Measuring PrEP Preferences Among U.S. Military Men Who Have Sex with Men: Results of an Adaptive Choice Based Conjoint Analysis Study


 BACKGROUND. Pre-exposure prophylaxis (PrEP) effectively prevents HIV infection, yet its uptake remains low among U.S. military men who have sex with men (MSM). Research shows that health services matching preferences produce favorable outcomes. Therefore, an early step for planning program design is to characterize and identify preferences for improved PrEP delivery within this population. METHODS. HIV-negative military MSM were recruited through a closed, LGBT military social media group. Participants completed an anonymous survey presenting five experimentally varied attributes of interest related to PrEP administration (dosing method, provider type, visit location, lab work evaluation location, and dispensing venue). Relative importance and part-worth utility scores were generated using Hierarchical Bayes (HB) estimation, and the randomized first choice model was used to examine participation interest across eight possible PrEP program scenarios.RESULTS. Among the 429 participants and the eight scenarios that varied the five attributes into delivery profiles, the most preferred scenarios featured a daily tablet, PrEP injection, or implant, along with a military provider, smartphone/telehealth visit, and on-base locations for lab evaluation and medication pick-up. Responses emphasized the importance for providers to be familiar with PrEP prescription knowledge, and to provide interactions sensitive to sexual identity and mental health. CONCLUSION: These results suggest that a daily tablet PrEP program within a military setting is preferred over a civilian or off-site one, yet more importantly, it points to a high value placed on the quality of clinical interactions. High interest in long-acting implants and injections also suggest potential markets for future PrEP engagement. TRIAL REGISTRATION. Not applicable.


Background
Each year, there are approximately 350 new cases of human immunode ciency virus (HIV) infections within members of the U.S. military; with those most affected being younger, Black, and men who have sex with men. (1)(2)(3)(4)(5)(6) When taken correctly, Pre-Exposure Prophylaxis (PrEP) effectively prevents HIV infection, (7)(8)(9)(10) yet the current PrEP cascade within the U.S. military suggests sub-optimal uptake. With an estimated 16% of at-risk military members taking PrEP and members of color remaining underrepresented in terms of uptake, (2) identifying the factors that drive uptake within this population remains a priority.
The eld of PrEP science has explored multiple avenues of PrEP access pathways, delivery mechanisms, and dosing methods to circumvent the geographic, psychological, and adherence barriers that exist and impede PrEP engagement. (11)(12)(13)(14)(15)(16)(17)(18) Currently, PrEP availability to military members largely depends on geographic proximity to a large medical facility with specialty services, as evidenced by 41% of all military PrEP prescriptions originating from military medical centers in three locations in the U.S.
Additionally, 60% of all military PrEP prescriptions also occur only after consultation with an infectious disease specialist. (2) Data are sparse on military MSM's most preferred program characteristics within a PrEP delivery program.
Health services designed around preferences in terms of product type, delivery method, and location settings have been shown to produce improved treatment outcomes and retention in care. (19)(20)(21) Stated preference methods, such as conjoint analysis, quantify preference data of new market entrants and product characteristics, also called attributes. (16,17,(22)(23)(24)(25)(26) The central theory of conjoint analysis is that products or programs are viewed as a composition of various attributes that possess a certain amount of value (part-worth utility score) determined by preference. By quantifying these part-worth utility scores for preferred attributes, these scores can then be entered into market simulation models to predict how respondents might respond to any potential combination of attribute levels. (16,17,(22)(23)(24)(25)(26) Using conjoint analysis, the purpose of this study is to identify the preferred attributes that are most in uential to at-risk U.S. military MSM's decision to take PrEP within the military healthcare system.

Methods
A convenience sample of self-reported HIV-negative, U.S. military MSM and trans-individuals were recruited between March and April 2020 through a closed Facebook group with an internal membership of over 7,000 LGBT U.S. military members. The group administrators placed monthly advertisements describing the study on the group's main forum. Those interested could click on a link to access an online study disclosure form with a 'click to consent' procedure. An option to provide an e-mail address that was not linked to survey responses was offered to participants who opted to receive $5 compensation for questionnaire completion. The study was approved by the Yale University Institutional Review Board.
To collect and quantify respondent preference data, an adaptive choice-based conjoint (ACBC) survey instrument was developed based on a starting set of PrEP program attributes resulting from review of the literature of previous PrEP preference conjoint experiments, which were then re ned through in-depth, qualitative interviews from PrEP experts and U.S. military MSM. (2-5, 11, 12, 14, 15, 27-39) With a focus on modi able PrEP program characteristics, the nal survey design was composed of ve different PrEP program delivery attributes of interest that included: dosing method (daily oral tablet, on-demand tablet regimen [two tablets before sex, one tablet for two days after], rectal douche [before sex], injection [every 2 months], implant [once a year]), provider type (military, civilian), visit location (on-base, off-base, smartphone app), dispensing venue (on-base, off-base, mail delivery), and lab evaluation (on-base, offbase, home-based mail kit). The survey was piloted by the author (JG) with a convenience sample of eleven military MSM members within the targeted social media group for concept testing, and the descriptions and wording of three attribute categories and two attribute level choices were revised for clari cation based on feedback. Figure 1 shows a sample item of the conjoint survey, and Table 1 describes the program attributes presented within the survey. Additionally, we collected demographical data to include age, race, ethnicity, rank type (o cer, enlisted or warrant o cer), military branch, geographic region, PrEP experience ("Have you ever used PrEP [Pre-Exposure Prophylaxis]?"), depressive symptoms with the Patient Health Questionnaire-2 (PHQ2),(40, 41) and the HIV Incidence Risk Index for MSM (HIRI-MSM).(42) Measures to explore levels of satisfaction with a current level of HIV protection and disclosure discomfort within interactions with a primary care provider were also collected.

Analysis
The nal survey instrument was loaded into Lighthouse Studio 9, and an experimental design module was used to pre-test the design with 500 simulated respondents for optimal choice task con guration. The nal design produced a survey where each level within an attribute was seen at least three times per respondent; achieving a high degree of precision at the individual level with a standard of error of <0.03 and all e ciencies reporting at 1.00.(43) Table 2 displays the CONSORT diagram of respondent enrollment and exclusion. To ensure the integrity of the data and eliminate random or duplicate responders, security features within the Sawtooth software and servers recognize returning study participants through the use of internet browser cookies and IP addresses. It also prevents repeated or duplicate attempts to retake the survey. (44) Additionally, as extensive pilot testing required at least 10 to 15 minutes, responses completed in less than 10 minutes (or if a respondent selected the same answer for all items) were excluded. Furthermore, the root likelihood (RLH) t statistic for each respondent was analyzed to evaluate within-respondent choice consistency. RLH, which has a probability value from 0 to 1.0, was used to discriminate between respondents who answered choice-questions consistently or randomly. (45) The survey design was tested by 1,000 computer-generated mock respondents to determine the median RLH for 'random responders' at the 95% percentile (0.5178 RLH). Survey respondents with an RLH below this score were excluded, as the inclusion of 'random responders' can affect the calculation of preference scores and participation rates. (45) For conjoint analyses, the Hierarchical Bayes (HB) procedure was used to estimate part-worth utility scores (PWUS) on an individual level for its accuracy and e ciency, (46,47) and was used to analyze the PWUS of the aggregated sample across all 16 attribute levels. The resulting PWUS of the levels under each attribute category are zero-centered; meaning that the sum of the level scores under each attribute category equal to zero. Scores that are further away from zero (0) indicate a stronger positive or negative preference for the level choice in relation to the other level choices under the same attribute.(39, 43, 47) After identifying each attribute level PWUS, the attribute relative importance scores (RIS) can then be calculated to characterize the magnitude of in uence that each attribute category has on the respondents preference decision-making. The RIS for this study was calculated by dividing the range of PWUS for levels under each attribute by the sum of the ranges, and then multiplying by 100. (48,49) Therefore, if an attribute RIS is 45%, then this means that 45% of an individual's decision making for program engagement will be in uenced by preferences within that attribute category. The PWUS were then used to predict the share of preference (participation interest) among eight hypothetical PrEP program scenarios. PrEP program scenarios were con gured after a variety of currently available or hypothetical PrEP program models. For this study, participation rates for these PrEP scenarios were generated using the randomized rst choice model; in which PWUS are summed across the levels corresponding to each option, and then exponentiated and rescaled, so they sum to 100. (48,49) This approach is based on the assumption that respondents or consumers will prefer a product with the highest composite utility (or value) adjusting for both attribute and program variability. (48) The randomized choice model accounts for variation in each participant's total utility for each option and error in point estimates of the utility, and has been shown to have better predictive ability than other shares of preference models. (49) All data analyses were performed using XLSTAT and Sawtooth Lighthouse Studio 9.0. Table 3 shows the descriptive statistics of the 429 respondents that met the required elapsed survey time and RLH consistency cut-off; strati ed by PrEP experience. Overall, the mean age was 30 years old. Participants mostly identi ed as white (72%), cis-gendered male (96.7%), of o cer rank (46.4%), and membership within the Army branch (48.7%). Additionally, the majority screened positive for depressive symptoms (62.7%) and scored as having a high objective risk for acquiring HIV (89.3%);(42) with 83.0% of participants reporting at least one episode of condomless receptive anal sex (CRAS) within the prior six months. In interactions with their primary care provider (PCP), 36.8% were "somewhat-" or "extremely" uncomfortable with talking about sex with their PCP.

Participants
When strati ed by PrEP experience, those with no previous PrEP experience were mostly of non-Hispanic ethnicity, were stationed on the Western region of the U.S., and were more likely to report having only a high school education. These participants also reported a lower number of male and HIV-positive partners within the last six months, and tended to report being less satis ed with their current level of HIV protection. Table 4 shows the relative importance scores (RIS) of the ve attributes and Table 5 shows the part-worth utility scores (zero-centered) for each attribute level. Overall, the dosing method was the most important attribute among all the participants. This suggests that the participant's decision-making process to participate in a PrEP program is mostly in uenced by the type of PrEP dosing method offered within a program; regardless of PrEP experience. For participants reporting experience with PrEP, the daily tablet was the most preferred dosing method option, followed by the on-demand tablet regimen. For those with no previous experience with PrEP, the bi-monthly PrEP injection was the most preferred dosing method option, with the yearly implant and daily tablet also preferred to a lesser degree. Among the remaining attributes, both groups generally preferred the option to see a military healthcare provider, to use a smartphone to conduct the PrEP visit, and to utilize an on-base location for laboratory evaluation and receipt of medication. In adaptive choice-based conjoint analysis experiments, an "opt-out" or "None" parameter score is calculated during the screener section of the experiment. This parameter represents the positive or negative magnitude in which a respondent is likely to select "None" or opt out of PrEP in any scenario despite program con guration.(39) Thus, respondents with no PrEP experience are less likely to select "None" compared PrEP-experienced individuals, and are more likely to initiate PrEP regardless of program con guration.

Preferences for PrEP Program Scenarios
To examine preferences for PrEP across programs with different combinations of features, we utilized the randomized rst choice model to estimate the percent of respondents that would prefer or have participation interest (share of preference) in each of the eight PrEP program scenarios based on the partworth utility scores collected from the conjoint analysis survey. (49) Table 6 describes the eight PrEP program con gurations and Table 7 displays the participation interest scores across each individual PrEP program scenario if it were available to them as an option. Programs with longer-acting PrEP options in the form of injectables and implants (Scenarios 4 through 6) were con gured for military members whose personal or work-related circumstances compel the individual to seek PrEP options with fewer dosing administrations or from a remote location with limited resources. When compared to the Scenario 1 baseline, an injectable PrEP option provided little impact on participation interest among PrEP-experienced individuals. However, these scenarios garnered a higher participation interest rate among those with no PrEP experience (80.5%) when injectable PrEP is offered within a smartphone- Off-base program con gurations (Scenario 7 & 8) were con gured to represent civilian-equivalent, offbase, PrEP programs that circumvent the military and on-base aspects of a PrEP program. These programs scored the lowest among the total sample, with a participation interest rate of 57.7% among all respondents for an off-base, civilian daily tablet PrEP program, and a participation interest of 40.5% for an off-base, civilian rectal douche option.

Discussion
The signi cant ndings from this study reveal that overall respondents prefer the convenience of daily tablet PrEP services at an on-base location compared with civilian and off-base settings, yet those with no previous PrEP experience have a stronger preference for longer-acting injectables and implants. Additionally, over half of all respondents have a positive screening score for depressive symptoms, the majority of respondents engage in risk behaviors that categorize them as having a high risk for acquiring HIV, and over one-third of respondents report a level of discomfort in discussing their sex life with a primary care provider. With a growing body of literature suggesting a link between depression and sexual risk behaviors among MSM, (50)(51)(52) it may be bene cial for PrEP-prescribing providers to provide PrEP clinics that are sensitive and inclusive to sexual identity and to remain vigilant to address factors related to mental and sexual health speci c to MSM.
Results of the conjoint experiment found the dosing method attribute to be the most critical and in uential preference factor within a PrEP delivery program, with a strong overall preference for a daily tablet among the total sample. Among those with no previous PrEP experience, a dominant preference for PrEP injectables and implants suggests that a demand for these longer-acting PrEP methods exists within this population once these alternatives become widely available. The preliminary results of cabotegravir injectable PrEP and MK-8591-eluting PrEP implants are promising (53,54), yet the bene ts to longer-acting agents is that it has high acceptability and lifts the burden of a daily tablet from a user; (55,56) an advantage for an individual with adherence concerns or an unpredictable work schedule. Military service's expeditionary nature often entails military members to relocate, deploy, or miss regular follow-up appointments due to speci c duties. (57,58) Therefore, the availability of alternative, longer-acting PrEP modalities that can sustain a protective level of medication with fewer medication administrations may be an attractive component for military members interested in PrEP. While not as important as the dosing methods, respondents additionally preferred seeing a military provider, interacting through a smartphone telehealth visit, and utilizing on-base locations for laboratory evaluation and medication pick-up.
A preference to see a military provider for PrEP services remains prevalent among the total sample, signaling that it is the military healthcare provider that will be central to the success of a military PrEP program. However, a survey of military health care providers regarding PrEP knowledge and prescription habits revealed that 49% rated their knowledge as poor and only 29% had ever prescribed it.(2) Additionally, most military members receive their PrEP prescription only after seeing an infectious disease specialist,(2) which suggests that military primary care providers may not feel comfortable prescribing PrEP or may be referring military members to infectious disease specialists for PrEP services. This lower level of PrEP knowledge and prescription practice may contribute to the heterogenous nature of PrEP availability that currently seems dependent on a military member's geographic location,(2) and could explain the statistically signi cant difference in respondents' regions of station within this study when strati ed by PrEP experience. However, it's been shown that an increase in PrEP knowledge has been associated with an increase in prescribing habits,(59) therefore supporting military primary care providers with the necessary training and resources to comfortably prescribe PrEP may help military members engage in a wider availability of PrEP services without the extra step of a referral to an infectious disease specialist. Further research will need to explore the preference for a military provider within this context, and how this preference can best be leveraged to improve uptake.
Finally, respondents with no previous PrEP experience tend to report fewer male sexual partners and fewer HIV-positive partners within the last six months, yet are also more likely to report being unsatis ed with their current level of HIV protection. Given that sexual contact with men and condomless sex have been found to be the most common indications for initiating PrEP among military MSM,(2) further studies will need to explore if a desire for more intimacy with male partners with the absence of condoms is not being met with the currently available options for PrEP through access or availability.
This study has limitations. First, this study utilized self-report measures from a convenience sample recruited from an online social media group comprised of U.S. military members who identify as LGBT.
While there was no way to verify actual eligibility for inclusion/exclusion criteria due to the anonymous nature of the survey, the literature examining MSM recruitment via online methods versus in-person had found similar samples of HIV/STI prevalence and HIV-testing patterns among MSM.(60, 61) Furthermore, these ndings may not be generalizable to at-risk military members who have sex with men yet do not identify as being MSM or LGBT, and further qualitative studies are needed to further explore the foundation of these preferences within this context. Also of note is the large number of respondents excluded to RLH cut-offs. There is a rising trend of "random responders" and "survey-bots" that attempt duplicate submission of surveys that provide nancial compensation and can impact preference data if included within nal the analyses.(45) These RLH standards follow evidence-based consistency cut-offs to eliminate these "random responders," yet further studies should examine additional methods to address this rising phenomenon. Finally, while quantifying preferences does not guarantee intention or ultimately behavior, it can be an innovative rst step and provide a foundation to successfully inform PrEP initiatives. Ultimately, subsequent steps will translate this preference data into an implementation science hybrid design that re nes service delivery approaches to optimize uptake and sustained use of this effective HIV-prevention tool.

Conclusion
This study provides an initial description of the preferences and interest for PrEP by U.S. military members with a high risk of acquiring HIV. Our results indicate that PrEP interest among this population is most likely to be successful when PrEP is offered as a daily tablet, injection, or implant, with a medical visit performed with a military healthcare provider through a telehealth smartphone app. Additionally, allowing on-base locations to provide laboratory samples and to receive PrEP medication can also facilitate program preference. PrEP engagement and availability will further be enhanced by ensuring that medical providers and facilities are knowledgeable and comfortable prescribing PrEP services. Offering an a rming environment sensitive to health care concerns related to mental and sexual health will also be important. Consequently, key populations, stakeholders, and policymakers will be better equipped for scale-up of PrEP among at-risk populations within the U.S. military.

Availability of data and material
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Social media group administrators placed monthly advertisement links describing the study on the group's main forum. Interested participants could click on a link to access an online study disclosure form with a 'click to consent' procedure.  Daily oral tablet means that you would have to take an oral tablet every day (daily) to maintain a protective level of PrEP medication.

PrEP injection
PrEP injection means that you get an injection or shot that would provide a protective level of PrEP medication for 2 months.

PrEP implant
PrEP implant means that you would get a small implant that would slowly release a protective level of PrEP medication for at least a year.

Ondemand regimen
On-demand regimen means you take two tablets 24 hours before sex and then one tablet daily for the next two days. This dosing method would protect you from HIV for that single sexual encounter only.

PrEP rectal douche
Rectal PrEP douche means that you would use a rectal douche or enema prior to having sex that leaves behind protective level of PrEP medication for that sexual encounter.

Military
Military provider means that you prefer a medical visit with a healthcare provider that is a member of the military.

Civilian
Civilian provider means that you prefer a medical visit with a healthcare provider that is a civilian or not a member of the military.

PrEP Visit Location
Smartphone Smartphone/mobile app visit means that you prefer to have a virtual medical visit with a healthcare provider through a smartphone call or mobile app.

On-Base
On-Base medical visit means that you prefer an in-person medical visit with a healthcare provider that is in a clinic on-base.

Off-Base
Off-Base medical Visit means that you prefer an in-person medical visit with a healthcare provider that is in a clinic off-base.

Lab Evaluation Location
Provide labs onbase Provide lab work on-base means that you prefer to do you lab work at a laboratory or clinic on-base.

Provide labs offbase
Provide lab work off-base means that you prefer to do you lab work at a laboratory or clinic off-base.

Homebased mailin kit
Home-based mail-in kit means that you prefer to receive a home-based lab testing kit in the mail. You will provide self-collected, small samples of blood and urine and mail the kit back to the laboratory for evaluation. Your PrEP provider would then see the lab results after processing.

PrEP Dispensing Venue
Receive PrEP onbase Receive PrEP on-base means that you prefer to pick up or receive your PrEP medication from a pharmacy/clinic on-base.

Receive PrEP offbase
Receive PrEP off-base means that you prefer to pick up or receive your PrEP medication from a pharmacy/clinic off-base.

Receive PrEP by mail
Receive PrEP by mail delivery means that you prefer to receive your PrEP medication in the mail at your home or APO.

Tables
Page 18/27        a: Zero-centered part-worth utility scores imply the positive or negative magnitude of the participant's preference for the level choice in relation to the other level options within the same attribute. b: The "None" parameter represents the positive or negative magnitude in which a respondent is likely to select "None" (not willing to take PrEP in any scenario despite program con guration) a: Scenarios descriptions reference Scenarios 1 through 8 in Table 7. a: Share of preference denotes the percent of respondents that would prefer or have interest to participate in the respective PrEP program scenario with a particular combination of program features based on utilities obtained during the conjoint survey.
b: Descriptions of PrEP scenarios 1 through 8 are explained in Table 6. Figure 1 Sample conjoint choice task screener item. Note: Attribute level options presented within programs change as choice tasks are made, which results in captured preference data on the most important attributes driving a respondent's selection of a particular program con guration.