressed if they were subjected to the same circumstances that the standardized patients portrayed.

Although the standardized patients were trained together and observed each other to standardize their portrayal, the actual interaction with patients and the medical students is unknown. Faculty directly observed only a small number of the encounters. It is possible that the standardized patients, in the presence of the students alone, may have portrayed their story and symptoms differently to the female than to the male students. The elderly female standardized patients may have seen more male physicians over their lifetimes and may have been more accepting of the male student physicians. This familiarity may have made them more willing to disclose their information, leading to the different diagnosis made by the male students. Their personal mannerisms and affect may have differed with the different students as well.

We may need to rethink our notion of standardization of standardized patients, with respect to gender. In addition, consistent with standard OSCE and standardized patient methodology, the case was specific to the individual patient history and manner. The extent to which findings derived from it can be generalized is, therefore, necessarily limited.

Clinical competency is assessed at the bedside throughout clinical clerkships, during school-based OSCEs and now nationally through the standardized USMLE Step 2 Clinical Skills Assessment. We need to be vigilant regarding the potential of gender bias in all of these evaluation tools. As research continues to investigate the variability of disease manifestation by gender, we will have to incorporate this evidence into our examination design. The gender of standardized patients, trainees, and practitioners may continue to influence diagnoses made, clinical management decisions, and outcome. This will need to be considered when we set performance standards for our students in standardized clinical encounters.