Sexually transmitted diseases (STDs) account for a substantial burden of morbidity and mortality in the United States, including acute illness, long-term sequelae (such as acquired immunodeficiency syndrome [AIDS], certain cancers, infertility, and other chronic conditions), and premature death. Clinical and public health measures against STDs include early detection and treatment as well as educational interventions aimed at reducing sexual risk.1 Evidence-based guidelines, such as those developed by the United States Preventive Services Task Force, recommend that physicians counsel their patients about how they can reduce their risk of acquiring STDs including human immunodeficiency virus (HIV).1 Recent guidelines on HIV also encourage primary care physicians to talk to their patients who are HIV positive about how they can help prevent secondary infections.1 Thus, physicians are in an excellent position to help prevent STDs and to slow the spread of the HIV epidemic.2
Despite the importance of physicians talking to patients about STDs, studies have shown that some physicians do not follow clinical practice guidelines for reducing STDs and that many practicing physicians and medical students do not feel comfortable assessing their patients' risks of sexually acquired illness or talking to their patients about risk-reduction practices.3–8 Thus, there is an ongoing need for training in sexual health in the medical curriculum and for continuing education opportunities to help primary care physicians integrate HIV and STD testing and counseling into their practices.
To date, few studies have examined the knowledge, attitudes, and behaviors of medical students regarding safe sex, including how these variables might change as the students advance through their medical education. Moreover, the relationship between the personal sexual health practices of medical students and their anticipated future practice is inadequately understood. To better understand the personal and clinical safe-sex–related knowledge, attitudes, and practices of medical students, we analyzed data from a study of students at 16 schools in the United States.
All medical students in the class of 2003 at 16 U.S. schools were eligible to complete three questionnaire administrations during their medical education (at freshman orientation, at entrance to wards, and in their senior year). We selected our judgment sample of schools to reflect all U.S. medical schools in terms of age (freshman average 24 years old versus 24 nationally), school size (average students per school 563 versus 527 nationally), National Institutes of Health research ranking (school average 64 versus 62 nationally), private/public school balance (51% private schools versus 41% nationally), underrepresented minorities (13% African Americans, Hispanics, and Native Americans, versus 11% nationally), sex (45% women versus 43% nationally), and geographic distribution.9–12 A 17th school was excluded from all analyses for nonadherence to protocol and resultant poor response rates.
Following our Emory institutional review board-approved protocol, the confidential questionnaires were administered to students in varied venues and at varied times of year outside of formal classroom or training time (most often given before or after a required event such as an orientation). The students were instructed that their participation was voluntary and that they could choose to withdraw from the study at any time. When necessary, we used Dillman's five-stage mailing process13 to maximize response rates. An individual's responses were linked across time using a unique identifier consisting of their mother's initials at her birth and their father's first two initials. At freshman orientation, 2,080 students were eligible to complete the survey, and 1,846 responded; 1,982 were eligible at entry to wards, and 1,630 responded; 1,901 were eligible at senior year, and 1,469 responded. Of the 2,316 students who provided responses, 71.6% (n=1,658) did so at more than one time point; 971 responded at three time points, 687 at two, and 658 at one. As suggested by Diggle et al, all available observations were used in analytic procedures that took into account repeated measurements. The overall response rate across the three time periods was 83% based on a denominator of the sum of student enrollment at each time point and a numerator of individual responses at those times. Rates of nonresponse for individual questionnaire items were a median of 3%. Our study was planned to have greater than 90% power to detect small differences (Cohen's effect size 0.1) when n=1,400 and alpha=0.05.
Our primary personal outcome was the response to the question, “Are you currently trying to practice safe sex when sexually involved?” Response choices were “No, not applicable,” “No, not trying,” “Yes, low priority,” and “Yes, high priority.” We examined the subset of students who were neither married nor a member of an unmarried couple for the association of intentions to practice safe sex with personal and professional characteristics: gender, race or ethnicity, self-determined risk of developing HIV infection, strength of religious identity, excessive alcohol consumption, time point in medical schooling, attitude about physicians' responsibility to promote prevention, relevance of safe-sex counseling to intended medical practice, and frequency of counseling on safe sex. If an unmarried, uncoupled respondent said that trying to practice safe sex was not applicable, we placed them in a sexually inactive category.
Our primary professional outcomes were two variables concerning medical students' talking to patients about safe sex: 1) “How relevant do you think talking to patients about safe sex will be in your intended practice?” and 2) “With a typical general medicine patient, how often do you actually talk about safe sex?.” The possible responses for relevance were “not at all/somewhat/highly,” and for frequency they were “never-rarely/sometimes/usually-always.” Relevance was queried at all three time points (T1, T2, and T3); frequency was measured only on the senior year survey (T3). Secondary safe-sex counseling outcomes of interest were the training received on safe sex (T2 and T3), the students' confidence in discussing the topic with patients (T2 and T3), and how important they thought it was for physicians to talk to patients about safe sex (T2). The same questions (on relevance, confidence, training, and performance) were asked concerning screening for chlamydia among sexually active women aged 25 years or younger. Knowledge about the human papillomavirus (HPV) was assessed using four true/false questions (all at T3 only): 1) HPV causes common, plantar, and anogenital warts; 2) certain types of HPV are associated with cervical cancer, dysplasia, and intraepithelial neoplasia; 3) most people with HPV infections develop symptoms; 4) more than half of all sexually active Americans are likely to have HPV infection in their lifetime.
Independent variables tested for association with the primary professional outcomes were personal characteristics (sex, intended specialty, race or ethnicity, marital status, political self-characterization, history of sexual abuse, strength of religious identity, perceived risk of acquiring HIV, their personal physicians' stress on prevention, and their intention to practice safe sex) and professional characteristics (intended specialty, their opinion on whether physicians have a responsibility to promote prevention, and whether they thought it was important for physicians to discuss safe sex with patients). The variable “any history of sexual abuse” combined information from family (parents/siblings/grandparents) (T1 and T2) and personal history (T1, T2, T3) responses. Variables were assessed at all three time periods unless otherwise noted.
Student clustering within schools and longitudinal data collection resulted in a lack of independence between observations; we adjusted variance estimates to account for this dependence using SUDAAN (Research Triangle Institute, Research Triangle Park, NC),14 treating each school as a cluster and each student's multiple responses as subclusters.
The univariable associations between our outcomes (trying to practice safe sex, relevance to intended clinical practice of discussions of safe sex and screening for chlamydia, and self-reported frequency of discussion/screening) and independent predictor variables were tested using the χ2 test. To control for possible confounding effects on counseling relevance and frequency results, these associations were adjusted for gender, race/ethnicity, and intended specialty using logistic regression models, and adjusted odds ratios (ORs) were reported. Further, we explored the possible impact of multipoint responders differing from less frequent responders in the associations of interest; associations did not differ significantly for one-point, two-point, or three-point responders. Because of the number of associations being tested, we limited our discussion of significant results to those with P<.01.
Among these U.S. medical students, 75% were single (this includes the 1% who were widowed, divorced, or separated but does not include those reporting themselves as a member of an unmarried couple). Forty percent of the students who were not part of a couple reported that trying to practice safe sex was not applicable to them (Table 1); we classified these students as being sexually inactive. Sexual inactivity was more common among female singles, those feeling they were at lower risk for acquiring HIV, those who had not been tested for HIV within the previous year, and those who did not report excessive drinking in the previous month. Among the comparisons of ethnic groups, Asians were significantly less likely to be sexually active than any other group (data not shown).
Among sexually active singles, three quarters reported safe sex to be a high priority. Highly prioritizing safe sex was reported by significantly more women than men and more commonly in the earlier years of medical schools. The decline over time in the priority of safe sex was consistent for both men and women and for all ethnic groups (data not shown). Other characteristics significantly associated with highly prioritizing safe sex were belief that one was at lower risk for acquiring HIV, having had personal physicians who emphasized prevention more than treating disease, and not drinking excessively (marginally associated, P=.02). Gender-adjusted ethnic differences were not significant, nor were strength of religious identity (neither shown) or recency of an HIV test.
Rates of HIV testing were low among all singles. Eighteen percent of sexually active singles reported a recent HIV test, and 52% had never had a test; these rates were 11% and 69%, respectively, among singles who were not sexually active (Pactive vs inactive=.0002).
The remaining results pertain to professionally related outcomes among all students regardless of relationship status. Table 2 provides students' responses, from freshman orientation to senior year, on perceived relevance of safe-sex counseling to their intended practice (by specialty area), the amount of training they had received in this counseling, and their confidence in and self-reported frequency of providing this counseling to patients. Overall, just more than half (55%) of students believed it would be highly relevant to counsel patients about safe sex, with student opinion varying significantly by intended specialty and time point. The pattern was that the highest relevance was reported by those going into obstetrics–gynecology, somewhat less for primary care, and least for nonprimary care. During senior year, perceived high relevance became less common for all specialties, with larger declines from the previous time point being observed among those not intending to go into obstetrics/gynecology. Forty-one percent of seniors reported extensive training, up from 26% at orientation to wards. More than half were highly confident about discussing safe sex with their patients regardless of time point. Women reported more confidence than men (61% of women were highly confident versus 49% of men, P<.001, data not shown). When queried at orientation to wards, 85% of students thought it highly important for physicians to talk to patients about safe sex (data not shown); in senior year, 17% usually-always and 65% sometimes discussed safe sex with patients.
Results for screening for chlamydia (for sexually active female patients, age 25 or younger), queried only in senior year, were similar: it was perceived by 32% as being highly relevant, 53% were highly confident in talking to patients about it, 35% reported extensive training on the topic, and 21% performed such screening usually-always (data not shown).
Odds of reporting high relevancy for safe-sex counseling were significantly higher among women, African Americans or Hispanics (versus all others), sexually active singles who made safe sex a high priority, and those with stronger religious identity (Table 3). Holding an opinion on their personal risk of HIV infection, either agreeing or disagreeing that they were at high risk, was associated with twice the odds of frequent counseling compared with replying “neither agree nor disagree” (ORagree 2.5, 95% confidence interval [CI] 1.3–4.6; ORdisagree 2.0, 95% CI 1.4–2.8). Characteristics not found to be associated with this outcome were marital status, political characterization, having had personal physicians who emphasized prevention, and any family or personal history of sexual abuse (latter not shown). Note that these and the following safe-sex counseling results have been controlled for possible confounding by intended specialty, gender, and ethnicity.
Higher odds of frequent safe-sex counseling of general medicine patients were associated with being African American or Hispanic (versus all others), highly prioritizing safe-sex practices (among sexually active singles but not married/coupled students), having a very strong religious identity, and having had personal physicians who emphasized prevention. Gender, marital status, political characterization, perceived HIV risk, and any family or personal history of sexual abuse (latter not shown) were not significant predictors of counseling frequency.
Senior students intending to specialize in obstetrics/gynecology had 11 times the odds of reporting that safe-sex counseling would be highly relevant to their future practice compared with those interested in nonprimary care specialties, and primary care had four times the odds of the nonprimary care specialties (Table 4). Obstetrics/gynecology also had significantly higher odds of perceived counseling relevance compared with primary care (OR 4.8, 95% CI 3.1–7.3). Similarly, frequent counseling was least likely among nonprimary specialties; however, obstetrics/gynecology and primary care specialties did not differ significantly on frequency (OR 1.8, 95% CI 0.9–3.6).
Holding the opinion that physician counseling on safe sex is highly important (queried at ward orientation) and agreeing that physicians have a responsibility to promote prevention are both associated with greater odds of reporting high relevance and frequent counseling.
Characteristics significantly associated with chlamydia screening (for sexually active female patients, age 25 or younger), were similar to those for counseling on safe sex, with three notable exceptions: 1) having a physician who emphasized prevention was not associated with either perceived relevance or frequency of screening, 2) women were more likely than men to frequently screen for chlamydia (OR 1.8, 95% CI 1.3–2.4), and3) African Americans, but not Hispanics, were more likely to screen than other ethnic groups (OR 2.1, 95% CI 1.2–3.6). [data on chlamydia not shown].
Senior medical students were asked four true/false questions that tested their knowledge of HPV: 1) “HPV causes common, plantar, and anogenital warts” was answered correctly by 90% of students; 2) “Certain types of HPV are associated with cervical cancer, dysplasia, and intraepithelial neoplasia” by 99%; 3) “Most people with HPV infections develop symptoms” by 92% (the only false item); and 4) “More than half of all sexually active Americans are likely to have HPV infection in their lifetime” was answered correctly by 86%. Seventy-three percent answered all four questions correctly. Human papillomavirus knowledge was not significantly associated with the characteristics gender, intended specialty (ob-gyn, primary care, nonprimary care), personally trying to practice safe-sex habits, and (for women) Pap test recency, nor with perceived relevance or frequency of safe-sex counseling.
We found that personally practicing safe-sex habits is a high priority for sexually active, single medical students, especially for women, African Americans, and those in the earlier part of their undergraduate medical education. Those who made safe-sex practices a higher priority considered themselves to be at lower risk of HIV infection. Medical students' personal safe-sex practices matter for their own well-being, but we also found that reporting safer personal sex practices oneself (and having a physician who emphasized prevention) positively correlated with patient safe-sex counseling practices. Interestingly, students may not appreciate this correlation: one small (N=315) study of second-year medical students showed that those with more risk-taking attitudes and behaviors (including sexual behaviors) were significantly less likely than their colleagues to believe their attitudes would affect their clinical practice.15 Our data provide yet another piece of evidence that it is worth improving the personal health practices of medical students because it makes them more likely to counsel patients on related topics. Through increased patient counseling, improved medical student health could provide an efficient way to improve the health of entire patient populations.
Students' likelihood of feeling extensively trained about counseling on safe sex increased to 41% in senior year from 26% at orientation to wards. However, perceived relevance declined between junior and senior years regardless of intended specialty; only 17% of seniors usually-always discussed safe sex with typical general medicine patients (18% never did so), and only about half of either juniors or seniors were highly confident about discussing safe sex with their patients. These findings underscore the need for identifying and disseminating evidence-based educational interventions to train medical students in how to help their patients reduce their risk of STDs and to reduce their risk of acquiring or transmitting HIV infection.
Several variables were significantly associated with the perceived relevance of counseling and with counseling frequency regardless of intended specialty. Women and African Americans, the demographically same medical students who were most interested in personally practicing safe-sex habits, were also most likely to perceive that counseling patients about safe sex would be highly relevant to their intended practices and to currently usually-always counsel typical general medicine patients about safe sex. Female medical students may be more comfortable talking with patients about how they can lower their risk of acquiring STDs or HIV/AIDS. They may also be more likely to perceive counseling about reduction of sexual risk as a routine part of clinical practice because of their own reproductive health care experiences; studies have shown that female physicians are more likely to talk with patients about sexual risk factors.7 Politically liberal students were more likely than very politically conservative students to think sex counseling relevant to their future practices (and were marginally more likely to currently usually counsel).
Unsurprisingly, students intending to go into obstetrics/gynecology (regardless of gender) reported the highest perceived counseling relevance and frequency; those going into primary care reported intermediate levels. Earlier surveys have examined the sexual risk assessment and counseling practices of obstetrician–gynecologists and found them to compare favorably with those of practicing primary care physicians.3,7 We also found that those students believing it highly important to promote prevention generally with their patients (and to counsel about safe sex specifically) were much more likely to think such counseling would be highly relevant to their future practice and currently to do so more frequently with their patients.
Our study indicates that U.S. medical students have good basic HPV-related knowledge; most, regardless of their gender, intended specialty, personal safe-sex practices, or (for women) time since most recent Pap test knew that HPV is associated with cervical cancer and anogenital warts. In a national survey of U.S. physicians, the majority reported that they were aware that genital HPV infection is common, asymptomatic, and that it increases risk for cervical dysplasia, cancer, and anogenital warts, but fewer than half knew that most genital HPV infections clear without medical intervention (a question not asked in the present study).15 Future surveys of medical students should explore more complex questions as HPV testing and HPV vaccines become more popular.
There are three limitations to our study. First, our sample was representative of medical schools nationally, but it was not randomly selected. Second, although students' attitudes were self-reported in an anonymous pencil-and-paper survey, a bias may have encouraged students to provide answers they thought would be more socially acceptable. This is particularly true for a survey regarding sex, and these biases may vary by respondent type (ethnicity, gender, religiosity, interest in prevention). Finally, the total response rate for all three time periods was 80.3%.
Earlier surveys have shown that many practicing physicians do not routinely counsel their patients about HIV/STD risk reduction.3 Possible reasons include discomfort with the topic, the physicians' lack of confidence in sexual risk assessment and counseling skills, and time constraints. Physicians are in a unique position to help slow the spread of HIV and STD epidemics, but many are not comfortable discussing risky behaviors and sexual practices with their patients.2,3 There is a compelling public health need for better undergraduate medical education and continuing professional education on how physicians can effectively counsel patients about sexual risk reduction.16 We found that about half of U.S. medical students studied believed that counseling their patients about safe sex would not be highly relevant, and this belief was highly influenced by personal characteristics that are marginally relevant to one's patient mix. These findings should be considered by those trying to interest a new generation of physicians in helping patients have safe-sex practices.
1. Phillips KA, Bayer R, Chen JL. New Centers for Disease Control and Prevention's guidelines on HIV counseling and testing for the general population and pregnant women. J Acquir Immune Defic Syndr 2003;32:182–91.
2. Coates TJ. HIV prevention in clinical practice. AIDS Clin Care 1999;11:1–2, 8.
3. Kerr SH, Valdiserri RO, Loft J, Bresolin L, Holtgrave D, Moore M, et al. Primary care physicians and their HIV prevention practices. AIDS Patient Care STDS 1996;10:227–35.
4. Hansen L, Barnett J, Wong T, Spencer D, Rekart M. STD and HIV counseling practices of British Columbia primary care physicians. AIDS Patient Care STDS 2005;19:40–8.
5. Irwin K. Doctors overlook sexual histories too often. AIDS Alert 2001;16:113–4.
6. Maheux B, Haley N, Rivard M, Gervais A. Do physicians assess lifestyle health risks during general medical examinations? A survey of general practitioners and obstetrician-gynecologists in Quebec. CMAJ 1999;160:1830–4.
7. Boekeloo BO, Marx ES, Kral AH, Coughlin SC, Bowman M, Rabin DL. Frequency and thoroughness of STD/HIV risk assessment by physicians in a high risk metropolitan area. Am J Public Health 1991;81:1645–8.
8. Boekeloo BO, Rabin DL, Coughlin SS, Labbok MH, Johnson JC. Knowledge, attitudes, and practices of obstetrician-gynecologists regarding the prevention of human immunodeficiency virus. Obstet Gynecol 1993;81:131–6.
9. American Association of Medical Colleges. Matriculant age at anticipated matriculation, 1992-2001. Available at: http://www.aamc.org/data/facts/archive/famg112001a.htm
. Retrieved July 27, 2007.
10. American Association of Medical Colleges. AAMC data warehouse: student records system as of December 12, 2000, reflecting those actively enrolled on October 31, 2000. Available at: http://www.aamc.org/data/facts/archive/famg82001.htm
. Retrieved July 27, 2007.
11. Stoll M. The best medical schools, 2000 edition. New York (NY): Princeton Review Publishing; 1999.
12. National Institutes of Health. Support to U.S. medical schools, Fiscal year 2000. Available at: http://report.nih.gov/award/Rank/medschrank00.txt
. Retrieved July 27, 2007.
13. Dillman DA. Mail and Internet surveys. The tailored design method. 2nd ed. New York (NY): John Wiley & Sons, Inc.; 2000.
14. Shah BV, Barnewell BG, Bieler GS. SUDAAN User's Manual. 6.4 ed. Research Triangle Park (NC): Research Triangle Institute; 1995.
15. Montaño DE, Kasprzyk D, Carlin L, Freeman C. Results from: HPV provider survey: knowledge, attitudes, and practices about genital HPV infection and related conditions. Available at: http://www.cdc.gov/std/HPV/HPVProviderSurveyExecSum.pdf
. Retrieved July 27, 2007.
16. Rabin DL. Adapting an effective primary care provider STD/HIV prevention training programme. AIDS Care 1998;10(suppl 1):S75–82.