Oral and vaginal HIV pre-exposure prophylaxis product attribute preferences among female sex workers in the Mexico-US border region

Abstract

To assess the potential uptake of HIV pre-exposure prophylaxis (PrEP) products among female sex workers (FSWs) vulnerable to HIV infection, we examined the influence of product attributes on willingness to use products among 271 HIV-negative FSWs in Tijuana and Ciudad Juarez, Mexico (2016–2017). Via five-point Likert scale ratings, participants indicated their willingness to use hypothetical products with six attributes: formulation (pill, gel, liquid, or ring), frequency of use (daily, on-demand, or monthly), cost per use (10 or 200 pesos), effectiveness (40% or 80%), side effects (none or mild), and access point (healthcare clinic or non-governmental organization). Conjoint analysis was used to determine the impact of attributes on product ratings and identify preferred product attributes. Multinomial logistic regression was used to identify factors associated with formulation preferences. In both cities, formulation and frequency of use had the greatest impact on ratings. Participants in Ciudad Juarez indicated a strong preference for oral pills, whereas participants in Tijuana indicated roughly equal preferences for oral pills and vaginal gels. Monthly product use was preferred in both cities. Compared to preferring oral pills (38%), preferring vaginal gels (28%) was associated with practicing vaginal lubrication (adjusted odds ratio = 2.08; 95% confidence interval: 1.07–4.04). Oral PrEP may be acceptable to many FSWs in Tijuana and Ciudad Juarez; however, continued development of behaviorally-congruent vaginal PrEP products may also facilitate uptake and ensure sufficient coverage.

Keywords

HIV Pre-exposure prophylaxis preferences female sex workers Mexico

Introduction

Daily oral tenofovir-emtricitabine (TDF/FTC) is an efficacious HIV prevention strategy,^1–4 recommended for use as pre-exposure prophylaxis (PrEP) by individuals at substantial risk of HIV infection.⁵ Monthly use of dapivirine-containing vaginal rings as PrEP provides moderate protection against HIV infection and is currently under regulatory review.^6,7 In addition, other systemic (e.g. injectable and implantable) and topical (e.g. gels, douches, films, and suppositories applied vaginally or rectally) antiretroviral-based PrEP products are in development. Although one trial documented moderate protection associated with on-demand (i.e. before and after sex) use of tenofovir-containing vaginal gel,⁸ in other trials neither daily nor on-demand gel use were protective.^9,10 However, across all PrEP trials, greater efficacy was observed in the context of more consistent product use.^1–4,6–8 Recent evidence that vaginal microbiota associated with bacterial vaginosis (BV) may metabolize tenofovir, and effectively reduce the efficacy of tenofovir-containing vaginal PrEP products, reinforces the need to maintain high levels of adherence, especially in the presence of BV.^11,12

Outside the United States (US), daily oral PrEP has not been widely implemented.¹³ Yet, despite the US Food and Drug Administration’s approval of Truvada (TDF/FTC) for daily oral use as PrEP in 2012, uptake remains low among women in the US.¹⁴ Implementation of sustained-release, dapivirine-containing vaginal PrEP rings could reduce barriers to uptake associated with oral PrEP’s daily use requirements and potential systemic side effects. However, the continued development of non-systemic, short-acting behaviorally-congruent vaginal PrEP products may also facilitate uptake and adherence among women. Female sex workers (FSWs) are particularly vulnerable to HIV infection with their odds of infection estimated to be 13.5 times those among reproductive-aged women.¹⁵ However, PrEP efficacy trials have largely excluded FSWs.¹⁶ Although intravaginal practices, including washing and lubrication, vary across regions and cultures,^17,18 FSWs often perform intravaginal practices.^19–24 Consistent use of behaviorally-congruent vaginal PrEP gels or douches may be less challenging for FSWs, particularly if they can be integrated into their existing practices.

Several studies have documented high acceptability of hypothetical PrEP products formulated as oral pills,^25–27 vaginal gels,^28,29 and vaginal rings²⁷ among FSWs globally. Conjoint analysis (CJA) is a market-based research technique used to identify consumers’ preferred product attributes and examine the impact of those attributes on consumers’ willingness to use a product.^30–32 Galea et al. employed CJA to explore preferred oral PrEP product attributes among key populations, including FSWs, in Peru, and identified preferences for low-cost, highly efficacious products without side effects.³³ Eisingerich et al. used CJA to examine PrEP product attribute preferences within an international sample of FSWs and found that the route of administration was most important to participants, with injections preferred to oral pills.³⁴ Few studies conducted with FSWs have simultaneously examined preferences between oral and vaginal (gel, liquid, ring) PrEP formulations.

In the present study, Proyecto PrEParado (Project PrEPared), we employed CJA to examine the impact of product attributes on willingness to use oral and vaginal PrEP products and identify preferred product attributes among FSWs in Tijuana and Ciudad Juarez, two northern Mexico border cities. HIV prevalence in Mexico is highest among key populations, including transgender women (17.4%), men who have sex with men (17.3%), sex workers (7.0%), and people who inject drugs (2.5%).³⁵ Along the Mexico-US border, HIV prevalence among FSWs (6%)³⁶ and FSWs who inject drugs (12%)³⁷ is roughly 15 and 30 times that among reproductive-aged women nationally (0.4%),³⁵ respectively. Similar to FSWs globally,³⁸ HIV risk among FSWs along the Mexico-US border is influenced by structural factors that limit their capacity to negotiate condom use with clients, which underscores their need for user-controlled HIV prevention methods, such as PrEP. Daily oral PrEP demonstration projects are currently underway with key populations in Mexico City and are expected to inform the implementation of oral PrEP nationally.¹³ Findings from the present study could expand our understanding of the potential uptake of oral PrEP and inform the continued development and future implementation of other PrEP products acceptable to FSWs in the Mexico-US border region and other similar low-income settings.

Methods

Study population

Proyecto PrEParado was conducted from 2016 to 2017 among FSWs in Tijuana and Ciudad Juarez participating in a randomized controlled trial evaluating a text message-based behavioral HIV prevention intervention. Trial recruitment was conducted by outreach workers at known sex work locations (e.g. bars, street corners). Eligible women were cisgender female, ≥18 years-old, HIV-negative, reported exchanging sex for money, drugs, or goods (past month), reported condom-unprotected vaginal/anal sex with a client (past month), agreed to accept free treatment if they tested STI-positive, and owned a cell phone. Participants provided written informed consent and institutional review boards at the University of California, San Diego, Xochicalco University in Tijuana, and SADEC-FEMAP in Ciudad Juarez approved the study.

Study procedures

Upon enrollment, participants completed baseline surveys and underwent STI testing. Within approximately one month, participants returned for their STI test results and completed Proyecto PrEParado supplemental surveys. All surveys were interviewer-administered in Spanish using computer-assisted personal interviewing.

Baseline surveys (∼50 min) collected information on socio-demographics (age, education, income), substance use (hazardous alcohol consumption [AUDIT score ≥8];³⁹ past month illicit drug use [types; frequency; routes of administration]), sex work characteristics (number of regular/non-regular clients [past month]; frequency of condom-protected/unprotected vaginal/anal intercourse with clients [past month]), and reproductive and sexual health (non-condom contraceptive use [past six months]; vaginal symptoms; pain during vaginal sex).

Proyecto PrEParado supplemental surveys (∼35 min) assessed past month vaginal washing (i.e. washing inside the vagina with commercial solutions, soap and water, or other household products using fingers, cloth, or a douching device to pump the solution inside the vagina) and lubrication (i.e. pushing or placing products, such as creams, oils, or sexual lubricants, inside the vagina using fingers, cloth, or an applicator) practices. Supplemental surveys also implemented CJA methods^30–32 to measure participants’ willingness to use hypothetical PrEP products. Product attributes and attribute levels were selected based on previous PrEP acceptability studies,^33,40,41 including those that employed CJA,^33,42–45 the cost of generic TDF/FTC available for antiretroviral therapy in low- and middle-income countries at the start of the study,^46,47 and side effects reported in trials of oral and vaginal PrEP products,^1–4,6–10 and included: formulation (oral pill, vaginal gel, vaginal liquid, or vaginal ring); frequency of use (daily, on-demand [before and after sex], or monthly); cost per use (10 pesos [∼55 US cents] or 200 pesos [∼11 US dollars]); effectiveness (40% or 80%); side effects (none or mild discomfort, such as nausea, headaches, or vaginal irritation); and access point (healthcare clinic or non-governmental organization [NGO]). Because these attributes could be combined into 192 (4 × 3×2 × 2×2 × 2) distinct hypothetical products, we used specialized algorithms to identify efficient fractional-factorial designs.⁴⁸ The smallest fractional-factorial design that was both balanced (i.e. attribute levels occur equally within each attribute) and orthogonal (i.e. all main effects can be estimated and are uncorrelated) consisted of 24 products. Since asking participants to consider 24 products would be too burdensome, we selected an efficient (D-efficiency = 95.25), balanced, and nearly orthogonal (correlation between formulation and cost, effectiveness, side effects, and access point = 0.33; all other attribute pairs were uncorrelated) fractional-factorial design consisting of 12 products (Table 2). Some of the hypothetical products contain attribute combinations that may not be feasible (i.e. daily or on-demand use of vaginal rings, monthly use of oral pills, vaginal gels, or vaginal liquids). However, the theoretical framework underlying CJA posits that consumers’ willingness to use a product can be decomposed into the specific effects of each product attribute.^31,32 The goal of CJA is not to identify attribute combinations most acceptable to consumers, but rather to understand trade-offs made by consumers with respect to product attributes when presented with products with a range of attributes. Including hypothetical PrEP products with potentially infeasible attribute combinations does not compromise the ability of our CJA to identify preferred product attributes or determine the influence of those attributes on participants’ willingness to use PrEP products.^30–32

Prior to initiating the CJA exercise, interviewers followed a script to inform participants of PrEP’s efficacy, potential side effects, and the need for routine HIV/STI testing and clinical monitoring while using PrEP. The script also defined the attributes under investigation and provided examples to facilitate comprehension. Participants were also shown pictures of potential product formulations. Next, interviewers presented 12 cards to participants in random order, with each card containing a different hypothetical product. Participants were then asked to (1) rank each product from 1 (most likely to use) to 12 (least likely to use) and (2) rate products by indicating their willingness to use products (“Would you be willing to use the PrEP product presented on the card to protect yourself against HIV?”) via five-point Likert scale responses (no, definitely not; probably not; unsure; probably; yes, definitely). Overall, interviewers indicated that participants comprehended the differences between the products presented and completed the CJA exercise with little difficulty in under 10 minutes.

HIV/STI testing procedures

At their enrollment visit, participants underwent HIV/STI testing. Participants who were rapid test positive for HIV (Advance Quality Anti-HIV 1&2; InTec Products, Inc.) and syphilis (Advance Quality Anti-TP; InTec Products, Inc.) underwent confirmatory HIV (Architect HIV Ag/Ab Combo; Abbott; Geenius™ HIV ½ Supplemental Assay; Bio-Rad) or syphilis (rapid plasma reagin [RPR] test: BD Macro-Vue™ RPR; Becton, Dickinson and Company; Treponemal [TP] assay: Architect Syphilis TP assay; Abbott) testing. Participants provided urine samples for the detection of Chlamydia trachomatis and Neisseria gonorrhoeae via nucleic acid amplification testing (Aptima Combo 2® Assay; Hologic). The San Diego County Public Health Laboratory conducted conﬁrmatory Chlamydia and gonorrhea testing, and participants returned for their results within approximately one month. Free STI treatment was offered to STI-positive participants according to Mexican treatment guidelines. HIV-positive women were not eligible to continue, but were referred to municipal clinics for free care.

Statistical analysis

We conducted a CJA to identify preferred levels within product attributes and determine the impact of those attributes on participants’ willingness to use PrEP products. First, five-point Likert scale product ratings reflecting participants’ willingness to use products were converted to a 100-point scale (i.e. 1 recoded as 0, 2 recoded as 25, etc.). Next, separate analysis of variance (ANOVA) models were fit for each participant to model product ratings as a function of the six attributes to obtain participant-specific part-worth utilities (PWUs) and importance values (IVs).⁴⁸ PWUs are equivalent to ANOVA parameter estimates, sum to zero for each attribute, and reflect relative preferences for levels within attributes. A positive PWU for an attribute level indicates a relative preference for that level, while a negative PWU for an attribute level indicates a relative aversion to that level. IVs were calculated for each attribute by dividing the range of PWUs for an attribute by the sum of PWU ranges for all attributes multiplied by 100. IVs sum to 100 across attributes and reflect the relative influence of each attribute on product ratings. Attributes with higher IVs have a greater impact on participants’ willingness to use a product. PWUs and IVs were then averaged across participants by study site. Because familiarity with similar products, existing vaginal practices, and motivations for performing those practices might impact formulation preferences, we examined whether non-condom contraceptive use, vaginal symptoms, pain during vaginal sex, STIs, and vaginal washing and lubrication practices were associated with formulation preferences using separate multinomial logistic regression models adjusting for intervention group assignment, study site, socio-demographics, substance use, and sex work characteristics. In the multinomial logistic regression models, we examined interaction terms between covariates and study site, but because none were statistically significant results are not stratified by study site. Analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC).

Analysis sample

The parent trial enrolled 602 FSWs (302 Tijuana; 300 Ciudad Juarez). Participants who returned for their STI test results and completed the Proyecto PrEParado supplemental survey were considered for inclusion in our analysis. Sample sizes of ∼300 are recommended for CJA.³⁰ Proyecto PrEParado was initiated at the same time as and six months after the initiation of the parent trial in Ciudad Juarez and Tijuana, respectively, and continued until ∼150 participants completed the survey in each city. Seventy-three percent (163/224) and 50% (150/300) of those enrolled in Tijuana and Ciudad Juarez, respectively, after the initiation of Proyecto PrEParado completed the survey. Participants who completed the survey were more likely to report vaginal symptoms (33% vs. 23%; p-value = 0.02) and condom use during a greater proportion of vaginal sex acts with clients (0.6 vs. 0.5; p-value = 0.002), and were less likely to report an average monthly income ≥3500 pesos (70% vs. 84%; p-value = 0.0003). No other statistically significant differences were observed. To ensure the relevance of baseline data to our analysis, we excluded 18 participants who completed the survey more than three months post-baseline. Because CJA cannot be applied to data from participants who rate all products equally, we excluded five participants with no variability in their ratings. Finally, we excluded 19 participants with ANOVA R² values <0.7, which indicates poor model fit suggesting they may not have taken the CJA exercise seriously.⁴⁸ Our analysis sample consisted of 271 FSWs (127 Tijuana; 144 Ciudad Juarez).

Results

Participants had a median age of 38 years (interquartile range [IQR]=30–46) (Table 1). A higher proportion of participants in Ciudad Juarez reported non-condom contraceptive use (81% vs. 15%; p-value < 0.0001), while a higher proportion in Tijuana reported vaginal symptoms (43% vs. 24%; p-value = 0.001), pain during vaginal sex (17% vs. 4%; p-value = 0.001), tested STI-positive (26% vs. 9%; p-value = 0.0002), and reported practicing vaginal lubrication (67% vs. 47%; p-value = 0.001) and washing (58% vs. 22%; p-value < 0.0001). A greater proportion of participants in Tijuana who reported practicing vaginal lubrication did so daily (55% vs. 33%; p-value = 0.01), while a similar proportion of participants reported daily vaginal washing in Tijuana (33%) and Ciudad Juarez (26%) (p-value = 0.64).

Table 1.

Characteristics of HIV-negative FSWs in Tijuana and Ciudad Juarez, Mexico.

	Tijuana (N = 127)		Ciudad Juarez (N = 144)		Total (N = 271)
	n	% or IQR	n	% or IQR	n	% or IQR	p^a
Sociodemographics
Median age (years)	38.0	31.0–48.0	37.0	29.0–43.5	38.0	30.0–46.0	0.18
Completed secondary school	55	43.3	67	46.5	122	45.0	0.59
Average monthly income ≥3500 pesos (∼172 USD)	56	44.1	138	95.8	194	71.6	<0.0001
Sex work characteristics
Median # of clients (past month)	43.0	20.0–75.0	48.0	28.0–77.5	45.0	24.0–75.0	0.11
Median # of regular clients	6.0	4.0–15.0	4.0	2.0–6.5	5.0	3.0–10.0	<0.0001
Median # of non-regular clients	28.0	10.0–60.0	44.0	21.5–70.0	32.0	15.0–65.0	0.0002
Median % of CPVS acts with clients (past month)	0.7	0.5–0.9	0.4	0.2–0.8	0.6	0.3–0.9	<0.0001
Median % of CPVS acts with regular clients	0.6	0.3–0.8	0.0	0.0–0.5	0.4	0.0–0.8	<0.0001
Median % of CPVS acts with non-regular clients	0.9	0.5–1.0	0.6	0.3–0.9	0.8	0.4–1.0	0.0002
Substance use
Past month illicit drug use	76	59.8	90	62.5	166	61.3	0.65
Methamphetamine	68	53.5	12	8.3	80	29.5	<0.0001
Cocaine	7	5.5	71	49.3	78	28.8	<0.0001
Marijuana	35	27.6	32	22.2	67	24.7	0.31
Heroin	32	25.2	26	18.1	58	21.4	0.15
Tranquilizers	17	13.4	27	18.8	44	16.2	0.23
Mexican speedball (methamphetamine + heroin)	16	12.6	0	0.0	16	5.9	<0.0001
Inhalants	3	2.4	6	4.2	9	3.3	0.51
Speedball (cocaine + heroin)	0	0.0	7	4.9	7	2.6	0.02
Ecstasy	3	2.4	2	1.4	5	1.9	0.67
Hazardous alcohol consumption (past 12 months)	47	37.0	77	53.5	124	45.8	0.01
Reproductive/sexual health
Used non-condom contraceptives (past six months)	19	15.0	117	81.3	136	50.2	<0.0001
Tubal ligation	3	15.8	75	64.1	78	57.4
Birth control pill	5	26.3	16	13.7	21	15.4
Intrauterine device	8	42.1	9	7.7	17	12.5
Hormone injection	1	5.3	10	8.6	11	8.1
Other	2	1.6	8	5.6	10	7.4
Current vaginal symptoms^b	54	42.5	35	24.3	89	32.8	0.001
Current pain during vaginal sex	22	17.3	6	4.2	28	10.3	0.001
Tested positive for an STI	33	26.0	13	9.0	46	17.0	0.0002
Chlamydia	23	18.1	11	7.6	34	12.6	0.01
Gonorrhea	15	11.8	2	1.4	17	6.3	0.001
Syphilis (titer ≥1:8)	6	4.7	1	0.7	7	2.6	0.05
Vaginal practices
Practiced vaginal lubrication (past month)	85	66.9	67	46.5	152	56.1	0.001
Practiced vaginal lubrication at least daily^c	47	55.3	22	32.8	69	45.4	0.01
Practiced vaginal washing (past month)	73	57.5	31	21.5	104	38.4	<0.0001
Practiced vaginal washing at least daily^d	24	32.9	8	25.8	32	30.8	0.64

Note: Numbers may not sum to column total due to missing data; percentages may not sum to 100 due to rounding or omission of one category for binary variables.

CPVS: condom-protected vaginal sex; FSW: female sex worker; IQR: interquartile range; STI: sexually transmitted infection; USD: United States dollars.

^aP-value from chi-squared, Fisher's exact, or Wilcoxon rank-sum test.

^bVaginal symptoms = burning or pain with urination, vaginal discharge, vaginal bleeding between periods, pelvic pain.

^cAmong those who practiced vaginal lubrication in the past month.

^dAmong those who practiced vaginal washing in the past month.

Table 2.

Hypothetical PrEP product ratings among FSWs in Tijuana and Ciudad Juarez, Mexico (N = 271).

	Mean product rating (SD)		Product attributes
Hypothetical PrEP product	Tijuana	Ciudad Juarez	Formulation^a	Frequency of use	Cost per use	Effectiveness, %	Side effects	Access point
1	77.56 (27.79)	76.91 (29.58)	Gel	On-demand	10 pesos	80	Mild	Clinic
2	71.85 (31.50)	59.72 (33.33)	Liquid	Monthly	200 pesos	80	None	Clinic
3	65.16 (34.53)	56.42 (35.08)	Ring	Daily	10 pesos	80	None	NGO
4	59.45 (38.70)	58.85 (39.05)	Pill	On-demand	200 pesos	80	None	NGO
5	57.87 (36.28)	52.60 (35.69)	Pill	Monthly	200 pesos	40	Mild	NGO
6	71.06 (35.69)	53.65 (35.23)	Ring	Monthly	10 pesos	80	Mild	Clinic
7	37.20 (34.05)	26.39 (30.26)	Ring	On-demand	200 pesos	40	None	Clinic
8	49.80 (34.14)	43.40 (31.01)	Liquid	On-demand	10 pesos	40	Mild	NGO
9	61.61 (34.90)	46.53 (31.79)	Gel	Monthly	10 pesos	40	None	NGO
10	49.41 (34.43)	32.47 (30.88)	Gel	Daily	200 pesos	40	Mild	Clinic
11	59.06 (36.96)	49.65 (29.71)	Liquid	Daily	200 pesos	80	Mild	NGO
12	57.87 (36.01)	53.82 (34.27)	Pill	Daily	10 pesos	40	None	Clinic

FSW: female sex worker; NGO: non-governmental organization; PrEP: pre-exposure prophylaxis; SD: standard deviation; on-demand: before and after sex; mild: mild discomfort, such as nausea, headaches, or vaginal irritation.

^aFormulation = oral pill, vaginal gel, vaginal liquid, or vaginal ring.

Most participants (94% Tijuana; 99% Ciudad Juarez) reported that they would probably/definitely use their highest ranked product (rank = 1), whereas few participants (15% Tijuana; 1% Ciudad Juarez) reported that they would probably/definitely use their lowest ranked product (rank = 12). Product 1 (vaginal gel, on-demand use, 10 pesos per use, 80% effective, mild side effects, accessible at healthcare clinics) received the highest mean rating in Tijuana (77.56; standard deviation [SD]=27.79) and Ciudad Juarez (76.91; SD = 29.58) (Table 2). In Tijuana, formulation had the greatest influence on product ratings (mean IV = 31.75, SD = 15.92) followed by frequency of use (mean IV = 20.40, SD = 11.62), effectiveness (mean IV = 17.24, SD = 15.14), and cost per use (mean IV = 12.22, SD = 10.49). In Ciudad Juarez, formulation (mean IV = 31.67, SD = 15.29) and frequency of use (mean IV = 18.28, SD = 9.79) also had the greatest influence on product ratings; however, the impact of effectiveness (mean IV = 16.60, SD = 14.85) and cost per use (mean IV = 16.71, SD = 13.11) on product ratings were approximately equal. Access point (Tijuana: mean IV = 9.75, SD = 7.60; Ciudad Juarez: mean IV = 8.71, SD = 6.00) and side effects (Tijuana: mean IV = 8.64, SD = 7.25; Ciudad Juarez: mean IV = 8.03, SD = 6.43) had the least influence on product ratings in both cities.

Participants in Ciudad Juarez indicated a strong preference for oral pills (mean PWU = 10.10, SD = 19.82) and a moderate preference for vaginal gels (mean PWU = 2.14, SD = 18.66) (Figure 1), whereas participants in Tijuana indicated roughly equal preferences for oral pills (mean PWU = 3.65, SD = 21.39) and vaginal gels (mean PWU = 3.81, SD = 16.32) (Figure 2). Participants in both Ciudad Juarez (mean PWU=–10.97, SD = 22.73) and Tijuana (mean PWU=–6.80, SD = 21.75) indicated a strong aversion to vaginal rings. Participants in both cities indicated strong preferences for products that are 80% effective (vs. 40%), moderate preferences for products that cost 10 pesos per use (vs. 200) and can be used monthly (vs. daily or on-demand), and slight preferences for products that can be accessed at healthcare clinics (vs. NGOs), but no obvious preferences (mean PWUs < 0.50) regarding side effects.

Figure 1.

Average part-worth utilities for attributes of hypothetical PrEP products among female sex workers in Ciudad Juarez, Mexico.

Figure 2.

Average part-worth utilities for attributes of hypothetical PrEP products among female sex workers in Tijuana, Mexico.

Among participants whose PWUs indicated a preference for one product formulation (N = 268), 38%, 28%, 18%, and 17% preferred pills, gels, rings, and liquids, respectively. Compared to preferring pills, preferring gels was associated with reporting pain during vaginal sex (adjusted odds ratio [AOR]=3.51, 95% confidence interval [CI]: 1.13–10.84) and vaginal lubrication practices (AOR = 2.08, 95% CI: 1.07–4.04) (Table 3).

Table 3.

Covariates associated with PrEP product formulation preferences among HIV-negative FSWs in Tijuana and Ciudad Juarez, Mexico (N = 268).

	Preferred PrEP product formulation^a
	Vaginal gel (N = 74)		Vaginal ring (N = 47)		Vaginal liquid (N = 45)
Covariates of interest^b	AOR	95% CI	AOR	95% CI	AOR	95% CI
Used non-condom contraceptives (past six months)
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	1.49	0.63, 3.55	1.02	0.39, 2.68	1.40	0.52, 3.79
Current vaginal symptoms^c
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	1.11	0.56, 2.18	0.69	0.31, 1.54	1.09	0.49, 2.39
Current pain during vaginal sex
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	3.51	1.13, 10.84	2.38	0.65, 8.69	1.65	0.40, 6.84
Tested positive for an STI
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	0.55	0.22, 1.37	0.33	0.11, 1.04	1.03	0.40, 2.65
Practiced vaginal lubrication (past month)
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	2.08	1.07, 4.04	0.77	0.37, 1.61	0.90	0.43, 1.89
Practiced vaginal washing (past month)
No	1.00	Ref	1.00	Ref	1.00	Ref
Yes	1.12	0.54, 2.29	1.81	0.79, 4.13	0.98	0.42, 2.30

AOR: adjusted odds ratio; CI: confidence interval; FSW: female sex worker; IQR: interquartile range; PrEP: pre-exposure prophylaxis.

^aMultinomial logistic regression reference group = female sex workers who prefer PrEP products formulated as oral pills (N = 102).

^bThe relationship between each covariate of interest and preferred PrEP product formulation (outcome) was examined in a separate multinomial logistic regression model adjusted for intervention group assignment, study site, socio-demographics (age, education, average monthly income), substance use (illicit drug use in the past month, hazardous alcohol consumption), and sex work characteristics (number of clients in the past month, percent of vaginal sex acts with clients in the past month that were condom-protected).

^cVaginal symptoms = burning or pain with urination, vaginal discharge, vaginal bleeding between periods, pelvic pain.

Discussion

We used CJA to investigate willingness to use hypothetical oral and vaginal PrEP products among FSWs in Tijuana and Ciudad Juarez, Mexico. Consistent with previous research suggesting high PrEP acceptability among FSWs,^25–29 nearly all participants were interested in using their highest-ranked PrEP product. Our findings build on previous research by providing insight on FSWs’ preferred PrEP product attributes that may affect uptake of oral PrEP and have implications for the continued development of vaginal PrEP rings, gels, and douches.

Product formulation had the greatest influence on overall product ratings among participants in both Tijuana and Ciudad Juarez suggesting that formulation preferences may drive future PrEP use within this population. Preferences for oral PrEP relative to vaginal products among participants in Ciudad Juarez and some participants in Tijuana indicate that uptake of oral PrEP products, which are currently being brought to scale globally, may be high among FSWs in the Mexico-US border region. This finding is consistent with preferences among women participating in a cross-over trial in the US whose preferences were largely explained by the widespread availability and use of medications formulated as oral pills in the US.⁴¹ Given the proximity of Tijuana and Ciudad Juarez to the US, preferences for oral pills within our sample may similarly be explained by FSWs’ familiarity with oral pill medications in the Mexico-US border region.^41,49,50 Preferences for PrEP formulated as vaginal gels among participants in Tijuana may be explained by a desire for more pleasurable sexual encounters as suggested by the relationship between preferring PrEP formulated as vaginal gels and reporting pain during vaginal sex and vaginal lubrication practices, both of which were reported by more Tijuana participants. As such, some FSWs may recognize vaginal PrEP gels as a behaviorally congruent HIV prevention strategy that can be easily incorporated into their existing vaginal practices to enhance lubrication and comfort during sex, which is consistent with research suggesting a relationship between increased sexual pleasure and vaginal PrEP gel acceptability.^51,52 The relatively strong aversion to vaginal PrEP rings may reflect infrequent use of contraceptive vaginal rings by women in Mexico. Nevertheless, no single product formulation was overwhelmingly preferred within our sample, with each formulation preferred by 17% to 38% of participants. Thus, our findings indicate a need for multiple PrEP product formulations, including behaviorally-congruent vaginal PrEP gels and douches, to facilitate uptake and ensure sufficient PrEP coverage among FSWs.

In Tijuana and Ciudad Juarez, frequency of use had the second greatest influence on PrEP product ratings. Participants preferred monthly product use to daily and on-demand use, which they may have perceived as less burdensome and easier to adhere to.^33,49 Of the formulations considered, only vaginal rings are being developed for monthly use. As such, uptake of daily oral PrEP may be low among FSWs who prefer products that can be used monthly. Given the aversion toward vaginal rings in our sample, future research should examine the acceptability of other sustained-release products, which require less frequent use. Although access point and side effects had little impact on product ratings within our sample, the systemic nature of other sustained-release products (e.g. injectable and implantable), which will require regular clinic visits and may have side effects that cannot be as easily reversed as those associated with non-systemic, short-acting products, may enhance the acceptability of more frequent PrEP product use. More specifically, because 31% and 45% of participants who reported vaginal washing and lubrication in the past month, respectively, did so at least daily, more frequent dosing may be acceptable for non-systemic, short-acting behaviorally-congruent vaginal PrEP gels or douches that can be integrated into FSWs’ existing practices.

While efficacy and cost influenced PrEP product ratings, they were not as important to FSWs in our sample as they were to participants from Peruvian key populations in Galea et al.’s study.³³ While the influence of product attributes (i.e. IVs) can be overestimated for attributes with more levels, IVs and PWUs reflect the influence of attributes and attribute preference, respectively, relative to the other attributes considered, as such findings across CJAs considering different attributes are not comparable.³¹ Thus, these discrepancies are likely explained by varying attributes in each study.

Our study has several limitations. First, non-random sampling may limit the generalizability of our findings. Second, due to social desirability bias, participants may have overstated their interest in using the PrEP products presented. Third, although participants were informed of the potential side effects of PrEP use and the need for regular clinical and laboratory monitoring for their occurrence, no distinction was made between systemic, non-systemic, long-acting, and short-acting products with respect to the nature of side effects or the necessary frequency of monitoring for those side effects. As such, our findings do not account for the potential impact of clinical and laboratory monitoring on willingness to use PrEP products. Moreover, characteristics of PrEP delivery programs (e.g. location and frequency of HIV/STI testing) could also affect PrEP product uptake. To the extent that this monitoring and these characteristics are unacceptable relative to other product attributes, our findings may overestimate FSWs’ willingness to use oral and vaginal PrEP products. Fourth, while CJA is useful for assessing preferred product attributes, qualitative research is needed to better understand attribute preferences. Fifth, while it is possible that FSWs could use vaginal PrEP products rectally, questions related to topical products (gel, liquid, ring) only referred to their vaginal use. Research examining this possibility is needed as it has implications for the development of multi-compartment (i.e. vagina and rectum) PrEP products. Finally, uptake may differ from hypothetical willingness to use PrEP.

In one of the first studies to examine willingness to use oral and vaginal PrEP products among FSWs in Mexico, we identified a strong interest in the use of these user-controlled HIV prevention strategies. Our findings provide important information regarding PrEP product attribute preferences and their relative impact on potential product uptake. Implementation of various PrEP product formulations, including behaviorally-congruent vaginal gels and douches, may enhance future PrEP uptake and adherence, and ultimately reduce the burden of HIV among FSWs.

Footnotes

Acknowledgements

The authors would like to thank the study participants and staff without whom this study would not have been possible.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institutes of Health, grants R01 DA039071 (TLP), K01 DA040543 (HAP), P30 AI036214 (UCSD CFAR Developmental Grant to HAP), and P30 AI094189 (Johns Hopkins University CFAR). SAS was supported by NIDA MERIT Award R37 DA019829. The funders had no role in the design of the study, collection, analysis, and interpretation of data, or manuscript preparation.

References

Baeten

Donnell

Ndase

et al . Antiretroviral prophylaxis for HIV prevention in heterosexual men and women. N Engl J Med 2012; 367: 399–410.

Choopanya

Martin

Suntharasamai

et al . Antiretroviral prophylaxis for HIV infection in injecting drug users in Bangkok, Thailand (the Bangkok Tenofovir Study): a randomised, double-blind, placebo-controlled phase 3 trial. Lancet 2013; 381: 2083–2090.

Grant

Lama

Anderson

et al . Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med 2010; 363: 2587–2599.

Thigpen

Kebaabetswe

Paxton

et al . Antiretroviral preexposure prophylaxis for heterosexual HIV transmission in Botswana. N Engl J Med 2012; 367: 423–434.

World Health Organization. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection: recommendations for a public health approach. 2nd ed., http://apps.who.int/iris/bitstream/10665/208825/1/9789241549684_eng.pdf?ua=1 (2016, accessed 2 February 2018).

Baeten

Palanee-Phillips

Brown

et al . Use of a vaginal ring containing dapivirine for HIV-1 prevention in women. N Engl J Med 2016; 375: 2121–2132.

Nel

van Niekerk

Kapiga

et al . Safety and efficacy of a dapivirine vaginal ring for HIV prevention in women. N Engl J Med 2016; 375: 2133–2143.

Abdool Karim

Frohlich

et al . Effectiveness and safety of tenofovir gel, an antiretroviral microbicide, for the prevention of HIV infection in women. Science 2010; 329: 1168–1174.

Marrazzo

Ramjee

Richardson

et al . Tenofovir-based preexposure prophylaxis for HIV infection among African women. N Engl J Med 2015; 372: 509–518.

10.

Rees

Delany-Moretlwe

Lombard

et al . (eds) FACTS 001 Phase III trial of pericoital tenofovir 1% gel for HIV prevention in women. Abstract number: 26LB. In: Conference on retroviruses and opportunistic infections, February 23–26, 2015, Seattle, WA.

11.

Klatt

Cheu

Birse

et al . Vaginal bacteria modify HIV tenofovir microbicide efficacy in African women. Science 2017; 356: 938–945.

12.

Hillier

Meyn

Bunge

et al . (eds) Impact of vaginal microbiota on genital tissue and plasma concentrations of tenofovir. Abstract number: 86LB. In: Conference on retroviruses and opportunistic infections, February 13–16, 2017; Seattle, WA.

13.

Ravasi

Grinsztejn

Baruch

et al . Towards a fair consideration of PrEP as part of combination HIV prevention in Latin America. J Int AIDS Soc 2016; 19: 21113.

14.

Mendoza

Huang

et al . Uptake of HIV preexposure prophylaxis among commercially insured persons-United States, 2010–2014. Clin Infect Dis 2017; 64: 144–149.

15.

Baral

Beyrer

Muessig

et al . Burden of HIV among female sex workers in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis 2012; 12: 538–549.

16.

Bekker

Johnson

Cowan

et al . Combination HIV prevention for female sex workers: what is the evidence? Lancet 2015; 385: 72–87.

17.

Braunstein

Van de Wijgert

Preferences and practices related to vaginal lubrication: implications for microbicide acceptability and clinical testing. J Womens Health (Larchmt) 2005; 14: 424–433.

18.

Hull

Hilber

Chersich

et al . Prevalence, motivations, and adverse effects of vaginal practices in Africa and Asia: findings from a multicountry household survey. J Womens Health (Larchmt) 2011; 20: 1097–1109.

19.

Bui

Tran

Ross

et al . Douching practices among female sex workers in Phnom Penh, Cambodia. Int J STD Aids 2015; 26: 238–242.

20.

Francis

Baisley

Lees

et al . Vaginal practices among women at high risk of HIV infection in Uganda and Tanzania: recorded behaviour from a daily pictorial diary. PLoS One 2013; 8: e59085.

21.

Hassan

Lavreys

Chohan

et al . Associations between intravaginal practices and bacterial vaginosis in Kenyan female sex workers without symptoms of vaginal infections. Sex Transm Dis 2007; 34: 384–388.

22.

Priddy

Wakasiaka

Hoang

et al . Anal sex, vaginal practices, and HIV incidence in female sex workers in urban Kenya: implications for the development of intravaginal HIV prevention methods. AIDS Res Hum Retroviruses 2011; 27: 1067–1072.

23.

Seidman

Rusch

Abramovitz

et al . Intravaginal practices among HIV-negative female sex workers along the US-Mexico border and their implications for emerging HIV prevention interventions. Int J Gynaecol Obstet 2016; 133: 212–216.

24.

Wang

Stanton

et al . Vaginal douching, condom use, and sexually transmitted infections among Chinese female sex workers. Sex Transm Dis 2005; 32: 696–702.

25.

Wei

Zou

et al . HIV pre-exposure prophylaxis interest among female sex workers in Guangxi, China. PLoS One 2014; 9: e86200.

26.

Reza-Paul

Lazarus

Doshi

et al . Prioritizing risk in preparation for a demonstration project: a mixed methods feasibility study of oral pre-exposure prophylaxis (PREP) among female sex workers in South India. PLoS One 2016; 11: e0166889.

27.

Peitzmeier

Tomko

Wingo

et al . Acceptability of microbicidal vaginal rings and oral pre-exposure prophylaxis for HIV prevention among female sex workers in a high-prevalence US city. AIDS Care 2017; 29: 1453–1457.

28.

Wang

Jiang

et al . Microbicide acceptability and associated factors among female sex workers and male clients in Kaiyuan county, Yunnan province, China. J Acquir Immune Defic Syndr 2010; 53: S93–S97.

29.

Robertson

Syvertsen

Martinez

et al . Acceptability of vaginal microbicides among female sex workers and their intimate male partners in two Mexico-US border cities: a mixed methods analysis. Glob Public Health 2013; 8: 619–633.

30.

Orme

Getting started with conjoint analysis: strategies for product design and pricing research. 2 ed. Madison: Research Publishers LLC, 2010.

31.

Rao

VR.

Applied conjoint analysis. Heidelberg, Germany: Springer, 2014.

32.

Bridges

Hauber

Marshall

et al . Conjoint analysis applications in health – a checklist: a report of the ISPOR Good Research Practices for Conjoint Analysis Task Force. Value Health 2011; 14: 403–413.

33.

Galea

Kinsler

Salazar

et al . Acceptability of pre-exposure prophylaxis as an HIV prevention strategy: barriers and facilitators to pre-exposure prophylaxis uptake among at-risk Peruvian populations. Int J STD Aids 2011; 22: 256–262.

34.

Eisingerich

Wheelock

Gomez

et al . Attitudes and acceptance of oral and parenteral HIV preexposure prophylaxis among potential user groups: a multinational study. PLoS One 2012; 7: e28238.

35.

UNAIDS. Country: Mexico, www.unaids.org/en/regionscountries/countries/mexico (2016, accessed 17 June 2018).

36.

Patterson

Semple

Staines

et al . Prevalence and correlates of HIV infection among female sex workers in 2 Mexico-US border cities. J Infect Dis 2008; 197: 728–732.

37.

Strathdee

Philbin

Semple

et al . Correlates of injection drug use among female sex workers in two Mexico-U.S. border cities. Drug Alcohol Depend 2008; 92: 132–140.

38.

Shannon

Strathdee

Goldenberg

et al . Global epidemiology of HIV among female sex workers: influence of structural determinants. Lancet 2015; 385: 55–71.

39.

Saunders

Aasland

Babor

et al . Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons with Harmful Alcohol Consumption – II. Addiction 1993; 88: 791–804.

40.

Cook

Downs

Marrazzo

et al . Preferred characteristics of vaginal microbicides in women with bacterial vaginosis. J Womens Health (Larchmt) 2009; 18: 1163–1167.

41.

Minnis

Gandham

Richardson

et al . Adherence and acceptability in MTN 001: a randomized cross-over trial of daily oral and topical tenofovir for HIV prevention in women. AIDS Behav 2013; 17: 737–747.

42.

Best

Tanner

Hensel

et al . Young women's contraceptive microbicide preferences: associations with contraceptive behavior and sexual relationship characteristics. Perspect Sex Reprod Health 2014; 46: 15–22.

43.

Kinsler

Cunningham

Nurena

et al . Using conjoint analysis to measure the acceptability of rectal microbicides among men who have sex with men in four South American cities. AIDS Behav 2012; 16: 1436–1447.

44.

Tang

Galea

Kinsler

et al . Using conjoint analysis to determine the impact of product and user characteristics on acceptability of rectal microbicides for HIV prevention among Peruvian men who have sex with men. Sex Transm Infect 2016; 92: 200–205.

45.

Tanner

Katzenstein

Zimet

et al . Vaginal microbicide preferences among midwestern urban adolescent women. J Adolesc Health 2008; 43: 349–356.

46.

Caceres

Borquez

Klausner

et al . Implementation of pre-exposure prophylaxis for human immunodeficiency virus infection: progress and emerging issues in research and policy. J Int AIDS Soc 2016; 19: 21108.

47.

UNAIDS. Oral pre-exposure prophylaxis: putting a new choice in context, www.unaids.org/en/resources/documents/2015/Oral_pre-exposure_prophylaxis_putting_a_new_choice_in_context. (2015, accessed 28 April 2018).

48.

Kuhfeld WF. Marketing research methods in SAS: experimental design, choice, conjoint, and graphical techniques, MR-2010, http://support.sas.com/techsup/technote/mr2010.pdf (2010, accessed 15 February 2018).

49.

Van der Elst

Mbogua

Operario

et al . High acceptability of HIV pre-exposure prophylaxis but challenges in adherence and use: qualitative insights from a phase I trial of intermittent and daily PrEP in at-risk populations in Kenya. AIDS Behav 2013; 17: 2162–2172.

50.

van der Straten

Stadler

Luecke

et al . Perspectives on use of oral and vaginal antiretrovirals for HIV prevention: the VOICE-C qualitative study in Johannesburg, South Africa. J Int AIDS Soc 2014; 17: 19146.

51.

Hoffman

Morrow

Mantell

et al . Covert use, vaginal lubrication, and sexual pleasure: a qualitative study of urban U.S. Women in a vaginal microbicide clinical trial. Arch Sex Behav 2010; 39: 748–760.

52.

Woodsong

Alleman

Sexual pleasure, gender power and microbicide acceptability in Zimbabwe and Malawi. AIDS Educ Prev 2008; 20: 171–187.