Acceptability and Effectiveness of a One-Hour Healthcare Provider Intervention Integrating HIV Pre-Exposure Prophylaxis and Cultural Competence Training

Abstract

Enhanced provider training could improve PrEP access and equity. We conducted a pilot randomized controlled trial comparing (a) a one-hour, group-based provider intervention integrating PrEP and Cultural Competence (PCC) training with (b) a standard HIV continuing medical education session (n = 56). PCC participants favorably rated the intervention and reported increased PrEP knowledge. The PCC intervention increased their confidence performing PrEP-related clinical activities and intention to prescribe PrEP. The percentage of participants discussing PrEP with patients increased marginally in both study conditions. The percentage of participants who prescribed PrEP and self-rated cultural competence did not change in either study condition.

Introduction

Although HIV pre-exposure prophylaxis (PrEP) has been federally approved in the US for over a decade (1) and is highly effective when taken as prescribed (2), only a fraction of people who are indicated for PrEP are accessing it (3). Healthcare providers can play a pivotal role in addressing low and inequitable PrEP access. Recent estimates from a national prescription database suggest that an increasing number of providers are prescribing PrEP in the US: 65,822 providers prescribed for 279,054 patients in 2019 compared with 9,621 providers for 22,278 patients in 2014 (4). However, deficits in providers’ PrEP knowledge, communication, and prescription persist and may be particularly consequential for communities most likely to benefit from PrEP (5, 6). In the current study, we examined the acceptability and preliminary effectiveness of a brief provider intervention that integrated PrEP and cultural competence training.

Stark disparities in HIV incidence and PrEP uptake that disproportionately affect Black and Latino/x/a people—particularly Black and Latino sexual minority men—suggest that it is essential to integrate cultural competence within provider training related to PrEP. Cultural competence refers to “the ability of systems to provide care to patients with diverse values, beliefs, and behaviors, including tailoring delivery to meet patients’ social, cultural, and linguistic needs” (7, p. V). A key component of cultural competence is clinical cultural competence, which manifests in patient-provider interactions (7). Past research with Black sexual minority men has found that culturally insensitive encounters with providers can discourage PrEP service engagement (6) and that culturally-tailored service provision can increase PrEP uptake (8). Additionally, some PrEP prescribers have explicitly called for additional cultural competence training to support PrEP service delivery, suggesting that such training could facilitate conversations about sexual health, particularly with sexual minority patients (9).

The current study extends previous work by evaluating a one-hour, group-based intervention for providers that combined PrEP and cultural competence training. Prior evidence suggests that brief, single-session educational interventions focused on PrEP are acceptable to providers and can have a favorable impact on providers’ PrEP knowledge, confidence enacting PrEP-related clinical activities, discussion of PrEP with patients, and PrEP prescription practices (10–12). Likewise, previous research has suggested that formal training on cultural competence in the form of standalone workshops can enhance providers’ cultural competence, at least in the short term (13). The separate successes of trainings related to PrEP and cultural competence offer promise that an intervention combining the two could effectively promote culturally competent PrEP service delivery. An integrated approach that tailors content on cultural competence to the context of PrEP service delivery—for example, incorporating evidence of biases in the clinical evaluation of PrEP candidates and strategies to combat such biases—could further facilitate providers’ application of cultural competence training to their provision of PrEP services.

We conducted a pilot randomized controlled trial evaluating the acceptability and perceived impact of a PrEP and Cultural Competence (PCC) intervention as well as its effectiveness compared with a standard HIV continuing medical education session. We evaluated several effectiveness outcomes, including confidence performing clinical activities associated with PrEP provision, intention to prescribe PrEP, discussion of PrEP with patients, PrEP prescription, and cultural competence. We hypothesized that the PCC intervention would be acceptable, would be perceived as increasing PrEP knowledge and intention, and would have a favorable impact on all effectiveness outcomes compared with the standard HIV CME session.

Methods

Participants/Procedure

We partnered with the New England AIDS Education and Training Center and the Yale CME Office to embed evaluation of the PCC intervention (described below) into an existing five-hour CME conference that took place in New Haven, CT, in 2019. The PCC intervention was tested during the first of five one-hour sessions held during the conference. The CME coordinator emailed an invitation to participate in the study to all providers who had registered for the conference. In addition, providers who had not already enrolled in the study were invited to participate during on-site registration the morning of the conference. Conference attendees who chose not to enroll in the study attended a standard HIV CME session as scheduled in the conference program.

Enrolled participants completed a survey at three time points: pre-intervention, immediately post-intervention, and three months post-intervention. All surveys were completed online except pre-intervention surveys completed by providers who enrolled in the study the morning of the conference, who completed a paper-and-pencil version. During the pre-intervention survey, all participants were randomly assigned to attend the PCC intervention (experimental condition) or the standard HIV CME session (control condition), which were held simultaneously in different rooms within the conference venue. The standard HIV CME session was a slides-based lecture that mentioned PrEP briefly among other HIV-related content. Participants were not informed in advance of which condition they were in. Rather, they were told they were in “Group A” or “Group B” during the pre-intervention survey and then, on the day of the intervention, directed to attend the session in a given room based on their earlier group assignment (A or B).

Participants were compensated with Amazon gift cards in the amount of $50 for the immediate post-intervention survey and $75 for the three-month post-intervention survey. Additionally, they received CME credit for their attendance of the CME session, whether assigned to the PCC intervention or standard HIV CME session. All study procedures were approved by the Yale University Human Subjects Committee (IRB #HSC-1308012487) and the George Washington University Office of Human Research (IRB #061636).

Intervention

The PCC intervention was designed as a one-hour, single-session, group-based training consisting of a slides-based lecture and role-play demonstration. It was co-facilitated by a psychologist (SKC) and infectious diseases specialist (DSK) and delivered in person as a CME session. The format and content of the intervention were decided based on review of relevant literature and assessment of providers’ objective and subjective training needs related to PrEP and cultural competence. Our needs assessment (published previously and not presented here) included interviews with early-adopting PrEP providers, focus groups with Black sexual minority men, focus groups with community healthcare providers, and an online survey of providers. Content covered during the lecture included a case study of a PrEP patient, explanation of how PrEP works in the body, review of PrEP efficacy trials and open-label studies, and discussion of both the benefits of PrEP and commonly expressed concerns (e.g., side effects, risk compensation). Additionally, the co-facilitators discussed patient eligibility for PrEP, including US federal guidelines and their limitations, and presented a clinical protocol for PrEP initiation and monitoring. The lecture also included content related to cultural competence in the context of PrEP care, including: historical and present-day HIV disparities, PrEP access disparities, explicit and implicit bias, evidence for bias impacting patient-provider interactions, evidence for potential provider biases related to PrEP, and guidance for preventing/addressing bias when providing PrEP services. The role play demonstration focused on how to communicate about sex with patients in a culturally sensitive way (e.g., avoiding assumptions about sexual behavior and sexual orientation) and featured examples of poor communication (prompting audience identification of problems) and good communication (prompting audience identification of improvements). Sample content from the intervention is available in Appendix A.

Survey Measures

Participants reported background information, including sociodemographic and professional characteristics; prior PrEP awareness (Ever vs. Never Heard of PrEP); and prior PrEP discussion and prescription (Ever with 1+ Patient[s] vs. Never) during the pre-intervention survey.

Participants evaluated several aspects of the PCC intervention related to acceptability during the immediate post-intervention survey. They rated statements that the session was well organized and interesting on a Likert scale ranging from (1) Strongly Disagree to (5) Strongly Agree. The first three response options (Strongly Disagree, Disagree, and Neither Agree nor Disagree) were coded as (0) Not Organized/Interesting, and the last two options (Agree and Strongly Agree) were coded as (1) Organized/Interesting. Participants also provided overall ratings of the training session and facilitators on a scale ranging from (1) Poor to (5) Excellent.

Participants reported perceived impact of the PCC intervention by rating whether the training increased their knowledge about PrEP and whether it increased their likelihood of prescribing PrEP during the immediate post-intervention survey on a Likert scale ranging from (1) Strongly Disagree to (5) Strongly Agree. The first three response options (Strongly Disagree, Disagree, and Neither Agree nor Disagree) were coded as (0) No Perceived Impact, and the last two options (Agree and Strongly Agree) were coded as (1) Perceived Impact. Participants also reported how the PrEP training had impacted their PrEP prescribing practices during the three-month post-intervention survey, with three response options indicating that the training decreased their likelihood of prescribing, neither increased nor decreased their likelihood of prescribing, or increased their likelihood of prescribing. The first two response options were recoded as (0) Did Not Increase, and the latter option was coded as (1) Increased.

Participants rated their confidence carrying out six clinical activities associated with providing PrEP on a scale ranging from (1) Not at All Confident to (5) Extremely Confident during all three surveys. The six PrEP-related clinical activities included: talking to a patient about their sexual health, screening a patient to determine whether they were an appropriate candidate for PrEP, answering a patient’s questions about PrEP, suggesting PrEP to a patient, counseling a patient about using other methods of protection, and effectively monitoring a patient on PrEP. We computed a mean score across the six items as an indicator of overall PrEP-related confidence (Cronbach’s α = .92 pre-, .93 immediately post-, and .93 three months post-intervention).

Participants rated their intention to prescribe PrEP, operationalized as their likelihood of prescribing PrEP for one or more patients in the next year, on a scale ranging from (1) Not at All Likely to (5) Extremely Likely during all three surveys. During the pre-intervention and three-month post-intervention surveys, participants reported whether they had discussed PrEP with one or more HIV-negative patients in the past three months (PrEP Discussed) and whether they had prescribed PrEP in the past three months (PrEP Prescribed). Participants who indicated that they did not discuss and/or prescribe PrEP during the follow-up period were asked to indicate the reason(s) why not based on a list of options that included an “Other” option with a prompt for specification in an open-response textbox.

Participants reported perceived cultural competence in response to the following question: “How would you rate your overall cultural competence when it comes to providing care for diverse populations?” Response options ranged from (1) Not at All Culturally Competent to (5) Extremely Culturally Competent.

Finally, immediately post-intervention, participants were asked to report which of the two sessions they attended. The focus and facilitators of each of the two sessions were specified in the response options to distinguish the sessions.

Analysis

We calculated frequencies and other descriptive statistics to characterize the sample and measures of interest. To test for differences between the two study conditions in attrition, background characteristics, acceptability, and perceived impact, we used independent samples t-tests and chi-square tests (or, when expected counts were less than five in one or more cells for a 2x2 analysis, Fisher’s two-sided exact tests). To examine the effectiveness of the PCC intervention, we assessed the effects of study condition (PCC intervention vs. standard HIV CME session), time, and study condition x time relative to our outcomes of interest. When the interaction effect between study condition and time was significant (p < .05) or marginally significant (p < .10) for a given outcome, we examined simple effects. We conducted mixed factorial ANCOVAs for continuous outcomes (PrEP-related confidence, intention to prescribe, and cultural competence) and binomial logistic regressions using generalized estimating equations for dichotomous outcomes (PrEP discussion and prescription). Time was a repeated measure that included three measurement points for continuous outcomes (pre-, immediately post-, and three months post-intervention) and two measurement points for dichotomous outcomes (pre- and three months post-intervention). When simple effects were significant for continuous outcomes, we conducted follow-up pairwise comparisons among the three different time points applying the Sidak adjustment for multiple comparisons. We adjusted all ANCOVA and regression models for any background variable that was significantly associated with study condition or 1+ outcome. Only prior PrEP awareness met this criterion. The sample was restricted to licensed prescribers for analyses examining effects on intention to prescribe and prescription.

Given the attrition that occurred between the immediate post-intervention survey (n = 44) and the three months post-intervention survey (n = 32), we repeated primary analyses imputing the data for the 12 missing participants to assess whether the same pattern of results emerged. We also repeated the ANCOVAs with the non-imputed data set limiting time to two points (pre- and immediately post-intervention) to assess whether the same pattern of results emerged.

Results

Sample Overview

Of the 56 participants who completed the pre-intervention survey, 44 (78.6%) attended the CME conference and completed the immediate post-intervention survey (16 in the PCC intervention condition and 28 in the standard HIV CME condition), and 32 (57.1%) also completed the three-month post-intervention survey (11 in the PCC intervention condition and 21 in the standard HIV CME condition).¹ We were unable to assess differences by study condition in attrition occurring between the pre-intervention survey and the two post-intervention surveys because study condition was determined by self-reported attendance during the immediate post-intervention survey; thus, pre-intervention survey completers who were not retained in the immediate post-intervention survey did not have a study condition assigned. Attrition occurring between the immediate post-intervention survey and the three-month post-intervention survey did not significantly differ by study condition (31.3% in the PCC intervention and 25.0% in the standard CME condition; Fisher’s exact test p = 0.732).

Background characteristics of the three survey samples, stratified by study condition, are summarized in Appendix B. The 44 participants who completed the immediate post-intervention survey ranged in age from 24–76 years (M [SD] = 52.7 [12.8]) and were predominantly non-Latino/x/a (97.7%), White (88.6%), heterosexual (95.3%), and women (81.8%). Most (70.5%) reported their clinical background to be nursing (RN, APRN, NP, or CNM), while 29.5% reported their background to be medicine (MD or DO). Prior to the study, 81.8% of participants who completed the immediate post-intervention survey had ever heard of PrEP, and 40.9% had ever discussed PrEP with a patient. Among those licensed to prescribe medication (n = 42 of 44), a minority (14.3%) had ever prescribed PrEP. There were no statistically significant differences in pre-intervention background characteristics between study conditions at either of the post-intervention survey time points.

Intervention Acceptability and Perceived Impact

All PCC participants (100%) and most standard HIV CME participants (82.1%) reported that their respective session was well organized, with no significant differences between the two conditions (Fisher’s exact test p = .141). Likewise, all PCC participants (100%) and most standard HIV CME participants (92.9%) reported that their respective session was interesting, with no significant differences between the two conditions (Fisher’s exact test p = .526). The mean overall rating of the PCC session (M = 4.8 out of 5.0, SD = 0.4) was significantly more favorable than that of the standard HIV CME session (M = 4.2 out of 5.0, SD = 0.9), t(41.7) = −2.9, p = .006. Likewise, the mean overall rating of the PCC facilitators (M = 4.8 out of 5.0, SD = 0.4) was more favorable than that of the standard HIV CME condition (M = 4.4 out of 5.0, SD = 0.8), t(41.7) = −2.6, p = .015.

A significantly greater percentage of participants in the PCC condition (100%) compared with the standard HIV CME condition (35.7%) perceived the session as having increased their PrEP knowledge when queried immediately post-intervention, X²(1, N = 44) = 17.4, p < .001. Similarly, among the subset of participants licensed to prescribe medication, a significantly greater proportion of participants in the PCC condition (86.7%) compared with the standard HIV CME condition (34.6%) perceived the session as having increased their likelihood of prescribing PrEP when queried immediately post-intervention, X²(1, N = 41) = 10.4, p = .001. When queried three months post-intervention, 63.6% of participants in the PCC condition and 57.9% of those in the standard HIV CME condition perceived the training to have increased their likelihood of prescribing PrEP (Fisher’s exact text p = 1.000).

Intervention Effectiveness

To compare the effects of our two study conditions, we conducted mixed factorial ANCOVAs for continuous outcomes (PrEP-related confidence, PrEP prescribing intention, and cultural competence) and binomial logistic regressions using generalized estimating equations for dichotomous outcomes (PrEP discussion and PrEP prescription). Results are summarized in Table I.

Table I.

Effectiveness of PrEP and Cultural Competence (PCC) Intervention vs. Standard HIV CME Session

	Overall PrEP-Related Confidence				Intention to Prescribe PrEPa				Cultural Competenceb
	Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post
Study Condition	M(SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)
PCC Intervention	2.8 (0.3)	3.6 (0.3)	3.3 (0.3)		2.2 (0.5)	2.9 (0.5)	2.6 (0.5)		3.5 (0.2)	3.5 (0.2)	3.6 (0.2)
Standard HIV CME	3.0 (0.2)	2.9 (0.2)	3.1 (0.2)		2.0 (0.3)	2.0 (0.4)	2.0 (0.3)		3.6 (0.2)	3.6 (0.2)	3.7 (0.1)
Effect	df	F	p	ηp 2	df	F	p	ηp 2	df	F	p	ηp 2
Study Condition	1,29	.0.4	.532	.01	1,27	1.1	.312	.04	1,29	<.01	.959	<.01
Time	2,58	2.2	.117	.07	2,54	3.0	.059	.10	2, 58	0.6	.534	.02
Study Condition × Time	2,58	4.4	.017	.13	2,54	4.1	.022	.13	2, 58	0.2	.819	.01
PCC Intervention (Simple)	2,28	6.1	.006	.31	2,26	5.0	.014	.28	-	-	-	-
Standard HIV CME (Simple)	2,28	1.0	.379	.07	2,26	.02	.985	<.01	-	-	-	-
	PrEP Discussed (Past 3 Mos.) b				PrEP Prescribed (Past 3 Mos.) a,b
	Pre	3 Mos. Post			Pre	3 Mos. Post
Study Condition	n (%)	n (%)			n (%)	n (%)
PCC Intervention	3 (27.3)	7 (63.6)			2 (18.2)	2 (18.2)
Standard HIV CME	6 (28.6)	10 (47.6)			2 (10.5)	3 (15.8)
Effect	aOR	95%CI	p		OR	95%CI	p
Study Condition (ref = Standard HIV CME)	1.1	0.2 – 5.5	.955		1.9	0.2 – 15.7	.557
Time (ref = pre-intervention)	2.3	0.9 – 6.3	.088		1.6	0.3 – 7.8	.564
Study Condition × Time	2.3	0.4 – 13.5	.345		0.6	0.1 – 3.1	.564
PCC Intervention (Simple)	-	-	-		-	-	-
Standard HIV CME (Simple)	-	-	-		-	-	-

Note. All statistical models were adjusted for prior PrEP awareness (pre-intervention) except for the PrEP Prescribed model (due to distribution of data). Means and standard errors reflect adjusted estimates.

^aRestricted to licensed prescribers (n =42 of 44 at three months post-intervention)

^bSimple effects are not shown because the interaction was non-significant.

PCC = PrEP and Cultural Competence CME=Continuing Medical Education

There was a significant interaction effect between study condition and time relative to mean PrEP-related confidence. Simple effects indicated increased confidence across time in the PCC condition but not the control condition. In the PCC condition, overall confidence was higher immediately post- vs. pre-intervention (p = .004); however, the overall gain had diminished somewhat three months later (p = .115).

Separate ANCOVAs examining the interaction between study condition and time relative to each of the six confidence items indicated that the PCC intervention increased participants’ confidence relative to most clinical activities (see Appendix C). Specifically, immediately following the intervention, PCC participants were more confident screening a patient to determine whether they were an appropriate candidate for PrEP (p = .033), answering a patient’s questions about PrEP (p < .001), suggesting PrEP to a patient (p = .003), counseling a patient about using other methods of protection (p = .019), and effectively monitoring a patient on PrEP (p = .013). The immediate gain in confidence remained statistically significant three months post-intervention for answering a patient’s questions about PrEP (p = .028) and marginally significant for suggesting PrEP to a patient (p = .061). The PCC intervention did not affect participants’ confidence talking to a patient about their sexual health history, which was already relatively high pre-intervention (M [SD]=4.0 [0.9] on a 1–5 scale).

There was a significant interaction effect between study condition and time relative to intention to prescribe. Simple effects indicated increased intention across time in the PCC condition but not the control condition. Prescribing intention was higher immediately post- vs. pre- intervention (p = .015), and this gain was sustained three months later (p = .044).

There was no significant study condition x time interaction effect on self-rated cultural competence or three-month PrEP discussion, suggesting that the PCC intervention did not uniquely affect these outcomes. However, across both conditions, the number of participants reporting PrEP discussion increased marginally, more than doubling (27.3% pre- vs. 63.6% post-intervention) in the PCC intervention condition. No impact of the PCC intervention on PrEP prescribing was observed, with only two (18.2%) of the PCC intervention participants prescribing PrEP during the three months preceding the intervention and the same number during the three months that followed.

A similar pattern of significant interaction effects emerged when we repeated the analyses with all participants who completed the immediate post-intervention survey (n = 44) after imputing data for the 12 participants who did not complete the three month post-intervention survey (Appendix D) except that: (1) the condition x time interaction effect was marginally significant (vs. significant) and (2) the increase in PrEP discussion across both study conditions was significant (vs. marginal). Likewise, a similar pattern of significant interaction effects emerged when we repeated the analyses with the non-imputed data set limiting time to two points (pre- and immediately post-intervention). Collectively, these replication analyses indicate that our results were robust with respect to the larger sample.

Appendix E summarizes the reasons PCC participants reported for not discussing PrEP (n = 4) and not prescribing PrEP (n = 9) three months post-intervention. The most common reasons for not discussing PrEP were: believing they were not seeing patients at high HIV risk (n = 3) and not having enough training to feel comfortable (n = 2). The most common reasons for not prescribing PrEP were perceiving PrEP prescribing to be outside of their purview (as suggested by their "other" open-text response: n = 6) and not seeing patients at high HIV risk (n = 4). Open-text responses suggesting PrEP to be outside of participants’ purview included statements about being a mental/behavioral health provider, referring patients to another type of provider, and working in a rehabilitation facility.

Discussion

This study evaluated a brief provider intervention that was novel in its integration of PrEP and cultural competence training. Findings partially supported our hypothesis, indicating that the intervention was acceptable to providers and prompted immediate gains in PrEP knowledge, confidence, and intention to prescribe PrEP, some of which persisted three months later. There was a marginal increase in PrEP discussion across both study conditions, with PrEP discussion more than doubling in the PCC condition. Contrary to our hypothesis, the PCC intervention did not impact PrEP prescription or cultural competence.

Our findings with respect to PrEP outcomes are encouraging. Consistent with the Information, Motivation, and Behavioral Skills (IMB) Model, the intervention significantly impacted the more proximal of the outcomes examined—information (PrEP knowledge), motivation (intention to prescribe PrEP), and behavioral skills (confidence performing PrEP-related activities), which have previously been associated with PrEP discussion and prescription among US providers (14). Although we did not find significant main effects of the PCC intervention on PrEP discussion and prescription, the observed malleability of information, motivation, and behavioral skills in response to this brief intervention bodes well for future interventions targeting these precursory factors.

The reasons that providers in the PCC intervention expressed for not discussing and prescribing PrEP post-intervention highlight content areas within the intervention that may require further elaboration. For example, although during the intervention we explicitly recommended that all patients be educated about PrEP as part of routine care, some providers reported that they did not discuss or prescribe PrEP because they did not perceive their patients to be at high HIV risk. Previous research has highlighted the misalignment in provider and patient viewpoints related to communicating about PrEP, with providers perceiving initiation of conversations about PrEP as optional (subject to their discretion) and patients perceiving it to be providers’ professional responsibility (5). In future interventions, it may be valuable to more fully explain to providers why educating all patients, irrespective of perceived risk, is important (e.g., doing so honors patient preferences, patients may not feel comfortable disclosing private information, the latest federal guidelines recommend discussing PrEP with all sexually active patients).

Regarding the lack of change in self-rated cultural competence, it is possible that covering PrEP and cultural competence in sufficient depth to impact outcomes associated with both was unrealistic in a one-hour time frame. We sought to confine our training to one hour in recognition of providers’ busy schedules, which often restrict their flexibility to pursue voluntary didactic activities. We envisioned this training being suitable for implementation in the context of CME sessions or staff meetings that providers already attended, thus avoiding the need for additional time or incentivization. Expanding the timeframe or targeting providers who already possess foundational knowledge and experience related to PrEP could allow for greater time allocated for content related to cultural competence.

Our analysis of cultural competence was also limited to providers’ perception of their own cultural competence and did not include objective indicators of cultural competence or patients’ perceptions thereof. Furthermore, providers self-rated their cultural competence broadly rather than relative to specific patient characteristics (e.g., race, sexual orientation, gender, age, religion). Although much of the cultural competence content of the PCC intervention was broad, several examples of disparities and bias were specific to race and sexual orientation. It is therefore possible that measures of cultural competence that were specific to these characteristics in particular would have detected change following the intervention.

Notably, our intervention focused on improving clinical cultural competence, targeting providers and their interactions with patients. Maximizing the downstream effects of clinical cultural competence initiatives on PrEP access and inequities may require complementary initiatives advancing organizational cultural competence (e.g., promoting racial minorities into healthcare leadership positions) and systemic cultural competence (e.g., offering interpreter services for non-English speakers) (7).

Our study has limitations. Participants were primarily primary care providers and psychiatry/mental health specialists. Findings may not generalize to other provider populations. Because many psychiatry/mental health specialists do not routinely prescribe non-psychiatric medications, the PCC intervention may have had less of an impact on the overall sample’s prescribing intentions and practices than it would have had with an audience composed entirely of primary care providers and infectious diseases specialists, for example. We recommend that future training with professionally diverse audiences affirms that prescribing PrEP is within the purview of any prescribing healthcare provider and highlights the value of providing PrEP in multiple health settings—including psychiatry, where clientele may be disproportionately affected by HIV (15)—to broaden access.

Attrition was fairly high, with only 78.6% of providers who completed the pre-intervention survey completing the immediate post-intervention survey and 57.1% completing the three-month post-intervention survey despite compensation and reminder emails. Other studies evaluating HIV educational initiatives with providers via post-intervention surveys have similarly documented low retention rates (11, 16). In the current study, replication of our primary findings, first with an imputed dataset and then restricting the dataset to pre- and immediate post-intervention time points only, suggest that results from the smaller, three-month post-intervention survey sample generally represented the larger sample. Replication of the intervention with a separate, larger sample implementing additional strategies to bolster retention may be beneficial.

Conference participants were informed in the study invitation that the CME session being evaluated was related to HIV prevention, which may have introduced selection bias. Additionally, all of the measures were self-reported and therefore subject to social desirability bias. Finally, although the randomized controlled design is a strength of the study, the standard HIV CME session was an imperfect control condition, varying from the experimental condition not only in its broader focus on HIV as intended, but also in several other ways (e.g., single facilitator vs. two facilitators, lecture only vs. lecture and role plays).

Results of this study are promising in their indication that a brief educational intervention with providers can be acceptable and meaningfully impact proximal PrEP outcomes, consistent with previous studies (10–12). Further research is needed to optimize such training and more comprehensively evaluate its impact on cultural competence, a critical target for addressing current inequities in PrEP access and ending the HIV epidemic.

Appendix A

Sample PrEP and Cultural Competence (PCC) Intervention Content²

Figure 1. Sample content covered during the lecture.

The lecture included a case study of a PrEP patient, explanation of how PrEP works in the body, review of PrEP efficacy trials and open-label studies, and discussion of both the benefits of PrEP as well as commonly expressed concerns (e.g., side effects, risk compensation). Additionally, the co-facilitators discussed patient eligibility for PrEP, including US federal guidelines and their limitations, and presented a clinical protocol for PrEP initiation and monitoring. The lecture also included content related to cultural competence in the context of PrEP care, including: historical and present-day HIV disparities, PrEP access disparities, explicit and implicit bias, evidence for bias impacting patient-provider interactions, evidence for potential provider biases related to PrEP, and guidance for preventing/addressing bias when providing PrEP services.

Figure 2. Notes and sample content from role play scripts.

The role play demonstration focused on communicating about sex with patients in a culturally sensitive way (e.g., avoiding assumptions about sexual behavior and sexual orientation) and featured examples of poor communication (prompting audience identification of problems) and good communication (prompting audience identification of improvements). Adapted from scripts developed by Cara Safon, MPH.

Appendix B

Study Sample Characteristics

	Pre-Interventiona,b	Immediate Post-Interventiona			3 Months Post-Interventiona
	Full Sample n (%)	Full Sample n (%)	PCC Group n (%)	Standard HIV Group n (%)	Full Sample n (%)	PCC Group n (%)	Standard HIV Group n (%)
Age
Under 30	1 (1.0)	1 (2.3)	1 (6.3)	0 (0.0)	1 (3–1)	1 (9.1)	0 (0.0)
30–39	3 (14.3)	7 (15.9)	1 (6.3)	6 (21.4)	7 (21.9)	1 (9.1)	6 (28.6)
40–49	10 (17.9)	9 (20.5)	2 (12.5)	7 (25.0)	7 (21.9)	1 (9.1)	6 (28.6)
50–59	15 (93.8)	12 (27.3)	4 (25.0)	8 (28.6)	9 (28.1)	4 (36.4)	5 (23.8)
60–39	17 (30.4)	11 (25.0)	6 (37.5)	5 (17.9)	6 (18.8)	3 (27.3)	3 (14.3)
70 or older	5 (8.9)	4 (9.1)	2 (12.5)	2 (7.1)	2 (6.3)	1 (9.1)	1 (4.8)
Gender c
Woman	45 (81.8)	33 (81.8)	12 (75.0)	24 (85.7)	27 (84.4)	8 (72.7)	19 (90.5)
Man	10 (18.2)	8 (18.2)	4 (25.0)	4 (14.3)	5 (15.6)	3 (27.3)	2 (9.5)
Country of Birth
USA	52 (92.9)	42 (95.5)	15 (93.8)	27 (96.4)	30 (93.8)	10 (90.9)	20 (95.2)
Other	4 (7.1)	2 (4.5)	1 (6.3)	1 (3.6)	2 (6.3)	1 (9.1)	1 (48)
Ethnicity
Non-Latino/x/a	55 (98.2)	43 (97.7)	15 (93.8)	28 (100.0)	31 (96.9)	10 (90.9)	21 (100.0)
Latino/x/a	1 (1.8)	1 (2.3)	1 (6.3)	0 (0.0)	1 (3.1)	1 (9.1)	0 (0.0)
Race
White	48 (85.7)	39 (88.6)	14 (87.5)	25 (89.3)	27 (84.4)	9 (81.8)	18 (85.7)
Asian	4 (7.1)	3 (6.8)	1 (6.3)	2 (7.1)	3 (9.4)	1 (9.1)	2 (9.5)
Black	2 (3.3)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Other	2 (3.6)	2 (4.5)	1 (6.3)	1 (3.6)	2 (6.3)	1 (9.1)	1 (4.8)
Sexual Orientation d
Heterosexual	51 (94.4)	41 (95.3)	14 (93.3)	27 (96.4)	29 (93.5)	9 (90.0)	20 (95.2)
Gay/Lesbian	2 (3.7)	1 (2.3)	1 (6.7)	0 (0.0)	1 (3.2)	1 (10.0)	0 (0.0)
Bisexual	1 (1.9)	1 (2.3)	0 (0.0)	1 (3.6)	1 (3.2)	0 (0.0)	1 (4.8)
Clinical Background e,f
Nursing	39 (69.3)	31 (70.5)	12 (75.0)	19 (67.9)	25 (78.1)	8 (72.7)	17 (81.0)
Medicine	13 (28.3)	13 (29.5)	4 (25.0)	9 (32.1)	7 (21.9)	3 (27.3)	4 (19.0)
Other	2 (3.6)	1 (2.3)	0 (0.0)	1 (3.6)	1 (3.1)	0 (0.0)	1 (4.8)
Type of Provided
Primary Care	25 (44.3)	21 (47.7)	7 (43.8)	14 (50.0)	16 (50.0)	6 (54.5)	10 (47.6)
Specialistf	33 (64.3)	28 (63.6)	13 (81.8)	15 (53.6)	20 (62.5)	8 (72.7)	12 (57.1)
Psychiatry/Mental Health	20 (35.7)	14 (31.8)	8 (50.0)	6 (21.4)	8 (25.0)	3 (27.3)	5 (23.8)
Obstetrics/Gynecology/Women's Health	5 (8.9)	5 (11.4)	1 (6.3)	4 (14.3)	5 (15.6)	1 (9.1)	4 (19.0)
Addiction Medicine	4 (7.1)	4 (9.1)	2 (12.5)	2 (7.1)	3 (9.4)	2 (18.2)	1 (4.8)
Infectious Diseases	3 (5.4)	3 (6.8)	2 (12.5)	1 (3.6)	3 (9.4)	2 (18.2)	1 (4.8)
Other Specialty	5 (8.9)	4 (9.1)	1 (6.3)	3 (10.7)	3 (9.4)	1 (9.1)	2 (9.5)
Practice Setting f
Private Practice/Physician Practice Group	23 (41.1)	10 (43.2)	8 (50.0)	11 (30.3)	11 (34.4)	4 (36.4)	7 (33.3)
Community Health Center	17 (30.4)	14 (31.8)	5 (31.3)	0 (32.1)	11 (34.4)	2 (18.2)	0 (42.0)
Hospital	9 (16.1)	7 (15.0)	1 (6.3)	6 (21.4)	6 (18.8)	1 (0.1)	5 (23.8)
University	5 (8.9)	3 (6.8)	2 (12.5)	1 (3.6)	2 (6.3)	1 (0.1)	1 (4.8)
Other	18 (32.1)	15 (34.1)	8 (50.0)	7 (25.0)	11 (34.4)	7 (63.6)	4 (10.0)
Licensure to Prescribe Medication in US
Licensed	54 (56.4)	42 (05.5)	16.0 (100.0)	26 (02.9)	30 (03.8)	11 (100.0)	19 (90.5)
Not Licensed	2 (3.6)	2 (4.5)	0 (0.0)	2 (7.1)	2 (6.3)	0 (0.0)	2 (9.5)
Prior PrEP Awareness (Pre-Intervention) g
Never Heard of PrEP	10 (18.2)	8 (18.2)	3 (18.8)	5 (17.0)	4 (12.5)	2 (18.2)	2 (9.5)
Heard of PrEP	45 (81.8)	36 (81.8)	13 (81.3)	23 (82.1)	28 (87 5)	9 (81.8)	19 (90.5)
Prior PrEP Experience (Pre-Intervention) f
Discussed PrEP with 1+ Patient(s)	21 (37.5)	18 (40.0)	7 (43.8)	11 (39.3)	14 (43.8)	6 (54.5)	8 (38.1)
Prescribed PrEP for 1 + Patient(s)h	8 (14.8)	6 (14.3)	2 (12.5)	4 (15.4)	5 (16.7)	2 (18.2)	3 (15.8)
Total	56 (100.0)	44 (100.0)	16 (100.0)	28 (100.0)	32 (100.0)	11 (100.0)	21 (100.0)

^aAll characteristics reported during pre-intervention survey

^bStratification by group (PCC Intervention vs. Standard HIV CME) not presented because group determined based on self-reported attendance during immediate post-intervention survey

^cResponse options included "woman," "man," "transgender woman." "transgender “man," "gender queer," “other" and "prefer not to say." No participants identified as a transgender woman, a transgender man, gender queer, or other, and the participant who selected "prefer not to say" was recoded as missing (denominators adjusted accordingly).

^dResponse options included "lesbian," "gay," "bisexual," "heterosexual (straight)", "pansexual," "asexual," "other," and "prefer not to say." No participants identified as pansexual, asexual, or other, and the two participants who selected "prefer not to say" were recoded as missing (denominators adjusted accordingly).

^eNursing backgrounds included Registered Nurse (RN), Nurse Practitioner (NP), Advanced Practice Registered Nurse (APRN), and Certified Nurse-Midwife (CNM). Medicine backgrounds included Doctor of Medicine (MD) and Doctor of Osteopathic Medicine (DO). Other backgrounds included Physician Assistant (PA) and Licensed Marriage and Family Therapist (LMFT).

^fCategories not mutually exclusive

^gn = 55 pre-intervention due to missing data (denominator adjusted accordingly)

^hSample restricted to participants who reported being licensed to prescribe medicine

PCC = PrEP and Cultural Competence

Appendix C

Effect of PCC Intervention on Confidence Performing Six PrEP-Related Clinical Activities

	Talking to Patient About Sexual Health Historya				Screening Patient to Determine PrEP Candidacy				Answering Patienťs Questions About PrEP
	Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post
Study Condition	M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)
PCC Intervention	4.0 (0.3)	3.7 (0.3)	3.8 (0.3)		2.8 (0.4)	3.6 (0.4)	3.4 (0.3)		2.2 (0.4)	3.3 (0.4)	3.0 (0.4)
Standard HIV CME	4.2 (0.2)	4.0 (0.2)	4.1 (0.2)		2.7 (0.3)	2.7 (0.3)	3.1 (0.2)		2.3 (0.3)	2.4 (0.3)	2.8 (0.3)
Effect	df	F	p	ηp 2	df	F	p	ηp 2	df	F	p	ηp 2
Study Condition	1. 29	0.6	.445	.02	1, 29	1.3	.265	.04	1, 29	0.6	.444	.02
Time	2, 59	0.2	.791	.01	2, 58	2.5	.094	.08	2, 58	2.4	.101	.08
Study Condition × Time	2, 59	0.2	.853	.01	2, 58	2.8	.070	.09	2, 58	5.3	.008	.15
PCC Intervention (Simple)	-	-	-	-	2, 28	3.7	.038	.21	2, 28	9.5	<.001	.40
Standard HIV CME (Simple)	-	-	-	-	2, 28	2.9	.074	.17	2, 28	2.4	.109	.15
	Suggesting PrEP to a Patient				Counseling Patient About Other Methods of Protection				Effectively Monitoring Patient on PrEP
	Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post
Study Condition	M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)
PCC Intervention	2.4 (0.4)	3.4 (0.4)	3.1 (0.4)		3.0 (0.4)	3.9 (0.4)	3.8 (0.4)		2.3 (0.5)	3.3 (0.4)	2.8 (0.4)
Standard HIV CME	2.5 (0.3)	2.6 (0.3)	2.9 (0.3)		3.8 (0.3)	3.5 (0.3)	3.5 (0.3)		2.3 (0.3)	2.2 (0.3)	2.4 (0.3)
Effect	df	F	p	ηp 2	df	F	p	ηp 2	df	F	p	ηp 2
Study Condition	1, 29	0.4	.519	.01	1, 29	<.01	.949	<.01	1r 29	1.1	.303	.04
Time	2, 58	1.7	.198	.05	2, 58	1.2	.319	.04	2, 58	1.4	.254	.05
Study Condition × Time	2, 58	3.9	.025	.12	2, 58	4.5	.015	.13	2, 58	5.3	.007	.16
PCC Intervention (Simple)	2, 28	6.8	.004	.33	2, 28	4.3	.024	.23	2, 28	4.6	.018	.25
Standard HIV CME (Simple)	2, 28	1.8	.183	.11	2, 28	0.8	.461	.05	2, 28	0.6	.550	.04

Note. All statistical models were adjusted for prior PrEP awareness (pre-intervention). Means and standard errors reflect adjusted estimates.

^aSimple effects are not shown because the interaction was non-significant.

PCC = PrEP and Cultural Competence CME=Contlnuing Medical Education

Confidence performing PrEP-related clinical activities at three time points in PCC Condition.

Means were adjusted for prior PrEP awareness (pre-intervention). Within the PCC intervention condition, participants’ confidence screening participants for PrEP candidacy, answering a patient’s questions about PrEP, suggesting PrEP to a patient, counseling a patient about using other methods of protection, and effectively monitoring a patient on PrEP increased immediately following the training. The immediate gain in confidence remained statistically significant three months post-intervention for answering a patient’s questions about PrEP and marginally significant for suggesting PrEP to a patient.

Appendix D

Replication of Effectiveness Analyses Using Imputed Dataset

	Overall PrEP-Related Confidence				Intention to Prescribe PrEPa				Cultural Competenceb
	Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post		Pre	Immed. Post	3 Mos. Post
Study Condition	M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)		M (SE)	M (SE)	M (SE)
PCC Intervention	2.6 (0.2)	3.5 (0.3)	3.4 (0.2)		1.8 (0.3)	2.6 (0.3)	2.4 (0.4)		3.5 (0.2)	3.6 (0.2)	3.6 (0.2)
Standard HIV CME	2.9 (0.2)	2.8 (0.2)	3.1 (0.2)		1.9 (0.3)	1.9 (0.3)	2.1 (0.3)		3.6 (0.1)	3.5 (0.1)	3.7 (0.1)
Effect	df	F	p	ηp 2	df	F	p	ηp 2	df	F	p	ηp 2
Study Condition	1, 41	0.8	.378	.02	1, 39	0.5	.491	.01	1, 41	0.0	.835	.00
Time	2, 72	5.3	.009	.12	2, 62	4.0	.031	.09	2, 82	0.8	.436	.02
Study Condition × Time	2, 72	6.6	.004	.14	2, 62	3.1	.062	.07	2, 82	1.1	.349	.03
PCC Intervention (Simple)	2, 40	16.9	<.001	.46	2, 38	10.5	<.001	.36	-	-	-	-
Standard HIV CME (Simple)	2, 40	1.2	.302	.06	2, 38	0.5	.590	.03	-	-	-	-
	PrEP Discussed (Past 3 Mos.)				PrEP Prescribed (Past 3 Mos.) a
	Pre	3 Mos. Post			Pre	3 Mos. Post
Study Condition	n (%)	n (%)			n (%)	n (%)
PCC Intervention	3 (18.8)	10 (62.5)			2 (12.5)	7 (43.8)
Standard HIV CME	7 (25.0)	13 (46.4)			3 (11.5)	4 (15.4)
Effect	aOR	95%CI	p		OR	95%CI	p
Study Condition (ref = Standard HIV CME)	0.7	0.2 – 3.1	.634		1.1	0.2 – 7.4	.926
Time (ref = pre-intervention)	2.6	1.0 – 7.0	.049		1.4	0.5 – 4.3	.563
Study Condition × Time	2.8	0.6 – 14.6	.212		3.9	0.7 – 23.3	.134
PCC Intervention (Simple)	-	-	-		-	-	-
Standard HIV CME (Simple)	-	-	-		-	-	-

Note. All statistical models were adjusted for prior PrEP awareness (pre-intervention) except for the PrEP Prescribed model (due to distribution of data). Means and standard errors reflect adjusted estimates.

^aRestricted to licensed prescribers (n = 42 of 44 at three months post-intervention)

^bSimple effects are not shown because the interaction was non-significant.

PCC = PrEP and Cultural Competence CME=Continulng Medical Education

Appendix E

PrEP and Cultural Competence (PCC) Participants’ Reasons for Not Discussing and Prescribing PrEP Three Months Post-Intervention

Reasona	3 Months Post-Intervention
	Did Not Discussb	Did Not Prescribec
	n (%)	n (%)
Not enough time during patient appointments	0 (0.0)	0 (0.0)
Have not seen any patients at high HIV risk	3 (75.0)	4 (44.4)
Not enough training to feel comfortable	2 (50.0)	2 (22.2)
Talking about HIV risk not part of my role	1 (25.0)	1 (25.0)
Donť want to deal with payment/insurance hassle	0 (0.0)	0 (0.0)
Donť want to deal with patient monitoring/follow-up	0 (0.0)	2 (22.2)
Believe patients are better off with other protection (e.g., condoms or abstinence)	0 (0.0)	0 (0.0)
Afraid patients will stop using condoms or have more partners	0 (0.0)	0 (0.0)
PrEP is too new too trust it works	0 (0.0)	0 (0.0)
Have brought it up but patients not interestedd	-	2 (22.2)
Othere	1 (25.0)	6 (66.7)

^aReasons were not mutually exclusive. All participants provided 1+ reason(s).

^bn = 4 of 11 participants in PCC condition reported not discussing PrEP with any patients during three-month post-intervention period

^cn = 9 of 11 participants in PCC condition reported not prescribing PrEP for any patients during three-month post-intervention period

^dResponse option only applicable to not prescribing

^eAll participants who specified another reason for not discussing or prescribing PrEP suggested PrEP was outside of their purview. For both PrEP discussion and prescription, one participant who specified this "other" reason also endorsed "Talking about HIV is not part of my role."