Supplemental Digital Content 1. Pictorial PK Result Tool Error! Reference source not found. Legend: The pictorial PK result tool was developed through expert consultations with PrEP researchers and site staff and pilot-tested at each site. The tool was designed to serve several purposes: (1) to provide graphic representation of a potentially complex concept using well-recognized, locally relevant objects (tea cups and tea pot), and to simplify specific average drug detection levels (e.g., 18%) into five broad categories (TFV never detected, TFV detected in 1-49% 50-74%, 7599%; and 100% of plasma samples, respectively); and (2) to minimize the likelihood of feeling judged by providing participants with their categorical adherence level rather than a point estimate of their individual result. Tea cups represent detection of TFV in plasma at a given visit. Tea cup rows represent pattern of use; by design, presentation was not specific to individual results. For analysis purposes, women were classified at 3 PK levels: low (0% of plasma samples with drug detected), inconsistent (1-74% of plasma sample with drug detected) and high (75100% of plasma samples with drug detected). Supplemental Digital Content 2. Rank ordered statement cards: by PK level, by product group, and overall (N=68 IDIs) PK level High (n=20) Health services Tablet (n=32) Side effects Product Gel (n=36) Side effects Card ranking 1 Low/Inconsistent (n=48) Side effects Total (n=68) Side effects 2 Harm Side effects Health services Harm Harm 3 Health services Harm Harm Health services Health services 4 5 6 7 8 9 10 11 12 13 14 15 16 Support Can’t prevent HIV Boring Challenge to take Too busy HIV stigma Told not to take Peer influence Partner disliked Family/friends Had to hide Not interested Living situation Told not to take Peer influence Not at risk Had to hide Boring HIV stigma Can’t prevent HIV Reimbursements Challenge to take Support Too busy Partner disliked Sexual pleasure Challenge to take HIV stigma Support Told not to take Boring Peer influence Can’t prevent HIV Family/friends Not interested Had to hide Didn’t need Too busy Living situation Too busy Can’t prevent HIV Support Partner disliked Boring Told not to take Peer influence HIV stigma Had to hide Not at risk Challenge to take Living situation Family/friends Support Told not to take Boring Can’t prevent HIV HIV stigma Challenge to take Peer influence Too busy Partner disliked Had to hide Family/friends Not at risk Not interested 17 18 19 20 Not at risk Didn’t need Sexual pleasure Reimbursements Family/friends Not interested Didn’t need Living situation Partner disliked Reimbursements Not at risk Sexual pleasure Sexual pleasure Not interested Reimbursements Didn’t need Living situation Didn’t need Sexual pleasure Reimbursements Legend: Following an open-format discussion, interviewers used a set of cards listing 20 predetermined statements intended to elicit additional information about study participation and adherence challenges. The statements were selected by the research team from previous qualitative research and recorded on a series of cards. Participants were asked to select the statements that were salient to them from their experience during VOICE and relative to their product adherence challenges (pile sorting), and to rank these in order of relevance. Cards were assigned a score according to ranking. A card ranked as most relevant was assigned a score of 20, a card ranked as second most relevant was assigned a score of 19, and so on. Non-relevant cards were assigned a score of 0. Average scores for each card were calculated and are presented in descending order, from most to least relevant. Participants were given an opportunity to indicate additional challenges during the exercise if they desired, though these were not scored. The following key provides the full text of each card: Boring It was boring to take the products daily; Can’t prevent HIV The products cannot prevent HIV; Challenge to take It was challenging to swallow big tablets/insert a gel in my vagina; Didn’t need I didn’t need to take the products; Family/friends My family/friends disliked the products/VOICE; Had to hide I had to hide when taking my products; Harm The products may be harmful; Health services I joined the study for health services provided by the clinic; HIV stigma Others would think I have HIV if I took the products; Living situation The place where I lived made it hard to take the products; Not at risk I wasn’t at risk of getting HIV; Not interested I was not interested in using the products; Partner disliked My partner disliked the products/VOICE; Peer influence [I heard that] other participants were not taking their products; Sexual pleasure The products decreased sexual pleasure; Side effects I experienced or was worried about side effects; Support I didn't have enough support from others to take the products; Told not to take Someone told me to not take the products; Too busy I was too busy to take products every day; Reimbursements I joined the study for the reimbursements. Supplemental Digital Content 3: Comparison between characteristics of VOICE-D stage 2 and VOICE participants not enrolled in VOICE-D at same sites (in Durban, Kampala, and Chitungwiza) Characteristics VOICE-D stage 2 Sample Non-stage 2 VOICE-D sample at same sites pvalue1 (N = 131) % (N =1311) % Age 25 or younger 77 59% 633 48% 0.03 Completed secondary school or more 46 35% 626 48% 0.006** At time of VOICE trial enrolment Earns an income of her own 88 67% 840 64% ns Parity (0-1 children) 59 45% 570 43% ns Married 59 45% 546 42% ns Had multiple sex partners in previous 3 months (ACASI) 43 33% 358 27% ns Used a condom at last sex (CRF) 93 71% 875 67% ns Current injectable contraceptive (CRF) 92 70% 848 67% ns Current oral contraceptive (CRF) 30 23% 298 23% ns Baseline behavior in VOICE Legend: 1Fisher’s exact test; **Chi-square test no longer significant when stratified by site/country CRF, Case Report Form; ACASI, Audio-Computer Assisted Self-Interviewing Supplemental Digital Content 4. Participants’ reactions to PK results and illustrative quotes Accepted PK results (high, inconsistent and low PK groups) I: Now, the results that came back from your blood showed that for most of the time your blood was checked, the product was found to be in your body and that you fall in line “A” R: Ehh and yet I used to miss? [laughs] Okay, that happened in the beginning but towards the end we had understood it well and it was swallowed daily. I: You could no longer forget? R: Yes, but that happened in the beginning and [then] we were given a calendar. (IDI, Uganda, high PK, Tablet) I: When we tested your blood, we found that you had been using the product because it was found in your blood. You fall under category "B". [..] What was your experience? What would you say helped you use the product even though your partner didn't like it? R: I would say that I'm the kind of person who cares about her health. It was important to get tested and know my status. We did PAP-smears and we had our blood drawn for testing. Some people continued in the study but others stopped. Luckily I was among those who continued. That motivated me a lot... I told myself to use the product because it might help me.” (IDI, South Africa, high PK, Gel) I: What this shows is that your blood samples show that you had traces of the product in your blood at times and sometimes you had none. […]What do you think about your results? R: Yes, that’s correct since on some days I would forget. That’s correct. (IDI, Zimbabwe, inconsistent PK, Gel) I am disappointed. Because I have been found out that I hadn’t been using the gel. (IDI, South Africa, low PK, Gel) We weren’t faithful. We weren’t honest enough to use the products (IDI, Zimbabwe, low PK, Gel) Did not accept initially and/or challenged PK results (all low or inconsistent PK groups) Initial reaction: R: Maybe something happened, they made a mistake with the results. I: Why do you say we made a mistake on your results? R: Ah because how can you not find the drug at all when I was using my tablets? Later in IDI: R: I feel guilty..[..] Sometimes I would forget [laughs]; other times I took them. (IDI, Zimbabwe, low PK, Tablet) Initial reaction: I: Now, they are informing you that you belong to Category E and not D as you had thought. Most of the time, there were no products in your blood R: I am not sure I: Why aren’t you sure? R: Because I used to use. Later in IDI: I: You used to insert the gel but why do you think the results came back like this? R: I don’t know doctor [referring to interviewer]. Maybe when they tested, that is what they found. You cannot argue and insist on it when they did not find any products. (IDI, Uganda, low PK, Gel) I became less motivated to take the products, as time went by. But I did use the products for a while. Because why are there no traces of the product at all, in my blood sample in the initial stages? We knew that there would be checks for product use, so I used the product a bit. So I am failing to comprehend why there was no evidence, whatsoever, of the product in my blood? I do not understand at all. [..] It looks like I did not use the product at all, but I did use it sometimes. That is what made me upset. (IDI, Zimbabwe, low PK, Gel) Denied results; some provided alternative explanations (all low or inconsistent PK groups) I feel bad because I used it [the tablets]. [..] I swear to God I never did [have a problem]. I would drink it whenever I was listening to a Xhosa story which starts at 19:15. (IDI South Africa, low PK, Tablet) I used to take a lot of tablets. Some for headache, chest pain and ulcers. I was swallowing tablets all the time. In my opinion, I think that might have led to the inability to detect the medication in my blood. (IDI, Uganda, low PK, Tablet) I used to drink beer a lot, maybe my drinking distorted the results. In fact I was drunk on most of my visits. (IDI, Zimbabwe, low PK, Tablet) How were you able to tell from my blood that it was working [had the drug] when the one I got was a placebo? […]They will say to you, “you, the one that you were using didn’t have the active drug”. In my case I was told that … (FGD, Zimbabwe, low PK, Gel)* Legend: I= interviewer; R=respondent. Quotes from respondents are italicized to distinguish them from statements by the interviewer. (*) we verified that this participant had been assigned to active gel during VOICE and that she had been correctly informed of her study arm by staff after trial completion. Supplemental Digital Content 5. Typology of product use patterns Non-initiation Discontinuation Misimplementation Temporary Visit-driven use Permanent Variable taking (e.g. missed set time) Adherence Modified dosing (e.g. extra or partial dose) Modified regimen (e.g. intermittent or episodic use) Legend: Product use patterns reported in VOICE-D can be described by this typology, which includes four major adherence behavior components (adapted from [14; 31; 32]: non- initiation (or nonengagement with drug regimen), discontinuation (temporarily or permanently), persistence with the product regimen but mis-implementation (or poor execution) and adherence (i.e. using product as instructed during the trial). Discontinuation and mis-implementation may or may not be intentional. Mis-implementation can result from various behaviors including visit-driven use (or white coat compliance); variable taking, when the product is not taken at the set time; modified dosing, when the correct amount of product is not taken with each dose (over- or under-dosing), and modifying the regimen or frequency of product use (i.e. non-daily, inconsistent, intermittent or episodic use).