Characterizing interest in and uptake of pre-exposure prophylaxis (PrEP) for HIV prevention among men in a US state correctional system

Abstract

Background:

People experiencing incarceration are disproportionately impacted by HIV and are potential candidates for HIV pre-exposure prophylaxis (PrEP). We explored factors associated with PrEP interest and PrEP uptake and described barriers to PrEP uptake among incarcerated men in a state correctional system.

Methods:

From September 2019 to July 2022, incarcerated men at the Rhode Island Department of Corrections (RIDOC) were screened for PrEP eligibility and referred to a PrEP initiation study. We used bivariate analyses and multivariable logistic regression models to explore factors associated with PrEP interest and uptake in the screening sample.

Results:

Of the men screened and determined to be eligible for PrEP, approximately half (50%) were interested in taking PrEP. Individuals identifying as men who have sex with men (AOR= 4.46, 95% CI [1.86, 11.4]) and having multiple female sex partners (AOR= 2.98, 95% CI [1.47, 6,27]) were more likely to express interest in PrEP (Interested/Not Interested) than those not reporting these behavioral factors. PrEP uptake (Yes/No) was 38%. Lack of PrEP interest, low self-perceived risk of HIV acquisition, and unpredictable lengths of incarceration were the most frequently encountered barriers to PrEP uptake.

Conclusions:

Men reporting sexual transmission behaviors were more interested in PrEP and had higher uptake than other men. PrEP interest and HIV risk factors were both moderately high, which suggests that men experiencing incarceration should be screened for and offered PrEP as part of standard clinical care. Study findings have important implications for research and practice to adapt PrEP care to correctional systems.

Short Summary

A study of PrEP implementation for men experiencing incarceration found specific behavioral factors associated with interest in PrEP and PrEP uptake while incarcerated.

Introduction

HIV disproportionately impacts marginalized communities globally and within the United States (U.S.), including men who have sex with men (MSM) and people who inject drugs (PWIDs).¹ Nationally, there were about 40,000 incident HIV diagnoses from January 2021 through March 2022 among persons aged 13 years and above, of which men accounted for 80%.¹ HIV pre-exposure prophylaxis (PrEP) is an effective HIV prevention approach among populations at increased risk, such as MSM, PWIDs, and sex workers.^2,3 Oral PrEP medications include Tenofovir Disoproxil Fumarate and Emtricitabine (TDF/FTC) and Tenofovir Alafenamide and Emtricitabine (TAF/FTC), and are taken prior to a potential exposure to HIV.²

There is substantial overlap between the populations in the U.S. that are impacted by HIV and those that experience incarceration.³ About one in seven persons with HIV in the U.S. will experience incarceration either in jail or prisons in their lifetime.⁴ Incarcerated people are disproportionately people of color, people with substance use disorders including people who inject drugs (PWIDs), sex workers, persons who are transgender and members of other sexual minority groups. ^5–7 During the immediate post-release period, formerly incarcerated individuals experience a greater risk of HIV acquisition due to factors such as a disruption of sexual networks and changes in substance use behaviors.^8,9 Recognizing the susceptibility of incarcerated populations to sexually transmitted diseases, the Center for Disease Control and Prevention (CDC) listed persons in correctional facilities as a special population in its Sexually Transmitted Infections Treatment Guidelines, 2021.¹⁰ The guidelines recommended that clinical providers identify persons at high risk of HIV acquisition, initiate them on PrEP, and provide linkage to PrEP care in the community prior to their release.¹⁰

The PrEP care continuum can be used to guide intervention development and to address potential challenges encountered during PrEP awareness, uptake, and adherence.¹¹ The PrEP care continuum has not been comprehensively characterized among incarcerated populations.¹² Only a few studies have assessed perceived barriers and facilitators to PrEP care among incarcerated populations, and there are a limited number of interventions to improve PrEP care outcomes that were conducted or evaluated in correctional settings.^13,14 Comparatively, barriers to PrEP use in community settings have been more thoroughly characterized. Known barriers include low awareness of PrEP, low self-perceived risk of HIV acquisition, stigma, and high cost of PrEP clinical care and insurance coverage.^15–17 Incarcerated populations likely face unique challenges to maintaining PrEP adherence when they re-enter society, such as substance use disorders, barriers to disclosing HIV risks to community care providers, and having competing health priorities.^12,18

The study team has previously published an adapted version of the PrEP care continuum for jail/prison settings.¹² Given that PrEP implementation in the correctional setting is still novel and largely untested, the goal of the present study was to explore the beginning of the continuum and specifically to explore factors associated with interest in and uptake of PrEP among men experiencing incarceration. The study seeks to inform the implementation of PrEP care to meet the needs of incarcerated populations, which can potentially reduce HIV incidence on a population-wide scale. To this end, the study team has published an adapted version of the socio-ecological model to present the multi-level barriers to PrEP implementation, ranging from the individual, interpersonal to the structural level, for both the correctional setting and the community setting.¹⁹ The use of the socio-ecological model offers the potential to better understand the factors that impact PrEP uptake among incarcerated men at increased risk of HIV acquisition.

Materials and Methods

Study setting

The study was conducted in the Rhode Island Department of Corrections (RIDOC), Rhode Island’s unified, statewide correctional system wherein pre-trial and sentenced populations both fall under the same medical and security administration. During the course of routine clinical care, men who are housed in five men’s facilities, including the intake facility, are regularly screened for HIV acquisition risk and potential PrEP candidacy. For example, upon commitment to the intake, or pre-trial, facility, all individuals undergo a standardized, nurse-led health screening which include a series of questions related to HIV including a history of HIV seropositivity, recent STI symptoms or diagnosis, sexual as well as substance use behaviors including injection drug use history. In the fiscal year (FY) 2022, the intake facility processed 8,477 commitments, which is about 706 commitments per month.²⁰ RIDOC data indicate that on a typical day, ninety-two percent of the total pre-trial population (N=704) were men (n=646). Among both sexes, 40% were White, 28% were Black, 27% were Hispanic, and less than 5% were Asian, Native American, of other races and unknown race.¹⁵

Study procedures

The study examined screening and referral data from a parent study funded by Gilead Sciences, which established a prospective cohort of men experiencing incarceration who expressed an interest in initiating PrEP while incarcerated and being linked to care in the community post-release.¹² In the RIDOC, PrEP could not be initiated outside of the context of the parent study. From November 2019 to July 2022, men who were identified as being potentially at increased risk of HIV acquisition through standard clinical care were referred to a physician on staff, who was also the RIDOC-designated “PrEP provider.” Men were referred either during the nurse-led screening process upon an individual’s commitment to the RIDOC intake facility or by nurses, staff physicians, or substance use treatment providers during the course of other routine clinical encounters. Nursing staff, clinical providers and substance use program personnel were provided information on indications for PrEP candidacy as well as risk factors related to increased HIV acquisition risk. Referrals were generated when relevant RIDOC staff identified an individual as potentially at increased risk of HIV acquisition. Men could also self-present for care, such as for HIV/STI testing, which would also prompt a PrEP referral. Referred individuals were scheduled for an appointment to meet with the PrEP provider.

During the visit with the PrEP provider, individuals were then screened for PrEP eligibility using the CDC criteria and were assessed for their interest in PrEP use through a series of standardized clinical questions.² If the individual was not interested in PrEP, the PrEP provider recorded the reasons for their lack of interest. During each clinical encounter, the PrEP provider attempted to obtain relevant demographic information and risk factors for HIV acquisition, such as injection drug history and identifying as a MSM (for full list, see Table 1), as well as presented the benefits and risks of daily oral use of (TDF/FTC) for PrEP. Participation in the study’s clinical encounters as well as other research activities was entirely voluntary with individuals able to decline to provide information or participate in clinical activities at any point. If individuals were not known to be living with HIV, met the CDC’s criteria for PrEP use, were interested in initiating PrEP while incarcerated and in being linked to PrEP care upon community re-entry, and had an anticipated period of incarceration of approximately 30 days, they were deemed “PrEP-eligible” and were offered the opportunity to enroll in the parent study’s prospective cohort. Individuals with a history of HIV infection or who did not meet other CDC criteria for PrEP use were excluded from enrollment.² For the purposes of this study, enrollment was considered equivalent to PrEP uptake. Individuals who agreed to enroll completed a written informed consent. Deidentified data from those who were screened but not consented were retrospectively reviewed upon parent study completion. Study protocols were approved by the Lifespan Institutional Review Board which is specially certified to review research protocols involving populations experiencing incarceration.

Table 1.

Demographic, behavioral characteristics and PrEP interest among PrEP-eligible men (N=260).

	Eligible for PrEP (N=260) n (%)
Race
White	121 (46.5)
African American	84 (32.3)
Asian	1 (0.4)
Native American	4 (1.5)
Multi-racial	4 (1.5)
Missing	46 (17.7)
Ethnicity
Latinx	53 (20.4)
Not Latinx	171 (65.8)
Missing	36 (13.8)
Age
Median (min, max)	30.0 (18.0, 68.0)
Referral source
Self-referred	153 (58.8)
Clinical provider-referred	107 (41.2)
Intake nurse*	52 (48.6)
MD, NP, PA*	37 (34.6)
Public health nurse*	12 (11.2)
Substance use treatment provider*	6 (5.6)
Injection drug use
Within past 12 months	16 (6.2)
Outside of past 12 months	69 (26.5)
Never	175 (67.3)
Multiple female sex partners
Yes	183 (70.4)
No	77 (29.6)
MSM
Yes	47 (18.1)
No	213 (81.9)
Inconsistent condom use
Yes	120 (46.2)
No	140 (53.8)
Recent STI history
Yes	62 (23.8)
No	198 (76.2)
Sexual contact with HIV positive person
Yes	4 (1.5)
No	256 (98.5)
Paid for sex†
Yes	9 (3.5)
No	251 (96.5)
Engaged in sex work
Yes	22 (8.5)
No	238 (91.5)
Partner is injection drug user
Yes	5 (1.9)
No	255 (98.1)
Shares needles
Yes	17 (6.5)
No	243 (93.5)
Total number of risk factors
Mean (SD)	2.13 (1.01)
(Min, Max)	(0, 5)
Interest
Yes	131 (50.6)
Possibly	27 (10.4)
No	101 (39.0)
Not reported	1 (0.4)

^*N=107.

^†Sex work here includes both transactional (consented to have sex with another individual out of expectation that the sex partner will provide the client with money, services, food, etc.) and commercial sex work.

Individuals who enrolled in the study underwent CDC-recommended clinical testing for PrEP initiation, including testing for HIV, STIs, Hepatitis B/C, and creatinine levels.² Participants who completed the necessary clinical evaluation to initiate PrEP and were confirmed to be eligible for PrEP received oral daily TDF/FTC in the correctional facility approximately seven days before their release. Enrolled individuals were provided with about a one month supply of PrEP prior to release to continue in the community pending linkage to a community PrEP provider. The clinical protocol was developed with input from the RIDOC’s health care services with the aim to initiate PrEP shortly before community re-entry to provide protection from HIV acquisition in the highly vulnerable period immediately post-release. Given that sexual behavior and substance use is prohibited within RIDOC’s facilities, and any sexual behavior is required to be reported to RIDOC’s security administration, PrEP was not generally offered during an individual’s incarceration period outside of the immediate pre-release period. If an individual was incarcerated beyond the initial anticipated 30-day period, retesting was made on a case-by-case basis in line with CDC recommendations. All study procedures were approved by both the Rhode Island Department of Corrections Medical Research Advisory Group and the Lifespan Institutional Review Board (IRB# 204119).

Analytic Plan

All statistical analyses were conducted using R studio.²³ Individuals who were not PrEP-eligible were excluded prior to analyses. Descriptive statistics, bivariate analyses (Chi-square tests and Fisher’s Exact Test), and multivariable regression analyses included PrEP-eligible men only (N=260). Due to the non-random distribution of missing race and ethnicity across PrEP uptake status and PrEP interest, observations with missing race and ethnicity were excluded from bivariate and multivariable analyses.

Multivariable binomial logistic regression models were used to identify factors associated with PrEP interest (Interested/Not interested) and PrEP uptake (Yes/No). Consistent with an empirical, step-wise model building approach, risk factors with p < 0.20 from bivariate analyses were added to the model.²¹ The model was adjusted for the following covariates: age, referral source (self/provider), and total number of risk factors. PrEP interest was not added to the regression model with PrEP uptake as the predictor variable because the relationship between PrEP uptake and PrEP interest is strongly significant and may displace other variables’ effects. Multi-collinearity was examined through the variance inflation factor (cutoff of >=5 for multi-collinearity)²² and statistical model fit was examined through McFadden R².

Results

Population characteristics

A total of three hundred individuals were referred to the parent study, out of which 260 (86.7%) were eligible for PrEP. Of the referred population, 40 (13.3%) were ineligible because they have tested positive for HIV or did not have HIV acquisition risk factors. The racial composition of the eligible population was: White (46.5%), African American (32.3%), Native American (1.5%), multiracial (1.5%), and Asian (0.4%). The race of the remaining eligible population was not documented (17.7%). In terms of ethnicity, about 20.4% (n=53) identified as Latinx/Hispanic. Out of the 260 individuals, 41.2% were referred by a clinical provider and 58.8% self-presented to care. Providers generating the highest rate of referrals were intake nurses (20.0%) and other medical providers (14.2%), which include physicians, nurse practitioners, and physician assistants (Table 1).

Individuals reported an average of two risk factors for HIV infection. The most frequently reported risk factors were having multiple female sex partners (70.4%), inconsistent condom use (46.2%), and having a lifetime history of injection drug use (26.5%) (Table 1). Approximately one fourth (23.9%) had a recent STI history and 18.1% identified as MSM. A minority paid for sex (3.5%) and/or identified as having engaged in sex work themselves (8.5%).

PrEP interest

Of the (N=260) PrEP-eligible men, (n=259) provided an answer to the question about their interest in PrEP. Of these, 50.6% individuals expressed interest in PrEP, 10.4% were non-committal (“possibly”), and 39.0% were uninterested (Table 1). Bivariate analyses showed that identifying as MSM (p = 0.001, Fisher’s Exact Test) and the total number of reported risk factors (t₂₅₇ =−2.91, p=0.004, two-tailed t-test) were significantly associated with interest in PrEP (Table 2). In the multivariable logistic regression model, identifying as MSM (AOR= 4.46, 95% CI [1.86, 11.4]) and/or having multiple female sex partners (AOR= 2.98, 95% CI [1.47, 6.27]) were significantly associated with PrEP interest (Table 3).

Table 2.

Bivariate analyses examining the relationship between demographic factors, HIV risk factors and interest in PrEP (N=259)^*.

	Interested (N=131) n (%)	Not interested (N=128) n (%)	p
Race†
White	60 (53.1)	60 (60.0)	0.253
African American	46 (40.7)	38 (38.0)
Other races or multiple races	7 (6.2)	2 (2.0)
Ethnicity‡
Latinx	33 (27.5)	19 (18.4)	0.151
Not Latinx	87 (72.5)	84 (80.6)
Age
Median [min, max]	31 [18.0, 68.0]	29 [18.0, 57.0]	0.145
Referral Source
Self-referred	71 (54.2)	81 (63.3)	0.175
Clinical provider-referred§	60 (45.8)	47 (36.7)
Intake nurse§	30 (50.0)	22 (46.8)	0.764
MD, NP, PA§	7 (11.7)	5 (10.6)
Public health nurse§	21 (35.0)	16 (34.0)
Substance use treatment provider§	2 (3.3)	4 (8.5)
Injection drug use¶
Within the past 12 months	6 (4.6)	10 (7.8)	0.385
Outside the past 12 months	38 (29.0)	30 (23.4)
Never	87 (66.4)	88 (68.8)
Multiple female sex partners
Yes	99 (75.6)	83 (64.8)	0.080
No	32 (24.4)	45 (35.2)
MSM
Yes	35 (26.7)	12 (9.4)	<0.001
No	96 (73.3)	116 (90.6)
Inconsistent condom use
Yes	62 (47.3)	58 (45.3)	0.841
No	69 (52.7)	70 (54.7)
Recent STI history
Yes	34 (26.0)	27 (21.1)	0.438
No	97 (74.0)	101 (78.9)
Sexual contact with HIV positive person
Yes	2 (1.5)	2 (1.5)	1
No	129 (98.5)	126 (98.4)
Paid for sex
Yes	6 (4.6)	3 (2.3)	0.500
No	125 (95.4)	99 (97.7)
Engaged in sex work
Yes	12 (9.2)	0 (7.8)	0.825
No	119 (90.8)	118 (92.2)
Partner is injection drug user
Yes	2 (1.5)	3 (2.3)	0.681
No	129 (98.5)	125 (97.7)
Shares needles
Yes	6 (4.6)	11 (8.6)	0.218
No	125 (95.4)	117 (91.4)
Total number of risk factors
Mean (SD)	2.31 (1.01)	1.95 (0.98)	0.004

^*One individual who did not report interest was excluded. Percentages presented in this table don’t add up to 100% due to rounding. Some explanatory variables have differing Ns as specified in footnotes.

^†N= 213. Fisher’s Exact Test. Individuals with missing race were excluded.

^‡N= 223. Fisher’s Exact Test. Individuals with missing ethnicity were excluded.

^§N=107. Fisher’s Exact Test.

^¶Those who denied injection drug use were grouped within “never”.

Table 3.

Multivariable binomial logistic regression that examines the association between interest in PrEP and selected risk factors (p<0.20) and covariates among PrEP-eligible individuals (N=259).

	OR (95% CI)	p	McR 2	AOR (95% CI)	p	McR 2
MSM
Yes	3.52 (1.78, 7.43)	<0.001	0.038	4.46 (1.86, 11.4)	0.001	0.073
No	Ref
Multiple female sex partners
Yes	1.68 (0.98, 2.89)	0.060	0.010	2.98 (1.47, 6.27)	0.003
No	Ref
Age (years)	1.02 (0.99, 1.04)	0.147	0.006	1.02 (0.99, 1.05)	0.135
Referral source
Provider	1.46 (0.89, 2.40)	0.138	0.006	1.38 (0.79, 2.44)	0.261
Self	Ref
Total number of risk factors	1.45 (1.13, 1.88)	0.005		0.96 (0.69, 1.34)	0.823
Identified race *
African American	1.21 (0.69, 2.12)	0.503	0.010	-	-	-
Other Races or multi-racial	3.50 (0.81, 24.1)	0.128
White	Ref
Identified ethnicity†
Latinx	1.68 (0.89, 3.22)	0.113	0.008	-	-	-
Not Latinx	Ref

^*N=213. Individuals with missing race were excluded

^†N=223. Individuals with missing ethnicity were excluded

PrEP uptake

Out of the eligible individuals who were interested in PrEP (N=131), a majority (71.8%) enrolled. Out of those who were non-committal and uninterested combined (N=128), a very small minority (4.7%) enrolled (Table 4). Bivariate analyses showed that interest in PrEP was significantly associated with PrEP uptake (p <0.001, Fisher’s Exact Test) (Table 4). The multivariable logistic regression model showed that identifying as MSM (AOR= 2.59, 95% CI [1.06, 6.48]) and having multiple female sex partners (AOR= 2.60, 95% CI [1.14, 6.17]) were significantly associated with PrEP uptake (Yes/No) (Table 5).

Table 4.

Bivariate analyses examining the association between demographic factors and HIV risk factors of eligible individuals with PrEP uptake (N=260).^*

	Enrolled (N=100) n (%)	Not Enrolled (N=160) n (%)	p
Race†
White	50 (52.6)	71 (59.7)	0.106
African American	38 (40.0)	46 (38.7)
Other or multiple races	7 (7.4)	2 (1.7)
Ethnicity‡
Latinx	26 (26.5)	27 (21.4)	0.464
Not Latinx	72 (73.5)	99 (78.6)
Age§
Median [min, max]	30.0 [18.0, 68.0]	30.0 [18.0, 63.0]	0.645
Referral Source
Self-referred	55 (55.0)	98 (61.3)	0.386
Provider-referred	45 (45.0)	62 (38.8)
Intake nurse¶	22 (48.9)	30 (48.4)	0.867
MD, NP, PA¶	17 (37.8)	20 (32.3)
Public health nurse¶	4 (8.9)	8 (12.9)
Substance use treatment provider¶	2 (4.4)	4 (6.5)
Interest #
Yes	94 (94.0)	37 (23.3)	<0.001
No	6 (6.0)	122 (76.7)
Injection drug use
Within the past 12 months	3 (3.0)	13 (8.1)	0.184
Outside the past 12 months	30 (30.0)	39 (24.4)
Never ♠	67 (67.0)	108 (67.5)
Multiple female sex partners
Yes	77 (77.0)	106 (66.3)	0.088
No	23 (23.0)	54 (33.8)
MSM
Yes	24 (24.0)	23 (14.4)	0.072
No	76 (76.0)	137 (85.6)
Inconsistent condom use
Yes	46 (46.0)	74 (46.3)	1
No	54 (54.0)	86 (53.8)
Recent STI history
Yes	27 (27.0)	35 (21.9)	0.427
No	73 (73.0)	125 (78.1)
Sexual contact with HIV positive person
Yes	2 (2.0)	2 (1.3)	0.640
No	98 (98.0)	158 (98.8)
Paid for sex
Yes	4 (4.0)	5 (3.1)	0.737
No	96 (96.0)	155 (96.9)
Engaged in sex work
Yes	10 (10.0)	12 (7.5)	0.634
No	90 (90.0)	148 (92.5)
Partner is injection drug user
Yes	1 (1.0)	4 (2.5)	0.651
No	99 (99.0)	156 (97.5)
Shares needles
Yes	4 (4.0)	13 (8.1)	0.293
No	96 (96.0)	147 (91.9)
Total number of risk factors ♥
Mean (SD)	2.28 (0.99)	2.04 (1.01)	0.058

^*Percentages in this table don’t add up to 100% due to rounding. Some explanatory variables have differing Ns as specified in footnotes.

^†N= 214. Fisher’s Exact Test. Individuals with missing race were excluded.

^‡N= 224. Chi-squared Test. Individuals with missing ethnicity were excluded.

^§N= 259. Two-tailed t-test.

^¶N= 107. Fisher’s Exact Test.

^#N=259. Fisher’s Exact test.

^♠Those who denied injection drug use were grouped within “never”.

^♥N= 260. Two-tailed t-test.

Table 5.

Multivariable binomial logistic regression examining the relationship between PrEP uptake, selected risk factors (p<0.20) and covariates among eligible individuals (N=260)

	OR (95% CI)	p	McR 2	AOR (95% CI)	p	McR 2
MSM
Yes	1.88 (0.99, 3.57)	0.052	0.011	2.59 (1.06, 6.48)	0.049	0.045
No	Ref
Multiple female sex partners
Yes	1.71 (0.97, 3.05)	0.066	0.010	2.60 (1.14, 6.17)	0.026
No	Ref
Injection drug use
Outside of past 12 months	1.24 (0.70, 2.18)	0.456	0.011	1.59 (0.73, 3.51)	0.243
Never	0.37 (0.08, 1.21)	0.134		0.43 (0.09, 1.55)	0.229
Within past 12 months	Ref
Referral source
Provider	1.29 (0.78, 2.15)	0.320	0.003	1.42 (0.81, 2.49)	0.223
Self	Ref
Total number of risk factors	1.27 (0.99, 1.64)	0.061	0.010	1.00 (0.72, 1.39)	0.993
Age (years)	0.99 (0.97, 1.02)	0.648	<0.001	0.99 (0.97, 1.02)	0.675
Identified race*
African American	1.17 (0.67, 2.06)	0.578	0.016	-	-	-
Other races or multi-racial	4.97 (1.15, 34.3)	0.051
White	Ref
Identified ethnicity†
Latinx	1.32 (0.71, 2.46)	0.373	0.003	-	-	-
Not Latinx	Ref

^*N= 214. Individuals with missing race were excluded

^†N= 224. Individuals with missing ethnicity were excluded

Barriers to PrEP uptake

Among PrEP-eligible individuals who did not enroll in the study (N=160), the most frequently encountered barriers to uptake were lack of interest in PrEP (76.3%), low self-perceived risk of HIV acquisition (31.3%), being released before enrolled (25.6%), sentenced beyond study period (16.9%), and being uninterested in the study (13.1%). Among the 49 individuals who indicated low self-perceived risk as an uptake barrier, almost all (98.0%) were not interested in PrEP. Among those whose lack of interest in the study was an uptake barrier (N=21), some expressed that they “did not want to be an experiment” and were not interested in research, some did not want to start a new medication, and one expressed having other health priorities. Other barriers to uptake included concern about side effects (7.5%), having already been prescribed PrEP in the community (5.0%), self-perceived adherence concerns (3.8%) and concerns about additional lab work and/or medication, which was expressed by individuals who did not want to undergo procedures beyond the care they requested to receive (1.9%).

Discussion

This was among the first studies to assess PrEP interest among a cohort of men screened for PrEP eligibility and to determine that PrEP interest and sexual risk factors were strongly associated with PrEP uptake in a correctional setting. Interest in PrEP (50%) among incarcerated men identified to be at-risk of HIV acquisition was notable. Similarly, PrEP uptake (38%) was also high overall, as a majority (76.7%) of those who were interested in PrEP ended up enrolling in the study.

PrEP interest in the study population was higher than that expressed in two earlier studies of men in the RIDOC (23%²³ and 31%²⁴ respectively). Certain studies conducted with MSM in the community yielded higher rates of interest: a study among HIV-negative trans-MSM reported an interest rate of 67.3% ²⁵, while a observational study of young Black MSM (YMSM) reported an interest rate of 63% (N=184) ²⁶. This suggests that while overall interest and uptake is relatively high in the carceral setting, more work can be done to boost PrEP use in this unique environment.

In terms of predictors of PrEP interest, men who reported sexual behaviors that placed them at higher risk of HIV (i.e. identifying as MSM, and/or having multiple female sex partners) were significantly more likely to express interest in PrEP, take steps to enroll in the parent study and initiate PrEP, compared to those not reporting these risk factors. In contrast, injection drug use history was not significantly associated with PrEP interest nor uptake. This is consistent with the extant literature, which has largely demonstrated that PWID have low rates of PrEP interest and uptake.^17,27 This might reflect the lack of adapted PrEP educational interventions and advertisement campaigns for PWIDs as compared to MSM and people with serodiscordant partners.²⁸ A qualitative study in the US exploring perspectives of PWIDs on PrEP revealed that PWIDs associated PrEP with gay men, which could be a potential barrier to PrEP use for individuals who don’t identify as gay and have a low self-perceived risk of HIV acquisition.²⁸

The study’s results highlight numerous barriers at the individual-, interpersonal-, and institutional- level to PrEP uptake among populations experiencing incarceration based on an adapted socio-ecological model, which serves as an organizational framework for contextualizing study findings on barriers to PrEP uptake.^19,29

Individual-level barriers

Despite relatively high levels of PrEP interest, there were other factors that prevented individuals from initiating PrEP. Low self-perceived risk of HIV acquisition was the second-most frequently reported barrier to PrEP uptake, highlighting a discordance between self-perceived risk and clinically assessed risk, which is reflective of the extant literature.^16,30 Low self-perceived risk was also associated with lack of PrEP interest in our sample; conversely, in a previous study in the RIDOC, men who were aware of potential elevated risk behaviors were interested in initiating PrEP prior to release.^24,31s

Anticipated adherence challenges post-release, which appear to be multi-factorial, were barriers to PrEP uptake. The challenges are reflective of the broader literature, such as unstable housing.^18,23 Among incarcerated persons in opioid agonist medication based treatment in the RIDOC, fearing side effects and worries about adhering to the daily regimen were reported barriers to PrEP uptake.²⁴

Interpersonal-level barriers

The interaction between the PrEP provider and men screened for study enrollment took place in a jail/prison setting, which may lack patient privacy and be punitive in nature. This environment likely impacted the interpersonal interaction between provider and patient, as well as the HIV risk assessment process, contributing to lower PrEP interest and uptake. First, these settings are highly structured and include rules that actively prohibit sexual activity and substance use. Second, a number of incarcerated men (6.5%) in the present study denied injection drug use despite track marks or indication of injection drug use in the electronic health record. This indicates that incarcerated men might experience anticipated stigma and discrimination from healthcare providers regarding their substance use, which then become barriers to disclosure of HIV risk and identification of individuals who would be good candidates for PrEP initiation.^32s,33s Furthermore, among incarcerated populations, disclosure of illicit drug use and participating in commercial sex work can be self-incriminating. LGBTQ+ individuals are also at a higher risk of experiencing sexual assault in carceral settings, resulting in a hesitation to disclose HIV acquisition risk behaviors or initiate a preventive therapy that is associated with the LGBTQ+ community.^31s,34s,35s Therefore, clinical providers should consider PrEP as a preventive option for individuals who do not self-report specific HIV-acquisition risk behaviors, but based on their clinical presentation, may be at an elevated risk of HIV acquisition. Additionally, individuals in carceral settings who express an interest in PrEP but do not necessarily report HIV acquisition risk factors, should still be considered for PrEP initiation given the potential hesitancy to disclose behaviors in this setting.

Of note, medical providers in this context cannot offer the same privacy and confidentiality as providers in the community. Instead, they are obligated to report certain activities associated with HIV acquisition risk including sexual contact with others while incarcerated. When offering PrEP in jail/prison settings, it is important to consider how the context can contribute to creating a stigmatizing care environment for PrEP discussions, so HIV preventive care must take a harm-reduction approach and be adapted to the unique administrative environment of these facilities.

Institutional-level barriers

Unpredictable lengths of incarceration for the pre-trial population, and subsequently the inability of the research team to coordinate PrEP clinical care including the completion of laboratory testing with legal proceedings, was a particularly notable barrier to PrEP uptake. More than half of individuals who were released before enrollment and sentenced beyond the study period, respectively, expressed some interest in PrEP, suggesting that institutional factors accounted for a loss of individuals who may have been good candidates for PrEP. In the RIDOC, the average length of stay was 31 days and the median length of stay of the pre-trial population was five days in fiscal year 2022.²⁰ This short median duration of incarceration can limit the ability of medical staff to screen individuals for medical interventions and to ensure that individuals initiate or complete clinical care before their release. Rapid PrEP initiation, such as same-day initiation which is currently recommended by the CDC,² would reduce the number of potential PrEP candidates who return to the community without initiating the medication. Finally, PrEP knowledge among clinical staff, and the ability to prescribe PrEP, plays an important role among the successful and timely initiation of PrEP for appropriate candidates in this unique setting. This was underscored by the referral process generating a notable number of referrals for individuals who either had previously been diagnosed with HIV or did not have any identifiable HIV acquisition risk factors. Having a greater number of providers who are PrEP knowledgable and empowered to prescribe PrEP would also increase the number of individuals who could undergo the clinical processes required for PrEP initiation.

Individuals who were incarcerated beyond the study period but were interested in PrEP could be at an increased risk of HIV acquisition on re-entry. Considering that persons in prisons are less likely to engage in HIV risk behaviors when serving their sentence, and that the CDC PrEP recommendations depend on the timing of HIV risk behaviors, how to effectively identify PrEP candidates among persons with a long sentence pre- release is a research gap to be filled.

In terms of limitations, this study lacked data on gender identity, pre-existing health conditions such as a chronic kidney disease, and socioeconomic factors such as housing status, income and insurance coverage. Collecting such information would be useful for monitoring the outcomes of future PrEP implementation efforts and preventing disparities in PrEP access. In addition, the sample size of racial groups other than White and Black were extremely small, which, although reflective of the general RIDOC population, does limit generalizability of our findings. While median age and the racial and ethnic composition of the enrolled population were similar to the correctional population in the RIDOC,²⁰ we cannot accurately infer the racial and ethnic composition of the non-enrolled population. This is because race and ethnicity data among the non-enrolled population was not consistently self-reported, reflecting one challenge of conducting research in a correctional setting, where participants may decline to share personal information with medical providers. This limited our ability to examine how uptake rates differed across racial and ethnic groups in the RIDOC.

Furthermore, given PrEP was only available during this time period through participation in the study, PrEP use outcomes may differ slightly when offered as part of routine clinical care as a small number of individuals reported not wanting to participate in research as their primary reason for declining PrEP care. Therefore, it is important that PrEP be made part of standard clinical care to facilitate accessibility as well as to understand PrEP use patterns in carceral settings outside of potentially more restrictive, and intrusive, research protocols.

Finally, the setting of the study is in a unified jail and prison system in Rhode Island. Protocols for risk-screening, referral, PrEP initiation and the cost of PrEP implementation may differ significantly from other correctional systems and prisons, limiting the findings’ generalizability. However, given the under-researched nature of PrEP care in corrections, findings on factors that influence PrEP uptake learned from the present study can offer value to other correctional systems.

Conclusions

Among incarcerated men screened for PrEP eligibility, a sizable population met the clinical criteria for PrEP use, were interested in PrEP and agreed to initiate PrEP while incarcerated, with subsequent linkage to PrEP care in the community post-release. This indicates that offering PrEP to men within a jail/pre-trial setting is feasible. However, additional research and implementation strategies are needed to adapt PrEP care to the correctional setting. Future research should seek to understand the best way to leverage factors to increase interest, uptake, and linkage to PrEP care in community settings among individuals experiencing incarceration. Future studies should be designed to allow for meaningful sub-group analyses given the disproportionate burden of HIV and incarceration on racial minority communities and other vulnerable populations. Additional research is also needed to develop evidence-based approaches to facilitate PrEP uptake among individuals who are incarcerated for prolonged periods of time, are interested in PrEP, or who might be good candidates when they return to the community. More specifically, our study revealed that PWIDs may require a more tailored approach given the differential nature of PrEP uptake.

Our study’s findings have many future practice implications, which include: training correctional health providers on taking sexual and substance use histories, and educating patients on a broad range of HIV risk behaviors and potential benefits of taking PrEP. Given that lack of interest in PrEP and low self-perceived risk of HIV were the most frequently encountered barriers to PrEP uptake, various provider-level and behavioral interventions could be deployed to enhance PrEP awareness and uptake. These include re-framing messaging around PrEP candidacy,^36s motivational interviewing,^37s and the “shared decision-making model” wherein providers guide conversations based on HIV indications and educate the patient on the benefits of PrEP.^38s,39s

Given our findings on the impact of perceived stigma based on sexual behaviors and substance use, creating non-stigmatizing clinical care environments in correctional settings will be important for future PrEP implementation efforts. HIV, STI and hepatitis testing strategies should also be adapted to the correctional setting and PrEP should be added to the formulary to facilitate the safe and rapid, or same-day initiation of PrEP. While this study highlights that there was a moderate level of interest in PrEP, as well as a substantial number of individuals who would benefit from initiating PrEP while incarcerated, a commitment from correctional leadership is necessary for effective implementation of PrEP care. Finally, a successful PrEP program must address PrEP adherence challenges especially when individuals are released into the community. Therefore, partnering with local or state departments of health can help correctional systems plan both for the efficient use of resources as well as the development of effective community linkage strategies for individuals on re-entry. Ending the HIV epidemic, particularly among populations impacted by incarceration, will require a renewed commitment and cross-agency collaboration.

Supplementary Material

Supplemental figure 1

Supplemental Figure 1. Barriers to PrEP uptake among men who did not enroll in the study, stratified by PrEP interest.

*Two people did not report any reasons, and 10 people reported more than one reason for not enrolling, therefore the n does not add up to 160. Released before enrolled, to be re-evaluated, and released early, were grouped into one category of “released before enrolled”.