Medical Distrust and the Intention to Initiate Pre-Exposure Prophylaxis in Black Cisgender Women

Abstract

Black women bear a disproportionate burden of the US HIV epidemic, compared with women of other racial groups. Pre-exposure prophylaxis (PrEP) is a highly effective prevention tool. Evidence indicates that Black women are interested in initiating PrEP, but low utilization persists in this population. Historical mistreatment of women of color and the resulting distrust erect barriers to communication and shared decision-making with health care providers. Using an institutional review board–approved questionnaire, we surveyed 186 adults (83% Black; 9% White; 3% American Indian/Alaskan Native; 4% Other) who were PrEP eligible, HIV seronegative, cisgender women in Washington, DC. We tested the interaction of patient racial identification and group-based medical mistrust on intentions to use PrEP, as mediated by intentions to discuss PrEP with a health care provider during the imminent clinical interaction. Results indicate significant moderated mediation of the interaction between race and distrust on intention to initiate PrEP at 3 months [index = −0.3093, standard error (SE) = 0.1886, 95% confidence interval (CI; −0.7455, −0.0122)] and 12 months [index = −0.3248, SE = 0.1987, 95% CI: (−0.7827, −0.0040)] through an anticipated discussion with a provider. When distrust is low, Black women had stronger intentions to utilize PrEP (relative to women of other racial groups). This was explained by stronger intentions to discuss PrEP with the provider during the clinical visit. These results underscore the critical importance of provider-initiated discussion of PrEP with women to improve health equity. This study was limited by the low number of non-Black participants (17%) as well as the requirement that subjects be English-speaking only.

Introduction

In the United States, Black cisgender women (hereafter, BW) disproportionately bear the burden of HIV compared with other racial and ethnic groups, constituting 47% of new infections among women nationally, despite comprising only 13% of the population. ¹ In Washington, DC, this epidemic is magnified, with BW accounting for 90% of women with HIV. ² One of the most effective tools to prevent transmission of HIV is pre-exposure prophylaxis (PrEP), including daily oral tenofovir disoproxil fumarate/emtricitabine and long-acting injectable cabotegravir, which can reduce HIV acquisition by up to 99% in women. ^{3 –5} Despite its effectiveness, PrEP is grossly underutilized by women, particularly BW, who comprise only 25.9% of cisgender female PrEP users with available race/ethnicity data. ⁶ Given the marked disparities between HIV incidence and PrEP uptake, understanding the factors contributing to these inequities is critically important.

Barriers to PrEP use among BW are multifaceted, encompassing intrapersonal, interpersonal, institutional, and structural domains as delineated in the literature. ^{7 –14} Among the multitude of social and structural factors that present barriers to women’s uptake of PrEP, the clinical interaction, in particular, plays a critically important role in women’s PrEP use. ^11,15 As a biomedical innovation that requires a prescription by a medical provider, the diffusion of PrEP depends on communication between providers and patients. Research has shown that racial bias creates barriers to providers’ communication about PrEP that disfavor Black women that can undermine the patient–provider relationship and erode trust. ^15,16 Evidence also demonstrates that women rely heavily on the recommendation of their health care provider when considering PrEP uptake. ¹¹ In light of this context, this study investigates medical distrust’s role in shaping Black women’s intentions to initiate PrEP.

We hypothesized that when BW experience distrust of the medical establishment, they are less inclined to discuss PrEP with clinicians, and as a result, less likely to intend to use PrEP for HIV prevention. To test this hypothesis, we conducted a cross-sectional survey from a convenience sample of women recruited from waiting areas of women’s health clinics in Washington, DC.

Methods

Study design

We recruited cisgender, PrEP-eligible, HIV seronegative female participants ≥18 years of age. Participants were recruited from women’s health clinics at a tertiary care medical center in Washington, DC, from August 2020 to September 2022. An institutional review board (IRB) approval was received from the MedStar Health Internal Review Board at both of our recruitment sites, in Washington, DC, before starting the recruitment process (IRB ID: STUDY00000733).

Members of the research team approached women in the women’s health clinics of the tertiary medical center. The demographics of the research team included individuals who were African American/Black and White, with varying roles/involvement in the health care field (research assistant, medical student, resident, and physician). All English-speaking women over the age of 18 were invited to complete the survey. Surveys were delivered using an anonymous, tablet or smartphone-based survey. The survey assessed sociodemographic characteristics (e.g., age, race, ethnicity, relationship status, level of education, income, health insurance, employment), perceived racism in the context of health care, communication, and intention to initiate PrEP. The survey included an embedded educational video to provide information about PrEP. ¹⁷ The survey took approximately 25 min to complete. Participants were compensated with a $10 gift card.

Survey measures

Behavioral risk factors

We assessed sexual history including the number of casual (i.e., uncommitted) and main (i.e., committed) sexual partners (i.e., none, some, a few, most, all), frequency of condom use (i.e., never, rarely, sometimes, often), intercourse with person who has HIV (i.e., past 12 months), transactional sex (i.e., ever), recent history of sexually transmitted infections, and injection drug use with shared equipment (i.e., ever).

Medical distrust

We assessed the 12-item Group-Based Medical Mistrust (GBMM) Scale ¹⁸ using 5-point Likert scales (1 = “strongly disagree�?; 5 = “strongly agree�?). The scale includes three subscales that evaluate distrust as experienced by a racial social group based on: (1) suspicion, (2) group disparities in health care, and (3) lack of support from health care providers (Cronbach’s α = 0.87). Although distrust and mistrust both refer to a lack of trust, they are not synonymous. Mistrust captures general unease or suspicion, while distrust is rooted in lived experiences with health care services, mistreatment, or knowledge of inadequate care in health care systems. ¹⁶ In the context of this study, we utilize the term “distrust�? as it more accurately characterizes the history of mistreatment of Black people within health care spaces in the United States.

Intention to discuss PrEP with a provider

We used a 5-point single-item measure to assess intention to discuss PrEP with a provider: “Do you plan to talk with your doctor about PrEP today?�? (1 =“definitely not�?; 5 = “definitely yes�?).

PrEP uptake behavioral intention

We assessed short- (i.e., 3 months) and long-term (i.e., 12 months) intention to initiate PrEP using 5-point Likert scale measures: “Do you plan on using PrEP to reduce your risk of getting HIV in the next (3 or 12) months?�?

Race

Participants’ racial identity was assessed, allowing respondents to select all that apply: (1) Black, (2) American Indian/Alaskan Native, (3) Asian (4), Native Hawaiian/Pacific Islander, (5) White, and (6) Other.

Analysis

Analyses were conducted using IBM SPSS Statistics Version 29. We analyzed descriptive statistics between groups to establish the sociodemographic characteristics of the sample, as well as the behavioral risk factor characteristics, using Chi-square for tests of significant differences between groups. This descriptive analysis included information regarding the age, relationship status, income level, education, employment, insurance of the respondents, sexual history, and injection drug use with shared equipment. We assessed the mean differences between groups of Black and non-Black women (White, American Indian/Alaskan Native, Asian, and Native Hawaiian/Pacific Islander) for the variables of GBMM, intention, as well as intention to discuss PrEP with a provider, using independent samples t-tests.

We tested the interaction of patient racial identification and GBMM on the intention to use PrEP, as mediated by the intention to discuss PrEP with a health care provider during the imminent clinical interaction using the PROCESS macro (Model 7). This model calculates an estimate of both direct and indirect effects of model predictors on the outcomes and also tests for moderation of mediated effect. Levels of medical distrust (low, moderate, and high) were derived using conditioning values of −1 standard deviation (SD) from the sample mean, sample mean, and +1 SD from the sample mean. We provided both unadjusted and adjusted models, with the adjusted model controlling for age, relationship status, income, education, employment, and insurance.

Results

Sample characteristics

A total of 353 eligible women initiated the study, of which 186 women completed all of the study items. The below patient demographics are representative of the clinic demographics.

The majority of participants were Black (83%), not married (82.8%), had a mean income <$20,000 (53.8%), were employed (53.8%), and had insurance (98.9%). The mean age of the study population was 30.6 years. The sample includes 154 Black women (83%) and 32 non-Black women (17%). Table 1 demonstrates the sociodemographic characteristics of the respondents.

Table 1.

Participant Characteristics

Participant characteristics	Non-Black n = 32 (17%) n (%)	Black n = 154 (83%) n (%)	Total N = 186 (%) n (%)
Age (mean, SD)	32.8 (±8.3)	30.1 (±14)	30.6 (±13.2)
Relationship status *
Married	14 (43.8)	18 (11.7)	32 (17.2)
Income Level *
$0–19,999	9 (28.1)	90 (59.2)	99 (53.8)
$20,000-$49,999	9 (28.1)	40 (26.3)	49 (26.6)
$50,000-$79,999	2 (6.3)	19 (12.5)	21 (11.4)
$80,000 and above	12 (37.5)	3 (2)	15 (8.2)
Education *
Less than high school degree	1 (3.1)	7 (4.5)	8 (4.3)
High school graduate	6 (18.8)	75 (48.7)	81 (43.5)
Some college	3 (9.4)	43 (27.9)	46 (24.7)
Associate degree (2-year)	2 (6.3)	6 (3.9)	8 (4.3)
Bachelor’s degree (4-year)	13 (40.6)	18 (11.7)	31 (16.7)
Advanced degree (MA, PhD, MD, JD)	7 (21.9)	5 (3.2)	12 (6.5)
Employment
Employed	22 (68.8)	78 (50.6)	100 (53.8)
Unemployed	9 (28.1)	71 (46.1)	80 (43.0)
Prefer not to answer	1 (3.1)	5 (3.2)	6 (3.2)
Insurance *
Insured	30 (93.8)	154 (100.0)	184 (98.9)
Uninsured	2 (6.2)	0	2 (1.1)
Behavioral Risk Factors
More than one sexual partner	4 (12.5)	25 (16.2)	29 (15.6)
Main sexual partner(s)	22 (68.8)	107 (69.5)	129 (69.4)
Casual sexual partner(s)	11 (34.4)	65 (42.2)	76 (40.9)
Inconsistent condom use	14 (43.8)	54 (35.1)	68 (36.6)
Intercourse with an HIV-positive partner	2 (6.3)	2 (1.3)	4 (2.2)
Transactional intercourse	0 (0)	2 (1.3)	2 (1.1)
Recent STI	0 (0)	10 (6.5)	10 (5.4)
Injection drug use and/or needle sharing	1 (3.1)	1 (0.6)	2 (1.1)

^* p < 0.001.

SD, standard deviation; STI, sexually transmitted infection.

In terms of potential HIV exposure, 15.6% of respondents indicated having more than one sexual partner, 40.9% indicated having a casual sexual partner, 36.6% reported inconsistent condom use with casual sexual partners, 2.2% had intercourse with an HIV-positive partner in the last 12 months, 1.1% indicated they exchanged sex for drugs or money, 5.4% reported a recent sexually transmitted infection (STI), and 1.1% indicated sharing intravenous drug use equipment. Significant differences between groups was evident, with observed differences in the Black subgroups in the categories of relationship status, income level, education, and insurance (Table 1).

Table 2 demonstrates that among the subscales, and in full scale for GBMM, Black women had higher reported levels of medical distrust, although these differences were not statistically significant. Black women reported a greater intention to discuss PrEP with a provider and higher intentions to initiate PrEP at 3 months and 12 months relative to non-Black women (Table 2).

Table 2.

Mean Differences Between Racial Groups

Moderator, mediator, and outcome variables	Non-Black n = 32Mean (SD)	Black n = 154Mean (SD)	Total n = 186Mean (SD)
GBMM_SG	2.20 (±1.00)	2.53 (±1.04)	2.47 (±1.04)
GBMM_DiH	3.31 (±1.11)	3.31 (±1.24)	3.16 (±1.21)
GBMM_LoS	2.44 (±1.13)	2.67 (±0.84)	2.63 (±0.90)
GBMM	2.54 (±0.84)	2.72 (±0.88)	2.69 (±0.88)
Intention to discuss PrEP w/ provider *	1.84 (±1.20)	2.36 (±1.32)	2.27 (±1.31)
Intention to initiate (12 months) *	1.81 (±1.15)	2.29 (±1.26)	2.21 (±1.25)
Intention to initiate (3 months) **	1.38 (±0.79)	2.27 (±1.28)	2.11 (±1.25)

^* p < 0.05.

^** p < 0.01.

GBMM_SG, Group-Based Medical Mistrust—Suspicious Group Subscale; GBMM_DiH, Group-Based Medical Mistrust—Disparities in Health Subscale; GBMM_LoS, Group-Based Medical Mistrust—Lack of Support from Health Care Providers Subscale; PrEP, pre-exposure prophylaxis.

As shown in Fig. 1, there was a significant indirect relationship between racial identification and intention to initiate PrEP in the next 3 months and 12 months, which was mediated by the intention to discuss PrEP with a provider. This mediated relationship was moderated by medical distrust in both unadjusted models for short- and long-term intention [unadjusted model at 3 months: −0.31 (0.19) 95% confidence interval (CI; −0.75, −0.012); unadjusted model in 12 months: −0.32 (0.20) 95% CI: (−0.79, −0.02)]. After controlling for relationship status, education, income, and insurance, the indirect relationship was no longer significant, but the patterns of the direct relationships between intention to discuss PrEP with a provider and intention to initiate PrEP in both 3 and 12 months are consistent with the short-term outcome.

FIG. 1.

Models representing the unadjusted relationships and adjusted models (both direct and indirect) of race on intention to initiate PrEP at 3 months and 12 months through comfort discussing with a provider as moderated by medical distrust. (a) Unadjusted model with intention to initiate PrEP within 3 months. (b) Unadjusted model with intention to initiate PrEP within 12 months. (c) Adjusted model with intention to initiate PrEP within 3 months. (d) Adjusted model with intention to initiate PrEP within 12 months. *p < 0.05; **p < 0.01; ***p < 0.001. PrEP, pre-exposure prophylaxis.

Finally, Table 3 illustrates the moderation of the indirect relationship by medical distrust. At low levels of medical distrust, BW (relative to non-Black) reported stronger intention to utilize PrEP, and this relationship is mediated through intention to discuss PrEP with their provider during the clinical visit (see Table 3). This effect decreases at higher levels of medical distrust, although the differences between groups at moderate and high levels of distrust are not statistically significant. Figure 2 illustrates the relationship between the intention to discuss PrEP with a provider and medical distrust among Black and non-Black women. The largest disparity between Black and non-Black women is when distrust is low.

Table 3.

Conditional Indirect Effect of Race on Intention to Initiate PrEP Through Intention to Discuss with a Provider at Levels of Low, Moderate, and High Levels of Medical Distrust

Indirect relationship	Level of distrust	Indirect effect (SE)	95% CI
Black→discuss→INT_3	Low	0.57 (±0.17)	0.26–0.91
Black→Discuss→INT_3	Moderate	0.30 (0.17)	−0.039 to 0.63
Black→Discuss→INT_3	High	0.033 (0.29)	−0.63 to 0.51
Index of moderated mediation	N/A	−0.31 (0.19)	−0.75 (−0.012)
Black→Discuss→INT_12	Low	0.60 (0.17)	0.28–0.93
Black→Discuss→INT_12	Moderate	0.32 (0.17)	−0.047 to 0.63
Black→Discuss→INT_12	High	0.034 (0.30)	−0.69 to 0.52
Index of moderated mediation	N/A	−0.32 (0.20)	−0.78 (−0.0040)

The results demonstrate that the indirect effect of race on intention to initiate PrEP at 3 months and 12 months is significantly moderated by medical distrust, as mediated through communication with a provider. The only significant mediated relationship between race and intention to initiate PrEP is present when medical distrust is low (indicated by row 1).

p < 0.05.

CI, confidence interval; PrEP, pre-exposure prophylaxis; SE, standard error.

FIG. 2.

Impact of racial grouping on intention to discuss PrEP with a provider as moderated by varying levels of medical distrust. PrEP, pre-exposure prophylaxis.

Discussion

This study explored the hypothesis that medical distrust among BW in health care environments decreases their intention to use PrEP for HIV prevention. Our analysis revealed that lower levels of medical distrust were associated with a stronger intention among BW to utilize PrEP compared to non-BW. This relationship was explained by a greater intention to discuss PrEP with their provider during the visit that day. This study builds upon previous research demonstrating that distrust inhibits PrEP communication between patients and providers. ^20,21 However, previous research has largely focused on cisgender men who have sex with men and transgender individuals, with limited attention to the unique barriers and facilitators to PrEP that are intersectionality specific to women’s experiences. ^7,9,10 Scholars, such as Tekeste et al., who have engaged in research focused on BW, have examined how distrust inhibits communication with providers about PrEP, but their work primarily focuses on communication outcomes and is limited in its extension to intentions or behaviors. ^20,22,23 Our findings contribute to the growing body of literature underscoring the need to center the experiences of those who are underserved by PrEP. This study aimed to understand how the dynamics of communication in clinical interaction may impede the equitable diffusion of PrEP for Black women.

The resulting impacts of medical distrust have been documented as a potential social determinant of health, and the research has cited multiple negative associations between distrust and activities as it pertains to HIV prevention such as decreased adherence to antiretroviral medication. ²⁴ Against the backdrop of the complex historical context surrounding marginalized communities in health care, an ethos of open and supportive provider-patient relationships has emerged as pivotal for enhancing patient engagement within health care settings. ^25,26 One of the most notable mechanisms underlying the outcome of our results is the relationship between the intention to discuss PrEP with a provider and the intention to initiate PrEP. In all of the models we tested (i.e., adjusted vs. unadjusted; short- vs. long-term), we identified a strong significant relationship between intention to discuss PrEP with a provider and intention to initiate PrEP. Once patients demonstrate a willingness to engage in discussions with their providers regarding PrEP, their motivation to initiate PrEP markedly increases, but as demonstrated in this study, this relationship is modified by barriers such as medical distrust. These findings suggest the importance of culturally competent care and the need for provider-initiated discussion of PrEP in the clinical setting, particularly within historically underrepresented populations. ^15,27,28 In particular, this study underscores that the clinical interaction is a crucially important aperture for intervention. Notable barriers to the provider’s initiation of PrEP discussion include the provider’s knowledge (related to patient eligibility, sufficient knowledge about the medication and insurance coverage), time during a clinic visit, comfort levels discussing sexual health, and the provider’s perceptions of the patient’s risk of STI acquisition. ^26,28,31,32 Clinic-level interventions are necessary to address these modifiable factors.

Studies quantifying the current landscape of clinician-initiated discussion are sparse, but one study showed that discussion is limited with estimates of provider-initiated discussion at around 30%, ³⁰ although this study focuses predominantly on a male population. It is likely that provider-initiated discussions of PrEP for women, BW specifically, are substantially lower. ¹⁹ Public health leadership and scholars recommend an expansion of practice such that PrEP is a routine part of clinical visits for all patients, not just those with presumed risk factors. ¹⁵ Current clinical guidelines recommend clinicians take a comprehensive medical history, including sexual history, as part of the well-woman visit. ²⁹ A recommended strategy includes the use of the 5P’s (Partners; Practices; Protection from STIs; history of STIs; and Pregnancy intention) and further inclusion of questions asking, “Are you aware of PrEP, a medicine that can prevent HIV?�? or “Have you ever considered using [PrEP]?�? as a framework to initiate discussions about PrEP use. Calabrese et al. elaborate to note that provider-initiated care is critical to achieving the following goals: (1) avoids missing potential PrEP candidates, (2) destigmatizes PrEP, (3) facilitates patient-centered care, and (4) transmits knowledge about PrEP to the broader community. ¹⁵ In jurisdictions such as Washington, DC, where the prevalence of HIV is high and disparities between Black women and non-Black women are stark, this recommendation becomes especially important in the pursuit of equitable care.

As supported in the literature, this study implies that efforts to raise PrEP awareness for women must not be limited to the provider introduction of PrEP through passive measures such as posters, flyers, and pamphlets. ³³ Rather, that research showed that intention to initiate PrEP was significant when a more robust discussion of PrEP with a provider took place, allowing for a collaborative, shared decision-making approach to risk assessment and PrEP utilization. ^33,34 In a study conducted by Dale et al., they found that BW who underwent motivational interviewing (including strategies that combined psychoeducation case management) regarding PrEP had a significantly higher likelihood of subsequently speaking with a provider about PrEP and utilizing PrEP in the follow-up assessments. ³⁴

The Centers for Disease Control and Prevention’s recommendations highlight specific female demographics that would benefit from PrEP including those with high-risk sexual behaviors and living in HIV-prevalent regions. ³⁵ The data have consistently shown clinical indications alone are insufficient to drive adequate PrEP uptake, necessitating a deeper exploration of the physician–patient relationship. Street et al. describe how “a physician who is skilled in informing, showing respect, and supporting patient involvement can transcend issues of race and sex to establish a connection with the patient that in turn contribute to greater patient satisfaction, trust, and commitment to treatment.�? ³⁶ Future efforts to increase PrEP uptake should focus on interventions that improve provider communication, dismantle stigma, and increase universal patient education about PrEP to achieve a greater, more equitable standard of care.

Limitations

As the study was conducted in an urban setting where women are already engaged in care, selection bias in the patient population may have been present, limiting the study’s overall generalizability. Additionally, the study design called for only English-speaking participants, again limiting the universal applications of these results. There was a much larger proportion of Black women enrolled in this study compared to non-Black women, which may impact the ability to establish the external validity of this study. Along these lines, our results only demonstrate the significance of our moderated mediation in the unadjusted models. When accounting for covariates such as age, relationship status, education, income, employment, and insurance, we see that the significance of the indirect relationship is no longer present, again potentially undermining the external validity of this study. This finding may be a result of the small sample size between our two independent variables (Black women vs. non-Black women) and calls for future work to evaluate these relationships at a larger scale. Finally, the use of self-reported data, as well as the cross-sectional data, also limits our ability to infer causality of our results. For future studies, it will be important to evaluate the actual behaviors (i.e., whether or not individuals uptake the medication) to offer a more direct assessment of medical distrust and PrEP utilization.

Authors’ Contributions

T.A., S.H., and R.K.S. contributed to the conception of the study and developed the conceptual models/methodology. T.A., H.S., J.Z., and P.M. participated in study investigation and resource collection. T.A. wrote the initial draft of the article, including data analysis and visualization, and all authors reviewed and edited subsequent versions. R.K.S. and S.H. provided critical feedback for the final version of the document. All authors have approved the final version of the article.

Author Disclosure Statement

R.K.S.: ViiV (grant/research support—managed by MedStar Health Research Institute).

Funding Information

S.H.’s time was supported by the National Institute on Drug Abuse under Award Number 1K01DA050496-A1. The funders had no role in the study design, collection, analysis, or interpretation of data, in the writing of the report, or the decision to submit the article for publication.