Improving adherence to daily preexposure prophylaxis among MSM in Amsterdam by providing feedback via a mobile application : AIDS

Secondary Logo

Journal Logo

EPIDEMIOLOGY AND SOCIAL

Improving adherence to daily preexposure prophylaxis among MSM in Amsterdam by providing feedback via a mobile application

van den Elshout, Mark A.M.a; Hoornenborg, Elskea; Achterbergh, Roel C.A.a; Coyer, Lizaa; Anderson, Peter L.b; Davidovich, Udia,c; de Vries, Henry J.C.a,d,e; Prins, Mariaa,d,f; van der Loeff, Maarten F. Schima,d

Author Information
doi: 10.1097/QAD.0000000000002949

Abstract

Introduction

MSM are disproportionally affected by HIV in the Netherlands, accounting for 61% of 525 new HIV diagnoses in 2019 [1]. Prevention strategies, such as barrier methods and viral load sorting rely, at least partly, on external factors, for example, correct condom use and adherence to antiretroviral therapy by a sex partner. By using preexposure prophylaxis (PrEP), partner-related external factors become less important for prevention of HIV acquisition. Both daily and event-driven regimens consisting of a combination of emtricitabine and tenofovir disoproxil (TDF) are highly efficacious in preventing HIV, as long as adherence is good [2]. Measuring drug levels [i.e. the concentration of the intracellular metabolite of TDF (tenofovir diphosphate or TFV-DP) in dried blood spots (DBS) is an objective way to ascertain longer term adherence to daily PrEP (dPrEP), because of TFV-DP's long half-life compared with tenofovir levels in plasma [3,4].

In several earlier studies, poor adherence was a major barrier to PrEP effectiveness [2]. Novel approaches to improve adherence often include use of mobile devices. Previous research has shown that sending text messages or emails can sometimes improve short-term adherence [5–9], and that personalized content could be more effective compared with basic and repetitious content, such as daily text reminders [10,11]. Therefore, there is continued need for engaging, evidence-based interventions to improve PrEP adherence [12].

We assessed whether automated visualized feedback on personal adherence [13] based on daily, self-reported PrEP use and sexual behaviour via a mobile application (app) can improve adherence to dPrEP over the course of 2 years, by performing a randomized clinical trial (RCT) nested in a cohort of MSM and transgender women (TGW) using PrEP in the Netherlands.

Methods

The CONSORT guidelines were used to report this RCT [14,15].

Settings and location of data collection

An RCT was nested in the Amsterdam PrEP (AMPrEP) study. AMPrEP (3 August 2015 to 1 December 2020), an open-label demonstration study including MSM and TGW, offered a choice between free-of-charge dPrEP or event-driven PrEP (edPrEP). AMPrEP aimed to assess PrEP uptake, determinants of PrEP regimen choice, adherence and incidence of sexually transmitted infections (STIs). Switching between the two regimens was allowed at each study visit. Study design, aim and procedures have been described previously [16]. Briefly, participants attended 3-monthly study visits at the STI outpatient clinic of the Public Health Service (GGD) of Amsterdam, the Netherlands. Eligible were HIV-negative MSM and TGW, who were at least 18 years old and in the 6 months prior to AMPrEP screening had a substantial likelihood to acquire HIV [16].

Eligibility and randomization

Participants of AMPrEP were eligible to participate in the RCT if they were using PrEP according to the daily regimen at the RCT screening visit and intended to continue dPrEP use. In addition, they had to use the AMPrEP app, comply with RCT procedures and give written informed consent. The study nurse/physician enrolled participants in the RCT at their first available AMPrEP study visit, that is, at 3, 6 or 9 months. Written informed consent was obtained if a participant was eligible and willing to participate. The study nurse/physician used a random-number generator integrated in a Microsoft Access 2007–2011 database (Microsoft Corporation, Redmond, Washington, USA) to randomize participants in a 1 : 1 ratio to either the control or intervention arm.

Amsterdam PrEP demonstration project app

Upon inclusion, all participants were offered a personal registration code for the AMPrEP app (Fig. 1a), in which they could register, on a daily basis, whether they had used PrEP, had had sexual intercourse, the type of sex partner they had anal sex with and use of condoms. The app also provided a basic reminder to take PrEP [17]. Use of the app was voluntary. Participants received automated messages at random intervals, monthly on average, with an additional motivating text when they had taken more than 90% of their doses. When information in the app had not been completed, a message was sent for up to 3 days requesting to complete the information and whenever indicated in the app that no medication was taken, a message was sent 3 days in a row reminding the participant to take the medication and register it in the app. General information was pushed on a monthly basis to all app users, and a reminder to use PrEP could be set by participants themselves. The information on pill use and sexual behaviour in the app was not accessible to the study nurses and doctors during the consultations and was not discussed with the participants.

F1
Fig. 1:
(a) Screenshot of the AMPrEP app overview tab, in which all participants could register on a daily basis whether they had used PrEP, had had sexual intercourse, the type of sex partner they had anal sex with and use of condoms. (b) Screenshot of the AMPrEP app trends tab (only available to participants in the intervention arm). It displays visualized feedback on the self-administered input in the AMPrEP app.

Intervention

Participants in the intervention arm got access to additional features in the AMPrEP app: visualized feedback, based on their self-administered input, in the form of graphs depicting trends in pill use and number of sex partners per week and per month. The app also provided feedback through bar charts indicating the proportion of days in the past month on which PrEP was used and the proportion of sex acts that was covered by PrEP, condoms or both (Fig. 1b). Participants in the intervention arm also got access to a more advanced alarm function with the option to set more than one daily reminder, and had a private tab for taking daily notes to which the researchers had no access. The intervention was developed in collaboration with a community engagement group set up for AMPrEP, consisting of key persons from the Amsterdam gay community, and a participant group.

Measures

Demographic, psychosocial and behavioural characteristics were collected via self-reported questionnaires [16]. Age, gender identity, ethnicity, place of residency, education level, employment status, income level, living situation, relationship status, sexual preference and history of bacterial sexually transmitted infections (STIs) in the 6 months prior to inclusion were asked at inclusion in AMPrEP. Number of sex partners and sex episodes, including partner type and condom use, were asked every 3 months. Sexual compulsivity was measured using the sexual compulsivity scale [18], with a score at least 24 being indicative of a greater impact of sexual thoughts on daily functioning and of an inability to control sexual thoughts or behaviours [19,20]. Chemsex was defined as the use of γ-hydroxybutyrate, γ-Butyrolactone, methamphetamine or mephedrone prior to or during sex. Symptoms of depression or anxiety were assessed using the MHI5 score, wherein a score of less than 60 indicated symptoms of depression or anxiety [21,22]. AUDIT [23] and DUDIT [24] questionnaires were used to assess problematic alcohol and drug use, respectively; a score at least 8 is interpreted as indicative of alcohol or drug-related problems.

Laboratory methods

At the 12-month and 24-month study visits, whole blood was collected via phlebotomy and spotted onto Whatman 903 Protein Saver Cards (GE Healthcare Bio-Sciences Corp, Piscataway, New Jersey, USA). After drying for at least 2 h on room temperature, the cards were sealed in plastic bags including a desiccant with a humidity indicator (MiniPax, Multisorp Technologies, Buffalo, New York, USA) and stored at −20°C. Samples were shipped in batches for analysis at the laboratory of the Skaggs School of Pharmacy and Pharmaceutical Sciences (University of Colorado Anschutz Medical Campus, Aurora, Colorado, USA) as described previously [3,25]. Intracellular TFV-DP was measured in fmol/punch; results were not reported to participants.

Primary and secondary outcomes

The primary outcome was poor adherence at 12 or 24 months after PrEP initiation, defined as TFV-DP less than 700 fmol/punch, corresponding to using less than four tablets per week on average. The complement – good adherence – was defined as TFV-DP at least 700 fmol/punch, corresponding to using four or more tablets per week, at both 12 and 24 months. Excellent adherence was the secondary outcome, which was defined as TFV-DP at least 1250 fmol/punch, commensurate with 7 doses/week on average [26,27], at both 12 and 24 months.

Statistical methods

Sample size calculations assuming a difference of 20% between intervention and control arm in good adherence, indicated that 75 participants per arm were needed (see Supplement, https://links.lww.com/QAD/C147).

The primary analysis was performed on the per-protocol study population, defined as randomized participants who provided DBS samples at both 12 and 24 months study visits, had used dPrEP in the 3 months prior to those study visits, and had not used direct acting antivirals (DAAs) for treatment of hepatitis C virus (HCV) infection between enrolment in the RCT and the 24-month visit as combined use of sofosbuvir containing DAAs and TDF is associated with elevated intracellular TFV-DP [28]. We compared baseline and behavioural characteristics between participants included versus not included in per-protocol analyses using χ2, Fisher exact or Wilcoxon rank-sum tests as appropriate.

We performed the following secondary analyses: an intention-to-treat analysis of the primary outcome on all randomized participants, including those who missed study visits or samples, switched to using edPrEP and those who used DAAs for treatment of HCV during the RCT. Missing TFV-DP results were interpreted as less than 700 fmol/punch. A per-protocol analysis using excellent adherence (TFV-DP ≥1250 fmol/punch at 12 and 24 months) as outcome.

Statistical analyses were performed in STATA 15.1 (StataCorp, College Station, Texas, USA). The study was approved by the ethics board of the Amsterdam University Medical Center, location AMC, Amsterdam, the Netherlands (NL49504.018.14) and registered at the Netherlands Trial Register (NL5413). The protocol is available online at https://www.trialregister.nl/trial/5413.

Role of funding source

The funders played no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript. The authors had full access to all data and were responsible for the decision to submit the manuscript for publication.

Results

Participants

AMPrEP participants were included in the RCT between 4 May 2016 and 12 September 2016. Study visits at which DBS were sampled took place between 28 July 2016 and 12 July 2018. A total of 351 AMPrEP participants were assessed for eligibility of whom 100 (28%) did not meet inclusion criteria and 22 (6.3%) declined to participate in the RCT (Fig. 2). Of 100 AMPrEP participants who did not meet inclusion criteria for the RCT, 89 were using event-driven PrEP at the time of screening, eight did not have a suitable phone, two did not want to use the app and in one case there was a language barrier. We randomized 229 persons of whom 111 (48.5%) were allocated to the control arm and 118 (51.5%) to the intervention arm. All allocated participants received the allocated intervention. During follow-up, two of 229 (0.87%) participants, both from the intervention arm, withdrew informed consent and were excluded from all analyses. Four out of 111 (3.6%) participants of the control arm and three of 118 (2.5%) participants of the intervention arm were treated with DAAs during the RCT and were excluded. At 12 months, there was no DBS sample from two of 106 (1.9%) in the control arm and two of 111 (1.8%) of participants in the intervention arm. At 24 months, DBS was missing from two of 101 (2.0%) and eight of 101 (7.9%) of participants, respectively.

F2
Fig. 2:
Flow diagram of AMPrEP study participants in each phase of the nested randomized clinical trial.

In total, 166 of 227 (73.1%) participants, 83 of 111 (74.8%) in the control arm and 83 of 116 (71.6%) in the intervention arm, attended both the 12-month and 24-month study visits and provided DBS samples both times, without having been treated for HCV with DAAs and were included in per-protocol analyses. Intention-to-treat analyses were performed on all randomized participants excluding two who withdrew informed consent (n = 227). Participants were all analysed within their originally allocated study arms.

Comparison of participants included versus not included in per-protocol analyses

In comparison to the 166 included in the per-protocol analyses, the 61 not included were younger [median 36 (interquartile range; IQR 27–43) versus 39 years (32–47); P = 0.007], had lower income (P = 0.02), more often lived with parents/flatmates instead of with a partner (P = 0.02), were less often in a steady relationship (P = 0.009) and more often showed symptoms of depression or anxiety (P = 0.006; Supplementary Table S1, https://links.lww.com/QAD/C147).

Baseline characteristics of the per-protocol cohort

Median age of the per-protocol cohort was 39 years (IQR 32–47). One participant identified as TGW, 165 as men. Ethnicity was self-declared as white by 88%. Three-quarters of participants reported their education level as college/university (Table 1).

Table 1 - Baseline characteristics of participants in the control arm and in the intervention arm of the adherence randomized clinical trial, who were included in per-protocol analyses.
Total (n = 166) Control app arm (n = 83) Intervention app arm (n = 83)
n %a N %a n %a
Demographic characteristics
 Age (years)
  Median [IQR] 39 [32--47] 37 [31–47] 41 [33–47]
 Age (categorized)
  20--34 53 32% 29 35% 24 29%
  35--44 54 33% 24 29% 30 36%
  45--73 59 36% 30 36% 29 35%
 Gender identity
  Male 165 99% 82 99% 83 100%
  Transgender woman 1 1% 1 1% 0 0%
 Self-declared ethnicity
  White 146 88% 75 90% 71 86%
  Nonwhite 20 12% 8 10% 12 14%
 Place of residency in the Netherlands
  Amsterdam 99 60% 43 52% 56 67%
  Other 67 40% 40 48% 27 33%
 Education level
  No college/university 41 25% 23 28% 18 22%
  College/university 125 75% 60 72% 65 78%
 Employmentb
  Employed 133 82% 65 79% 68 84%
  Unemployed 5 3% 3 4% 2 2%
  Other (retired, volunteer, disabled, student) 25 15% 14 17% 11 14%
 Net monthly income in Euro'sc
  ≤1700 35 22% 19 24% 16 20%
  1701--2950 78 49% 35 44% 43 54%
  >2950 45 28% 25 32% 20 25%
 Living situation
  Alone 83 50% 40 48% 43 52%
  With partner 62 37% 26 31% 36 43%
  With parents/flatmates 21 13% 17 20% 4 5%
 Steady relationshipd
  No 83 50% 46 56% 37 45%
  Yes 82 50% 36 44% 46 55%
 Sexual preferenced
  Exclusively homosexual 126 76% 61 73% 65 79%
  Not exclusively homosexual 39 24% 22 27% 17 21%
Sexual behaviour
 Number of sex partnersf
  Median [IQR] 20 [10–36] 20 [10–38] 20e [10–34]
 Number of condomless anal sex episodes with casual partnersf
  Median [IQR] 12 [5–24] 12 [5–23] 13e [6–24]
 Sexually transmitted infectionsg
  No 99 60% 47 57% 52 63%
  Yes 67 40% 36 43% 31 37%
Mental health characteristics and drug use
 Sexual compulsivity scale
  Score <24 (no indication of sexual compulsivity) 129 78% 67 81% 62 75%
  Score ≥24 (indication of sexual compulsivity) 37 22% 16 19% 21 25%
 Chemsexh e
  No 94 57% 47 58% 47 57%
  Yes 70 43% 34 42% 36 43%
 Depression or anxiety symptoms
  MHI5 score ≥60 (no symptoms)i 145 87% 76 92% 69 83%
  MHI5 score <60 (symptoms)j 21 13% 7 8% 14 17%
 Alcohol use disorder identification test (AUDIT)d
  Score <8 (no indication)k 117 71% 60 73% 57 69%
  Score ≥8 (indication)l 48 29% 22 27% 26 31%
 Drug use disorder identification test (DUDIT)
  Score <8 (no indication)m 99 60% 52 63% 47 57%
  Score ≥8 (indication)n 67 40% 31 37% 36 43%
 Level of concern about acquiring HIVo
  Low 25 15% 13 16% 12 14%
  Neutral to high 141 85% 70 84% 71 86%
 Level of importance to prevent HIVo
  Less than very important 41 25% 19 23% 22 27%
  Very important 125 75% 64 77% 61 73%
Inclusion
 AMPrEP study visit at which included into RCT
  3 months 75 45% 39 47% 36 43%
  6 months 69 42% 30 36% 39 47%
  9 months 22 13% 14 17% 8 10%
AMPrEP study, Amsterdam, 2015--2018. AMPrEP, Amsterdam PrEP demonstration project; IQR, interquartile range; MHI5, Mental Health Inventory-5; PrEP, preexposure prophylaxis; RCT, randomized clinical trial.
aPercentages may not total 100 because of rounding.
bThree missing.
cEight missing.
dOne missing.
eTwo missing.
fIn the past 3 months prior to inclusion into the RCT.
gChlamydia, gonorrhoea and/or syphilis within 6 months prior to inclusion into AMPrEP.
hUse of γ-hydroxybutyrate, γ-Butyrolactone, methamphetamine or mephedrone prior to or during sex in the 3 months prior to inclusion into AMPrEP, two missing.
iNo indication of an anxiety or depressive mood disorder.
jIndication of an anxiety or depressive mood disorder.
kNo indication of an alcohol use disorder.
lIndication of an alcohol use disorder.
mNo indication of a drug use disorder.
nIndication of a drug use disorder.
oScale 1–7, dichotomized.

Adherence

Overall median TFV-DP in the per-protocol cohort was 1291 (IQR 1058–1603) fmol/punch; at 12 months 1304 (IQR 1122–1640) fmol/punch and at 24 months 1264 (IQR 997–1590) fmol/punch. Twelve months after PrEP initiation, the control arm had a median of 1265 (IQR 1010–1544) fmol/punch and the intervention arm 1391 (IQR 1158–1782) fmol/punch (P = 0.0255). At 24 months, the median TFV-DP was 1202 (IQR 989–1512) fmol/punch in the control arm and 1349 (IQR 1005–1654) fmol/punch in the intervention arm (P = 0.135).

Poor adherence

The primary outcome, poor adherence, was observed in nine of 83 (11%) of participants in the control arm, and in 13 of 83 (16%) of participants in the intervention arm (P = 0.36). The crude odds ratio (OR) for poor adherence in the intervention arm versus the control arm was 1.5 (95% CI 0.61–3.8). Two factors were significantly associated with the outcome: symptoms of depression/anxiety (OR 3.2; 95% CI 1.1–9.5) and low concern for acquiring HIV (OR 4.3; 95% CI 1.6–12; Table 2).

Table 2 - Baseline characteristics associated with poor adherence (tenofovir diphosphate <700 fmol/punch in dried blood spots at 12 or 24 months) in logistic regression analysis.
Univariable
Poor adherencea OR (95% CI) P value
Overall 22/166 (13%)
Demographic characteristics
 Age (per 10 years) at inclusion in the RCT 0.87 (0.58--1.3) 0.49
 Age (categorized) at inclusion in the RCT
  20--34 7/53 (13%) 1 0.89
  35--44 8/54 (15%) 1.1 (0.38--3.4)
  45--73 7/59 (12%) 0.88 (0.29--2.7)
 Gender identity
  Male 22/165 (13%)
  Transgender woman 0/1 (0%)
 Self-declared ethnicity
  White 19/146 (13%) 1 0.81
  Nonwhite 3/20 (15%) 1.2 (0.32--4.4)
 Place of residency in the Netherlands
  Amsterdam 11/99 (11%) 1 0.33
  Other 11/67 (16%) 1.6 (0.64--3.9)
 Education level
  No college/university 5/41 (12%) 1 0.82
  College/university 17/125 (14%) 1.1 (0.39--3.3)
 Employmentb
  Employed 19/133 (14%) 1 0.76
  Unemployed 0/5 (0%)
  Other (retired, volunteer, disabled, student) 3/25 (12%) 0.82 (0.22--3.0)
 Net monthly income in Euro'sc
  ≤1700 6/35 (17%) 1 0.67
  1701--2950 9/78 (12%) 0.63 (0.21--1.9)
  >2950 5/45 (11%) 0.60 (0.17--2.2)
 Living situation
  Alone 12/83 (14%) 1 0.83
  With partner 8/62 (13%) 0.88 (0.33--2.3)
  With parents/flatmates 2/21 (9.5%) 0.62 (0.13--3.0)
 Steady relationshipd
  No 11/83 (13%) 1 0.98
  Yes 11/82 (13%) 1.0 (0.41--2.5)
 Sexual preferencee
  Exclusively homosexual 18/126 (14%) 1 0.29
  Not exclusively homosexual 3/39 (7.7%) 0.5 (0.14--1.8)
Sexual behaviour
 Sexually transmitted infectionsf
  No 13/99 (13%) 1 0.96
  Yes 9/67 (13%) 1.0 (0.41--2.6)
 Number of sex partnersg 1.0 (0.99--1.0) 0.95
 Number of condomless anal sex episodes with casual partnersg 0.99 (0.96--1.0) 0.48
Mental health characteristics and drug use
 Sexual compulsivity scale
  Score <24 (no indication of sexual compulsivity) 16/129 (12%) 1 0.55
  Score ≥24 (indication of sexual compulsivity) 6/37 (16%) 1.4 (0.49--3.8)
 Chemsexh
  No 9/94 (10%) 1 0.16
  Yes 12/70 (17%) 2.0 (0.77--4.9)
 Depression or anxiety symptoms
  MHI5 score ≥60 (no symptoms)i 16/145 (11%) 1 0.034
  MHI5 score <60 (symptoms)j 6/21 (29%) 3.2 (1.1--9.5)
 Alcohol use disorder identification test (AUDIT)e
  Score <8 (no indication) 17/117 (15%) 1 0.48
  Score ≥8 (indication)l 5/48 (10%) 0.68 (0.24--2.0)
 Drug use disorder identification test (DUDIT)
  Score <8 (no indication) 12/99 (12%) 1 0.60
  Score ≥8 (indication)n 10/67 (15%) 1.3 (0.52--3.1)
 Level of concern about acquiring HIVo
  Neutral to high 14/141 (10%) 1 0.0047
  Low 8/25 (32%) 4.3 (1.6--12)
 Level of importance to prevent HIVo
  Less than very important 8/41 (20%) 1 0.18
  Very important 14/125 (11%) 0.52 (0.20--1.3)
RCT allocation
 Control app 9/83 (11%) 1 0.36
 Intervention app 13/83 (16%) 1.5 (0.61--3.8)
AMPrEP study, Amsterdam, 2015-2018. AMPrEP, Amsterdam PrEP demonstration project; IQR, interquartile range; MHI5, Mental Health Inventory-5; PrEP, preexposure prophylaxis; RCT, randomized clinical trial.
aPercentages may not total 100 because of rounding.
bThree missing.
cEight missing.
dOne missing.
eOne missing.
fChlamydia, gonorrhoea and/or syphilis within 6 months prior to inclusion into AMPrEP.
gIn the past 3 months prior to inclusion into AMPrEP, OR per increase of 1 of the natural log of the number of sex partners/episodes.
hUse of γ-hydroxybutyrate, γ-Butyrolactone, methamphetamine or mephedrone prior to or during sex in the 3 months prior to inclusion into AMPrEP, two missing.
iNo indication of an anxiety or depressive mood disorder.
jIndication of an anxiety or depressive mood disorder.
kNo indication of an alcohol use disorder.
lIndication of an alcohol use disorder.
mNo indication of a drug use disorder.
nIndication of a drug use disorder.
oScale 1–7, dichotomized.

Intention-to-treat analysis

Among the 227 participants included in the intention-to-treat analysis, 41 of 116 (35.3%) participants in the intervention arm had poor adherence or missing data at 12 or 24 months after PrEP initiation, compared with 32 of 111 (28.8%) participants in the control arm (P = 0.29).

Excellent adherence

In the per-protocol cohort, 66 of 166 (40%) of participants had excellent adherence. Of those, 26 of 83 (31%) were in the control arm and 40 of 83 (48%) in the intervention arm (P = 0.026), resulting in an OR of 2.0 (95% CI 1.1–3.8; Supplementary Table S2, https://links.lww.com/QAD/C147). Reporting chemsex was found to be negatively associated with excellent adherence in univariable analysis (OR 0.47; 95% CI 0.24–0.90; P = 0.022).

Discussion

This is the first RCT assessing whether providing visualized feedback on self-reported sexual behaviour and PrEP use via an app improves adherence to PrEP, in which we compared a control app with one that provided visualized feedback over a course of 24 months. In this cohort of MSM who use PrEP in Amsterdam, the Netherlands, a very high proportion of participants had good adherence to daily PrEP. No significant effect of the intervention on good adherence was found. The intervention increased the proportion of MSM with excellent adherence in univariable logistic regression [48% of participants of the intervention arm versus 31% of control arm (P = 0.026)].

We hypothesized that using an app to record PrEP use and sexual behaviour could be a low-cost, accessible way to engage PrEP users and support adherence to PrEP. However, an effect on good adherence could not be shown in our trial, perhaps because of high overall adherence in both arms. This high level of adherence could result from the characteristics of the included participants. Included participants had to be in possession of a good phone, willing to provide DBS, to consistently attend study visits and fill in the app. Inclusion might have selected for highly adherent individuals, hence creating a ceiling effect throughout the trial. Furthermore, the RCT participants were a subgroup of the AMPrEP cohort consisting of early adopters, highly motivated to use PrEP [16]. Last, the high adherence level could be an effect of the features included in both the control and the intervention app, in which PrEP use and sexual behaviour are registered on a daily base, which reminds one to take one's pills, and thus contributes to adherence by itself.

Another study of a cohort of highly adherent MSM and TGP in Southern California – which aimed to improve adherence by sending daily text messages – also reported that a high overall adherence could be a barrier to find improved adherence of the intervention [8]. Various other strategies are being attempted to improve adherence to PrEP. Songtaweesin et al.[29] performed an RCT among young MSM and TGW in Thailand to investigate the impact of self-assessment of HIV risk, point rewards and reminders for PrEP on adherence via an app but found no additional benefit of these app features on adherence. Liu et al.[30] are developing an app, which uses artificial intelligence combined with daily reminders, a sexual diary and visual feedback on covered sexual encounters, which was very promising in the pilot phase in improving adherence among the included population of young MSM. In a study of people with HIV, graphical representation of level of antiretroviral therapy medication and immune protection via an app was associated with higher self-reported adherence and lower viral load [31].

Future research on adherence improvement using mobile applications should focus on populations known to experience barriers to adherence — such as younger people — ideally when they have a concomitant high affinity with mobile technology. Currently, apps specifically designed for young MSM and TGW are being developed and investigated [30,32,33]. In addition, our study suggests that adherence to PrEP may wane over the course of 2 years of follow-up. More research is needed to determine whether this downward trend is real and whether this differs by (sub)population. Appropriate interventions can then be developed to safeguard adherence of long-term PrEP users and a run-in period may be considered before introducing the intervention.

Strengths of this study include the use of an objective outcome measure, TFV-DP measured in DBS, rather than self-reported adherence, as was used in many previous studies [34]. DBS drug levels are less sensitive to white-coat adherence (i.e. increased adherence to treatment regimens prior to a study appointment), because of the long half-life of TFV-DP in erythrocytes of 17 days [4]. Second, we used a combined end point of adherence at both 12 and 24 months for each participant. This is more robust than looking at only one measurement and it includes a longer period of PrEP use. Third, the app was designed in cooperation with community members and PrEP users, tailoring it to the needs of the end-users. Last, the limited difference in design between the intervention and control arms enabled us to specifically assess whether providing visualized feedback on self-reported PrEP use within an app can affect adherence. If we would have compared the app with all its features with a control arm without an app, it would not have been possible to assess the independent effects of – in our case – visualized feedback, as was recommended by Saberi and Johnson [11].

We acknowledge the following limitations in this study. First, all participants used an app and there was limited difference between the intervention app and the control arm. In this RCT, we could not assess the effect of using an app versus not using an app, which might have been substantial. Second, adherence in the control arm was very high (89% good adherence), leaving very limited room for improvement. Hence, in our sample-size calculations, we underestimated the adherence levels in the control arm, which reduced the power to find an effect of the intervention. Third, participants were included in the RCT at their 3-month, 6-month or 9-month study visits and at their 12 month study visit the first DBS sample was taken. Therefore, some participants had been exposed to the intervention for 3 months at the time of the DBS, whereas others had been exposed for 6 or 9 months. As this was similar between study arms, this will not have affected the comparison. Last, the AMPrEP cohort from which the RCT participants were recruited, mostly consists of relatively older, white, cisgender and highly educated people and PrEP was dispensed free of charge. Therefore, results may not be generalizable to the broader MSM and TGP populations in real-world settings where PrEP often is not obtained for free.

In conclusion, in this cohort of generally highly adherent PrEP users, we could not establish that visualized feedback on self-reported adherence through an app increases the proportion of participants with good adherence. There was a nonsignificant increase in the proportion of participants adhering excellently. Using an app in which sexual behaviour and PrEP use can be recorded and visualized could be an accessible approach to aid long-term adherence but more research is required to assess who could benefit from it the most.

Acknowledgements

The authors wish to thank all AMPrEP participants and Homeyra Amir Khosravi Aghchay, Marjo Broeren, Ertan Ersan, Princella Felipa, Merel Goos, Kees de Jong, Myra van Leeuwen, Dominique Loomans, Ilya Peters, Jason Schouten, Marjolein Stam, Adriaan Tempert and Kenneth Yap for their support of the study. Furthermore, we thank the members of the AMPrEP advisory board and the community engagement group, and all of those who contributed to the H-TEAM.

Data sharing: AMPrEP data are owned by the Public Health Service of Amsterdam. Original data can be requested by submitting a study proposal to the steering committee of AMPrEP. The proposal format can be obtained from [email protected]. Request for further information can also be submitted through the same e-mail address. The AMPrEP steering committee verifies each proposal for compatibility with general objectives, ethical approval, and informed consent forms of the AMPrEP study, and potential overlap with ongoing studies. There are no restrictions to obtaining the data, and all data requests will be processed in a similar way.

Authors’ contributions: R.A., E.H., U.D., H.d.V., M.P. and M.S.v.d.L. contributed to the design and implementation of the study. M.v.d.E. and L.C. performed the data analysis and M.v.d.E. drafted the manuscript. All authors contributed to the interpretation of the results, critically revised, have read and approved the final manuscript.

Funding: The AMPrEP study received funding as part of the H-TEAM initiative from ZonMw (grant number: 522002003), the National Institute for Public Health and the Environment, GGD research funds, Gilead Sciences and the H-TEAM. The study drug and an unrestricted research grant for AMPrEP was provided by Gilead Sciences. The H-TEAM initiative is supported by the Aidsfonds Netherlands (grant number: 2013169), Stichting Amsterdam-Dinner Foundation, Gilead Sciences Europe Ltd (grant number: PA-HIV-PREP-16-0024), Gilead Sciences (protocol numbers: CONL-276-4222, CO-US-276-1712), Janssen Pharmaceuticals (reference number: PHNL/JAN/0714/0005b/1912fde), M.A.C. AIDS Fund and ViiV Healthcare (PO numbers: 3000268822, 3000747780).

Conflicts of interest

The study drug was donated by Gilead Sciences. E.H. obtained advisory board fees from Gilead Sciences, which were paid to her institute. H.d.V. report grants from ZonMW, grants from National Institute for Public Health and the Environment and GGD research funds, nonfinancial support and grants from Gilead Sciences, grants from Aidsfonds Netherlands, grants from Stichting Amsterdam Dinner Foundation, grants from Gilead Sciences Europe Ltd, grants from Janssen Pharmaceuticals, grants from M.A.C. AIDS Fund and grants from ViiV Healthcare provided to his institute during the conduct of the study. U.D. reports nonfinancial support and unconditioned grants for conduction of AMPrEP and No-more C studies and speaker fees, all provided/paid to his institute. M.P. reports unrestricted research grants and speaker fees from Gilead Sciences, Roche, Abbvie and MSD, all of which were paid to her institute, during the conduct of the study. P.A. received personal fees and research funding from Gilead Sciences paid to his institute. The other authors reported no conflicts of interests.

References

1. van Sighem AI, Wit FWNM, Boyd A, Smit C, Matser A, Reiss P. Monitoring Report 2019. Human immunodeficiency virus (HIV) infection in the Netherlands. Amsterdam: Stichting HIV Monitoring; 2020.
2. Fonner VA, Dalglish SL, Kennedy CE, Baggaley R, O’Reilly KR, Koechlin FM, et al. Effectiveness and safety of oral HIV preexposure prophylaxis for all populations. AIDS 2016; 30:1973–1983.
3. Castillo-Mancilla JR, Zheng J-H, Rower JE, Meditz A, Gardner EM, Predhomme J, et al. Tenofovir, emtricitabine, and tenofovir diphosphate in dried blood spots for determining recent and cumulative drug exposure. AIDS Res Hum Retrovirus 2013; 29:384–390.
4. Anderson PL, Liu AY, Castillo-Mancilla JR, Gardner EM, Seifert SM, McHugh C, et al. Intracellular tenofovir-diphosphate and emtricitabine-triphosphate in dried blood spots following directly observed therapy. Antimicrob Agents Chemother 2017; 62:e01710–e01717.
5. Fuchs JD, Stojanovski K, Vittinghoff E, McMahan VM, Hosek SG, Amico KR, et al. A mobile health strategy to support adherence to antiretroviral preexposure prophylaxis. AIDS Patient Care STDs 2018; 32:104–111.
6. Liu AY, Vittinghoff E, von Felten P, Rivet Amico K, Anderson PL, Lester R, et al. Randomized controlled trial of a mobile health intervention to promote retention and adherence to preexposure prophylaxis among young people at risk for human immunodeficiency virus: the EPIC Study. Clin Infect Dis 2018; 68:2010–2017.
7. Nguyen LH, Tran BX, Rocha LE, Nguyen HLT, Yang C, Latkin CA, et al. A systematic review of eHealth interventions addressing HIV/STI prevention among men who have sex with men. AIDS Behav 2019; 23:2253–2272.
8. Moore DJ, Jain S, Dube MP, Daar ES, Sun X, Young J, et al. Randomized controlled trial of daily text messages to support adherence to preexposure prophylaxis in individuals at risk for human immunodeficiency virus: the TAPIR Study. Clin Infect Dis 2018; 66:1566–1572.
9. Grinsztejn B, Hoagland B, Moreira RI, Kallas EG, Madruga JV, Goulart S, et al. Retention, engagement, and adherence to preexposure prophylaxis for men who have sex with men and transgender women in PrEP Brasil: 48 week results of a demonstration study. Lancet HIV 2018; 5:e136–e145.
10. Park LG, Howie-Esquivel J, Dracup K. A quantitative systematic review of the efficacy of mobile phone interventions to improve medication adherence. J Adv Nurs 2014; 70:1932–1953.
11. Saberi P, Johnson MO. Technology-based self-care methods of improving antiretroviral adherence: a systematic review. PLoS One 2011; 6:e27533.
12. Maloney KM, Bratcher A, Wilkerson R, Sullivan PS. Electronic and other new media technology interventions for HIV care and prevention: a systematic review. J Int AIDS Soc 2020; 23:e25439.
13. Demonceau J, Ruppar T, Kristanto P, Hughes DA, Fargher E, Kardas P, et al. ABC project team. Identification and assessment of adherence-enhancing interventions in studies assessing medication adherence through electronically compiled drug dosing histories: a systematic literature review and meta-analysis. Drugs 2013; 73:545–562.
14. Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche PC, Devereaux P, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. Int J Surg 2012; 10:28–55.
15. Schulz KF, Altman DG, Moher D. CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. Trials 2010; 11:c332.
16. Hoornenborg E, Achterbergh RC, van der Loeff MFS, Davidovich U, van der Helm JJ, Hogewoning A, et al. Men who have sex with men more often chose daily than event-driven use of preexposure prophylaxis: baseline analysis of a demonstration study in Amsterdam. J Int AIDS Soc 2018; 21:e25105.
17. Finkenflügel RN, Hoornenborg E, Achterbergh RC, Marra E, Davidovich U, de Vries HJ, et al. Amsterdam PrEP Project team in the HIV Transmission Elimination AMsterdam Initiative (H-TEAM). A mobile application to collect daily data on preexposure prophylaxis adherence and sexual behavior among men who have sex with men: use over time and comparability with conventional data collection. Sex Transm Dis 2019; 46:400.
18. Grov C, Parsons JT, Bimbi DS. Sexual compulsivity and sexual risk in gay and bisexual men. Arch Sex Behav 2010; 39:940–949.
19. Kalichman SC, Johnson JR, Adair V, Rompa D, Multhauf K, Kelly JA. Sexual sensation seeking: acale development and predicting AIDS-risk behavior among homosexually active men. J Pers Assess 1994; 62:385–397.
20. T. Parsons DB, Perry N, Halkitis, Jeffrey. Sexual compulsivity among gay/bisexual male escorts who advertise on the Internet. Sex Addict Compulsivity 2001; 8:101–112.
21. Kelly MJ, Dunstan FD, Lloyd K, Fone DL. Evaluating cutpoints for the MHI-5 and MCS using the GHQ-12: a comparison of five different methods. BMC Psychiatry 2008; 8:10.
22. McCormack S, Dunn DT, Desai M, Dolling DI, Gafos M, Gilson R, et al. Preexposure prophylaxis to prevent the acquisition of HIV-1 infection (PROUD): effectiveness results from the pilot phase of a pragmatic open-label randomised trial. Lancet 2016; 387:53–60.
23. Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro MG. AUDIT: the Alcohol Use Disorders Identification Test. Guidelines for use in primary care. Geneva: World Health Organization; 2001.
24. Berman AH, Bergman H, Palmstierna T, Schlyter F. Evaluation of the Drug Use Disorders Identification Test (DUDIT) in criminal justice and detoxification settings and in a Swedish population sample. Eur Addict Res 2005; 11:22–31.
25. Zheng J-H, Rower C, McAllister K, Castillo-Mancilla J, Klein B, Meditz A, et al. Application of an intracellular assay for determination of tenofovir-diphosphate and emtricitabine-triphosphate from erythrocytes using dried blood spots. J Pharm Biomed Anal 2016; 122:16–20.
26. Anderson PL, Glidden DV, Liu A, Buchbinder S, Lama JR, Guanira JV, et al. Emtricitabine-tenofovir concentrations and preexposure prophylaxis efficacy in men who have sex with men. Sci Translat Med 2012; 4:151–225.
27. Grant RM, Anderson PL, McMahan V, Liu A, Amico KR, Mehrotra M, et al. Uptake of preexposure prophylaxis, sexual practices, and HIV incidence in men and transgender women who have sex with men: a cohort study. Lancet Infect Dis 2014; 14:820–829.
28. Brooks KM, Castillo-Mancilla JR, Blum J, Huntley R, MaWhinney S, Alexander K, et al. Increased tenofovir monoester concentrations in patients receiving tenofovir disoproxil fumarate with ledipasvir/sofosbuvir. J Antimicrob Chemother 2019; 74:2360–2364.
29. Songtaweesin WN, Kawichai S, Phanuphak N, Cressey TR, Wongharn P, Saisaengjan C, et al. Youth-friendly services and a mobile phone application to promote adherence to preexposure prophylaxis among adolescent men who have sex with men and transgender women at-risk for HIV in Thailand: a randomized control trial. J Int AIDS Soc 2020; 23 (S5):e25564.
30. Liu AY, Laborde ND, Coleman K, Vittinghoff E, Gonzalez R, Wilde G, et al. DOT Diary: Developing a novel mobile app using artificial intelligence and an electronic sexual diary to measure and support PrEP adherence among young men who have sex with men. AIDS Behav 2020; 25:1001–1012.
31. Perera AI, Thomas MG, Moore JO, Faasse K, Petrie KJ. Effect of a smartphone application incorporating personalized health-related imagery on adherence to antiretroviral therapy: a randomized clinical trial. AIDS Patient Care STDs 2014; 28:579–586.
32. LeGrand S, Knudtson K, Benkeser D, Muessig K, McGee A, Sullivan PS, et al. Testing the efficacy of a social networking gamification app to improve pre-exposure prophylaxis adherence (P3: Prepared, Protected, emPowered): protocol for a randomized controlled trial. JMIR Res Protoc 2018; 7:e10448.
33. Whiteley L, Mena L, Craker LK, Healy MG, Brown LK. Creating a theoretically grounded gaming app to increase adherence to preexposure prophylaxis: lessons from the development of the viral combat mobile phone game. JMIR Serious Games 2019; 7:e11861.
34. Marcus JL, Buisker T, Horvath T, Amico KR, Fuchs JD, Buchbinder SP, et al. Helping our patients take HIV preexposure prophylaxis (PrEP): a systematic review of adherence interventions. HIV Med 2014; 15:385–395.
Keywords:

dried blood spot testing; eHealth; HIV; MSM; mHealth; mobile applications; preexposure prophylaxis; treatment adherence and compliance

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.