HIV Preexposure Prophylaxis Training in Family Medicine Residencies: A National Survey

In 2012, the Food and Drug Administration approved the use of emtricitabine/tenofovir disoproxil fumarate (Truvada) as preexposure prophylaxis (PrEP) for HIV-negative individuals. Since that time, emtricitabine/tenofovir alefenamide (Descovy) has been approved for HIV-negative men and transgender women, and other therapies are under investigation as well. PrEP reduces the risk of contracting HIV by 44%-88% in at-risk populations,1-5 with risk reductions of 92%-100% among individuals most adherent to PrEP.6-8 However, dissemination and uptake of this HIV-reduction opportunity requires clinicians to be trained in its use and willing to prescribe PrEP. Primary care providers (PCPs) should be equally, if not more, knowledgeable of the utility of PrEP and practiced in its delivery within primary care settings, as many HIV specialists may not care for a large number of patients who are HIV negative. Furthermore, patients in rural areas often lack access to this preventative treatment due to barriers that include large distances to nearest PrEP provider, inadequate funding to support PrEP distribution in those areas, and insufficient telehealth resources within those communities.9-10

Access to PrEP requires health care providers to be aware of its existence, competent to discuss and prescribe it for appropriate patient populations, or comfortable to refer patients elsewhere. Many of the studies on use of PrEP have evaluated potential barriers for providers. One study identified that the HIV providers who have the knowledge and skills necessary to prescribe PrEP may see few HIV-negative individuals.11 In contrast, PCPs may encounter high-risk HIV-negative patients, but may be unaware of PrEP or uncomfortable prescribing it. Other study participants have also referred to concerns about effectiveness of PrEP in real-world settings, potential risk disinhibition, medication toxicity, and antiretroviral drug-resistance induction.12-19 Providers often do not identify high-risk behavior for various reasons, including stigma and communication barriers.19 Interventions to improve access to PrEP are often developed by scientists, and implemented by health care providers, who are not part of the marginalized groups they are intended to serve.20 Some studies have referenced a general lack of knowledge about PrEP among at-risk people and by some medical providers, indicating that they do not feel sufficiently trained to prescribe PrEP. Others express the need for initiatives to better educate providers on the proper indications for PrEP. Yet, there has been no study directly assessing the acceptance of and barriers to PrEP utilization among primary care residency programs as it varies between rural and nonrural areas. Therefore, we sought to assess the factors associated with PrEP prescribing in family medicine training programs and their associations on perceived PrEP competencies.

Our primary hypothesis was that nonrural family medicine residency programs would be more likely to teach PrEP versus those in more rural settings. Our second hypothesis was that programs with significant training would have more prescribing within their practices. Our third hypothesis was that programs with significant training graduate residents with greater PrEP competency.

The survey questions were part of a larger omnibus survey conducted by the Council of Academic Family Medicine Educational Research Alliance (CERA). The CERA Steering Committee evaluated questions for consistency with the overall subproject aim, readability, and existing evidence of reliability and validity. Pretesting was done on family medicine educators who were not part of the target population. Questions were modified following pretesting for flow, timing, and readability. The American Academy of Family Physicians Institutional Review Board approved the project in January 2018. Data were collected from January to February 2018.

The sampling frame for the survey was all Accreditation Council for Graduate Medical Education-accredited US family medicine residency program directors as identified by the Association of Family Medicine Residency Directors. Email invitations to participate were delivered via the online program SurveyMonkey. Six follow-up emails to encourage nonrespondents to participate were sent after the initial email invitation. There were 549 program directors at the time of the survey; 13 had previously opted out of CERA surveys.

Items included in CERA surveys are developed by workgroups who submit items within their area of inquiry for review by CERA, internal piloting, and revision under the mentorship of senior researchers. For the program director surveys, all measures were self-reported estimates provided by the program director from each residency surveyed. While no single person in a residency can speak to the behaviors and performance represented by the developed items, program directors are the stakeholder group most representative of the academic and clinical issues facing family medicine graduate medical trainees.

Initially, we planned to define programs as above or below 1 million in population. This led to a large imbalance in subgroups, as only 47 of 274 programs (17.2%) were in areas >1 million. We then decided to use the US Census Bureau definition of urbanity. It uses ≥50,000 as its definition of an “urbanized area.”21 The CERA omnibus survey had demarcations of 30,000 and 75,000, so we decided to test our hypothesis using both 30,000 and 75,000 as cutoffs. One analysis compared communities >30,000 to those ≤30,000; a second, separate analysis compared communities >75,000 to those ≤75,000.

The level of PrEP training was used as both an independent and a dependent outcome in our analyses. We compared those with either no formal training in PrEP or some formal training (eg, as part of a lecture/didactic) with those who identified significant training, (eg, a lecture/didactic with routine reinforcement in clinical settings). For PrEP prescribing, we compared answers to the question “percent of patients who are appropriate for PrEP and being cared for by residents are receiving PrEP,” contrasting those with ≤50% prescribed PrEP appropriately by residents to those with >50% of patients. A final outcome used to explore our hypotheses included a measure of programs’ self-estimation of graduating residents’ level of competency in prescribing PrEP. Program directors were asked to estimate residency graduates’ competency across the following categories: “None (will refer to outside provider)”; “Basic (able to screen for risk factors and overall PrEP need)”; “Independent (able to screen, order appropriate labs, and prescribe PrEP)”; or “Advanced (able to screen, order appropriate labs, prescribe PrEP, and to effectively teach others how to properly prescribe PrEP).” We compared programs reporting “none” or “basic” with programs reporting “independent” or “advanced.”

We used two items to assess the largest and second-largest barriers to PrEP prescribing in residencies. Respondents were asked to identify among “Lack of resident knowledge or training,” “Lack of faculty expertise,” “Not enough high-risk patients,” “Inadequate screening by providers,” “Insufficient interest or willingness of patients,” “Provider resistance to prescribing,” or “Cost of medication.” We calculated the proportion of respondents who identified each potential barrier as either the biggest or second-biggest barriers to produce an overall ranking of barriers to PrEP prescribing.

Fisher’s exact testing and expected values were calculated for assessing the association with community size, as no rural communities reported significant PrEP training for their residents. Since the numerator was 0 for that subgroup, an odds ratio calculation was not possible. For other comparison, we used logistic regression to compute odds ratios to estimate the associations between dependent and independent measures. We employed a level of statistical significance set at P=.05, recognizing that tests of statistical significance are approximations that serve as aids to interpretation and inference. We used Stata software (Stata Corp, College Station, TX) for the analysis.

Of 549 program directors identified, 13 had previously opted out of CERA surveys, and 14 emails could not be delivered. The final sample size was therefore 522. The overall response rate for the larger omnibus survey was 57.1% (298/522). An additional 22 programs did not complete PrEP items and were excluded from analyses leaving a final sample of 276/522 for a response rate of 52.9% for the analyzed PrEP items. Table 1 illustrates the demographics of respondents’ residency programs.

A total of 15.8% of respondents reported no PrEP training in their residency and no provision of PrEP clinically for patients. Another 10.7% of programs reported no PrEP training, but did have clinical PrEP prescribing for their patients. The majority of programs (52.0%) reported some PrEP training in their residency with 13.8% of programs reporting significant PrEP training.

For our first hypothesis, we found that no programs training in rural communities reported significant PrEP training for their residents. Table 2 illustrates the distribution of training by community size. For our second and third hypotheses, highly significant associations were noted for the association between significant resident training in PrEP and reports of the majority of PrEP-appropriate patients receiving PrEP (OR=7.3; P<.001) and residents achieving independent or advanced competencies in PrEP (OR=18.3; P<.001). Table 3 illustrates the associations between reports of significant residency PrEP training and outcome measures.

Table 4 illustrates the combined proportion of barriers to PrEP prescribing. Lack of faculty expertise (23.2%) and not enough high-risk patients (20.5%) were the most commonly endorsed barriers to PrEP prescribing among respondents.

The data presented illustrate a concerning proportion of training programs that include no PrEP training (26.5%) and a significant number of programs that provide no clinical access to PrEP for their patients (15.8%). To our knowledge, this study is the first to assess the utilization of PrEP among family medicine residency programs in rural and nonrural settings. The primary findings of our study included that residency programs that reported significant training were seven times more likely to have the majority of their PrEP-appropriate patients receiving PrEP. We also found that programs that reported significant training were 18 times more likely to graduate residents described as competent at the independent or advanced level in managing PrEP-appropriate patients. Lastly, we found that no rural family medicine programs (<30,000 population) reported significant PrEP training.

HIV PrEP lies firmly within the realm of primary care and has been given a Grade A recommendation by the United States Preventive Services Task Force.22 Therefore, we sought to assess if there were geographic disparities among family medicine residencies regarding their acceptance of and barriers to PrEP utilization. Our survey respondents had a large distribution across geographic regions and community size, with the greatest percentage coming from community based, university-affiliated residency programs. These community-level gaps in residency PrEP training risk producing family medicine physicians who are unprepared to offer the full range of HIV prevention options to appropriate patients. This may ultimately contribute to avoidable HIV infections.

While we did not have a specific hypothesis about the most likely barrier to PrEP prescribing, we found that faculty expertise and having enough high-risk patients were the most likely to be the top choice of respondents. These data fit with existing literature describing how a lack of provider knowledge about PrEP leads to discomfort prescribing it.23,24 Current literature also shows regional disparities in PrEP prescribing, which is relatively low in the South/Southeastern United States.25,26 HIV disproportionately affects those living in the Southeast United States, yet this is also a region where PrEP is prescribed with relatively low frequency.26 While the disparity between regions with high HIV prevalence and relatively low PrEP prescribing is multifactorial, it begs the question: are providers truly taking sexual histories that assess for HIV risk behaviors? If there is a lack of faculty expertise regarding PrEP, patients are less likely to be accurately screened for PrEP-appropriateness. This could, in turn, contribute to the misperception of lacking enough high-risk patients. Each of these barriers potentially drives lower levels of training for family medicine residents, perpetuating an unpreparedness to prescribe PrEP.

The limitations of this study include the limits of self-reported data and biases related to social desirability. For example, program directors identifying prescribing rates to eligible patients may overestimate this variable if their clinic underscreens for appropriate patients. While program directors are the stakeholder group best able to respond to the breadth of items included in the survey, there are limitations to the ability of respondents to provide valid estimates of resident competencies and clinical delivery of PrEP services. Given the impact of social desirability, we assume that our results are conservative estimates of the educational challenges related to PrEP training illustrated by our data. While the survey had a reasonable response rate, it is always possible that nonresponders would have been systematically different from responders resulting in bias. While we would like to make inferences for primary care, it is very possible and even likely that family medicine training programs may be different from other primary care training professions. While we believe that our community size comparisons are meaningful, the limits of the available survey response options did not allow us to map standard community size measures for rural versus urban comparisons. Finally, cross-sectional data can only suggest associations, and we must be cautious in making causal assumptions from our findings.

Potential next steps include an improved delineation of best practices for training future primary care providers in the assessment of HIV risk and the use of PrEP. In addition to developing and augmenting existing training curricula, continuing medical education venues will need to adopt similar training priorities in order to enhance spread and uptake among already-practicing providers. Providers should be encouraged to work as part of interprofessional teams, including care navigators, pharmacists, and other health care providers as resources. At-risk populations should be empowered to ask for PrEP, and tools should be developed along with these populations for ways to bridge communication gaps and reduce stigma. Additionally, in clinical settings with lower-risk populations, residents could be encouraged to participate in rotations with higher-risk populations to increase their exposure to appropriate PrEP prescribing.

In conclusion, we found that rural family medicine residency programs lack PrEP training and programs that provide significant PrEP training are more likely to prescribe it to appropriate patients as well as graduate residents who are prepared to do the same. HIV is a national epidemic, and requires a national response. The COVID-19 pandemic requires us to rededicate ourselves to prevention of disease transmission, recognizing that prevention is the most powerful tool in controlling epidemics. PrEP has the potential to impact the incidence of new HIV infections significantly, which disproportionately affect young MSM of color and transgender women of color. As such, PrEP prescribing is a tool of social justice. In partnership with patient populations they are intended to serve, family medicine residencies must commit themselves to improving PrEP uptake and adherence. Family physicians are at the frontlines of preventative health throughout the country and, given the appropriate training, have the power to curb rising rates of HIV infections among marginalized and vulnerable populations, and expand the reach of PrEP overall.