Recent mobile health interventions to support medication adherence among HIV-positive MSM

Purpose of review We describe recent mobile health (mHealth) interventions supporting antiretroviral therapy (ART) medication adherence among HIV-positive MSM. Recent findings Keyword searches (1 January 2016–13 May 2017) identified 721 citations. Seven publications reporting on six studies met inclusion criteria. Five studies focused on MSM. Interventions primarily employed text messaging (n = 4), whereas two focused on smartphone apps and one on social media. Three studies measured intervention impact on adherence and found increased ART use intentions (n = 1), self-reported adherence (n = 1), and viral suppression (n = 1, no control group). Other mHealth interventions for HIV-positive MSM focused on status disclosure and reducing sexual risk. Summary mHealth interventions to support ART adherence among MSM show acceptability, feasibility, and preliminary efficacy. No recent mHealth interventions for MSM measured impact on viral suppression compared with a control condition despite earlier (pre-2015) evidence for efficacy. Studies are underway that include multiple features designed to improve adherence within complex smartphone or internet-based platforms. Areas for future growth include overcoming measurement and engagement challenges, developing tools for coordinating patient and provider adherence data, testing combination interventions, and adapting efficacious interventions for new languages and geographic settings.


INTRODUCTION
Globally, gay, bisexual, and other MSM are disproportionately impacted by HIV [1,2 & ,3]. MSM are estimated to account for 18-49% of all new HIV infections worldwide [4], with larger proportions in high-income countries such as the United States (67% of all new diagnoses in 2014) [5]. Treatment with antiretroviral therapy (ART) to achieve viral suppression improves individual health [6] and prevents onward transmission [7,8 && ,9 && ,10]. Sustained adherence to ART is critical for long-term viral suppression but remains suboptimal among MSM, ranging from 16 to 74% across 12 countries [11]. In the United States, only half of all HIV-diagnosed MSM are estimated to achieve viral suppression [12], with significantly lower rates documented among young MSM (6-7%) [13,14].
Societal stigma and discrimination within healthcare systems impede access to quality HIV care for MSM and influence adherence and viral suppression [15][16][17]. Many MSM manage multiple adherence barriers (e.g., mental health and unstable housing) [16,18,19], with each factor increasing the chances of having a detectable viral load [16,20 && ]. Despite clear need, few ART adherence interventions have focused explicitly on MSM [21 ]. The flexibility of mHealth interventions allows the delivery of tailored content for each user's needs [23] and accessibility helps reduce some societal and structural barriers [26]. mHealth also offers strong capability for scalability and diffusion across geographic locations, including within resource-limited settings

METHODS
We searched PubMed from 1 January 2016 to 13 May 2017 for English language publications, using combinations of keywords and MeSH terms: HIV, AIDS, antiretroviral, ART, therapy, medication, adherence, compliance, nonadheren Ã , technology, technology-based, SMS, text messag Ã , texting, online, internet, web, Web 2.0, social media, social networking, app Ã , application Ã , smartphone Ã , cell phone Ã , cellular phone Ã , mobile phone Ã , eHealth, mHealth, video conferenc Ã , videoconferenc Ã , Twitter, Grindr, Jack'd, Facebook, computeriz Ã , computer-based, virtual reality, VR, MSM, men who have sex with men, gay, homosexual, bisexual, LGBT Ã , sexual minorit Ã . Publications were excluded that did not include: an intervention description, an mHealth intervention component, an ART adherence behavioral (e.g., self-reported medication taking, device recorded medication taking) or biomedical (e.g., viral load, viral suppression, and drug levels) outcome, or a focus on MSM populations or report including MSM subpopulations. Three authors reviewed citations and full texts were pulled for all citations noted as relevant by at least one reviewer. Of the 721 articles extracted, seven articles reporting on six interventions met inclusion criteria (Table 1).

RECENT ANTIRETROVIRAL THERAPY MOBILE HEALTH ADHERENCE INTERVENTIONS TAILORED FOR OR INCLUSIVE OF MSM
For core mHealth components, most interventions used text messaging, including three pilot studies, one randomized controlled trial (RCT) [ ]. Below we highlight a few findings and novel features from these interventions.

KEY POINTS
Recent mHealth intervention tools to support HIV medication adherence among MSM show acceptability, feasibility, and preliminary efficacy at increasing ART use intentions, self-reported adherence, and viral suppression.
Additional research is needed in the form of rigorous large-scale trials, evaluation of commercially available medication adherence apps, and development of new methods to accurately measure adherence and assess which clinically relevant barriers can be addressed using technology-based tools.
Further development needs for technology-based adherence interventions include end-user engagement features, tools for coordinating patient and provider adherence data (e.g., dashboards), and adaptation of efficacious interventions for new languages and geographic settings. Messages4Men tested brief, one-way messages in a single-session pilot study that aimed to increase intentions for behavior change among HIV-positive and HIV-negative Black and Latino MSM in three US cities [31 & ]. Eligible HIV-positive men were: 18 and older, diagnosed for at least six months, and not currently in another HIV-related study. Messages for HIV-positive MSM focused on the benefits of ART adherence to oneself and one's partner and condom use. No overall association between viewing the messages and increased intention for ART use was found. Among participants with detectable viral load, exposure to messages about benefits of ART to oneself was associated with reporting messages provided new information [adjusted odds ratio (AOR) ¼ 2.32; 95% confidence interval (CI) ¼ 1.04-5.26] and exposure to messages about ART benefit to one's partner was associated with intentions for ART use (AOR ¼ 7.69; 95% CI ¼ 1.01-50.0). These findings highlight the importance of message tailoring and targeting for achieving specific desired impact among different subpopulations of MSM -in particular continuing to emphasize the benefits of adherence for both individual health as well as its importance for reducing transmission.
Two interventions combined two-way -or bidirectional -text messaging with another intervention component. The first, by Rana et al. [32 && ], included 32 people living with HIV in the United States who were either new to care, had experienced a break in care of a year or more, or who were considered by their provider to be at-risk for ART or appointment nonadherence (14/32 had detectable viral load at baseline, 13/32 identified as MSM). Participants received 6 months of SMS about adherence and HIV care. At enrollment, participants were prompted to identify their own barriers and facilitators to care or choose from preselected barriers and facilitators adapted from the US CDC's Medical Monitoring Project. Study staff called once per month to discuss and adjust (if needed) frequency and content of SMS. Comparing baseline to followup, the proportion of virally suppressed participants increased from 56 to 78% (P ¼ 0.002 ] of a text messaging and peersupport intervention to improve retention in care and adherence for Black MSM in the United States. Peer mentors -themselves HIV-positive Black MSM -received a 2-h training on information and skills related to ART and HIV care and how to respond to participants ('mentees') in nonjudgmental, motivational ways. Participants were assisted in sending their first message to peer mentors at study enrollment. For 1 month, mentors communicated with mentees via SMS. Mentors were asked to respond to texts within 12 h and to make contact with mentees at least every 3 days. Five out of seven participants were satisfied/very satisfied with the intervention. Mentees wanted more frequent contact, whereas both mentors and mentees wanted more personalized contact (e.g., an introductory phone call). Social media was recommended as additional form of contact that could be used. No  Thus, there is interest, public health need, and scientific rationale for multicomponent, complex mHealth interventions to support ART adherence among MSM. Interventions should be responsive to current best practices for mHealth-supported ART adherence and tailored to men's expressed needs and preferences. Intervention developers may want to consider incorporating gamification to improve engagement [45,46] and dynamic tailoring based on frequent assessments -features that effectively promote ART adherence [30,47]. Several interventions are underway that do this for MSM. Thrive with Me is a theoretically based (IMB model), multicomponent ART adherence intervention for MSM age 18 years and older (majority substance users) that includes asynchronous peer communication (similar to Facebook), tailored adherence information, and SMS dose reminders, and SMS adherence, mood, and substance use self-monitoring. An 8-week pilot study showed high feasibility (90%), acceptability, and significant improvement in self-reported ART adherence [48], and the intervention is currently undergoing an RCT in New York City [49]. AllyQuest is an HIV medication adherence and social support app [for Android and iPhone (Apple Inc., Foxconn, Pegatron, Cupertino, California, USA)] designed for MSM age 18-29 years. The app components were informed by SCT and the Fogg Behavioral Model [50] of persuasive technology. In a 4-week pilot study, greater app use was associated with significant increases in knowledge about HIV and confidence in ability to reliably take ART [51]. A third example is the Youth mHealth Adherence Intervention for HIV þ YMSM (Dowshen, personal communication May 2, 2017). This tailored intervention app for Android phone will support ART adherence and engagement in care for MSM age 14-24 years and includes a customizable avatar, medication reminders and tracking, and social support features (anonymous chat, ability to send 'kudos'). Gamification features include a point system, leader board, and ability to unlock new accessories. A feasibility study (n ¼ 10) anticipated beginning enrollment in 2017 (Dowshen, personal communication May 2, 2017).
Alongside ART adherence, mHealth interventions can be used to support sexual risk reduction among HIV-positive MSM. This is a critical need for curtailing the epidemic among MSM given suboptimal adherence and viral suppression in the context of high rates of condomless anal sex ]. Second, as with most adherence interventions, the accurate measurement of adherence barriers and adherence, and connection to clinical outcomes remains a challenge. The development of mHealth interventions typically follows the pattern of formative work to identify barriers to adherence and desired mHealth intervention features, followed by usability or pilot testing, and eventually an RCT. Yet common self-report adherence barriers (e.g., 'I forgot') that arise in formative work may not correlate with clinical outcomes [ ,60]) may offer new venues in this regard. Furthermore, whereas mHealth may facilitate more frequent recording of adherence, at least one study to date found discordance between app-measured adherence and self-report and pharmacy refill, with app-measured adherence being lower [61]. This will be the case if participants do not use the app every day, emphasizing the importance of creating highly engaging apps, considering whether users should be able to back-fill app-based adherence data, including biological outcome measures (viral load), and identifying parameter estimates for adjusting app-reported data. At least two studies in the field are currently gathering these metrics [25 && ,49]. Third, as mHealth interventions become more complex, individually tailored, and multifeature, the complexity of measuring meaningful or standardized intervention exposure/dose also becomes more complicated. Extremely limited work has been done in this area for HIV interventions (Bauermeister et al., current issue), hampering the ability to identify which intervention components and features are associated with desired outcomes.
Relatedly, the chronicity of HIV raises questions about the use of mHealth to maintain longer term ART adherence beyond a typical 3-12-month intervention period. Extended utility would require resources and innovation to create content and features that are continually new, relevant, and aligned with the evolving evidence base. Further challenges stem from the need to keep up with software and hardware upgrades. The use of opensource software and existing platforms that also evolve over time and the adaptation of existing platforms for new populations can mitigate some of these resource demands. Effective, open-source interventions could also be embedded into commercial platforms (e.g., Apple Health) to maximize uptake. On the side of scientific rigor, there are unanswered questions about whether people actually need to use mHealth tools for long periods of time -or whether the tools can establish selfsustaining skills and patterned behaviors. Furthermore, in trials, participants are asked to use (or be exposed to) mHealth interventions continuously over a set amount of time. However, in real-world settings, some of these interventions may be better employed in an 'as needed' fashion. Implementation studies and dynamic intervention designs (e.g., Sequential Multiple Assignment Randomized Trials) may offer insight by providing data about long-term mHealth use and effectiveness and testing combinations of mHealth intervention approaches.
Perhaps the greatest area for growth in mHealth interventions for MSM is the exchange of evidencebased mHealth interventions or intervention components across settings and populations. This includes the tailoring and testing of effective general population mHealth interventions for use with MSM, as well as the cultural adaptation and translation to other languages of successful MSM-focused apps, SMS-based interventions, and social media approaches.

CONCLUSION
Several recent mHealth interventions developed specifically for MSM show promise for supporting ART adherence and viral suppression. However, none of the reviewed interventions for MSM measured the impact of an mHealth intervention on viral suppression compared with a control condition. Several studies are planned or underway that will assess multiple features within complex smartphone or internet-based platforms providing the opportunity to advance the evidence base for mHealth supported ART adherence as well as expand the science in mHealth measurement and engagement challenges. Combination interventions that use multiple delivery modalities are a future important area for growth alongside adapting efficacious interventions for new languages and geographic settings. Global saturation of basic phones -and increasingly smartphones -has laid the groundwork for widely scalable mHealth interventions to support ART adherence to achieve viral suppression among significantly more HIV-positive MSM. Presents the design, development, and usability testing of Epic Allies -an MSMtailored app based on information motivation and behavioral skills (IMB) theory and utilizing game mechanics, social networking features, personalization, tailored feedback messages, functional ART support (e.g., editable medication alarm, daily medication tracker), and privacy features. Participants preferred an adherence app that was informational, interactive, social, and customizable. In usability testing, the app was easy to understand and navigate, and was rated highly. This intervention is currently being tested in a US-based five-site randomized controlled trial (RCT). :S121-S127. Reports on the design, development, and pilot testing of a two-way SMS app (for Android phones) and peer support for HIV-positive Black MSM who have missed an HIV appointment, experienced a break in care or report being less than 95% ART adherent. For 1 month, mentors provide via SMS informational and motivational support about HIV; help mentees build skills to stay in care and take ART (based on IMB). Mentors' protocol: respond to texts within 12 h; contact at least every 3 days. The pilot was conducted with three mentors and eight mentees. The intervention was feasible and acceptable: 5/7 were satisfied/very satisfied. Mentees wanted more frequent contact; mentors and mentees wanted more personalized contact (e.g., introductory phone call). Social media was recommended as additional contact. No HIV care outcomes are reported. Describes the design and development of weCare -a health educator guided intervention for racially and ethnically diverse MSM age 13-34 years that will utilize participants' existing social media preferences (Facebook, in-app instant messenger) and SMS to support the full HIV Continuum of Care from prevention to treatment. weCare was developed using CBPR methods, based on social cognitive theory (SCT) and theory of empowerment education. The adherencefocused components include ART refill and daily medication reminders, problem solving for ART prescription and adherence barriers, and positive reinforcement of desired behaviors. An implementation and evaluation trial was planned to start in late 2016.

38.
&& Kanters S, Park JJ, Chan K, et al. Interventions to improve adherence to antiretroviral therapy: a systematic review and network meta-analysis. Lancet HIV 2017; 4:e31-e40. This comprehensive systematic review and meta-analysis searched the Cochrane Central Register of Controlled Trials, Embase, and MEDLINE as well as major conference databases (2013-2015) for studies published up through July 2015 that include RCTs of interventions to improve ART adherence. A major strength of this study is its inclusivity of interventions outside of the United States. A total of 85 eligible trials were identified. Key findings of effective interventions include -for ART adherence: the superiority of text messages over standard of care overall, as well as specifically in low-income and middle-income settings, and the benefit of multiple interventions over single interventions; for viral suppression: cognitive behavioral therapy and supporter interventions were significantly better than standard of care overall but not in low-income and middle-income settings. This internet-based and phone-based RCT aims to reduce sexually transmitted infections and condomless anal intercourse among HIV-positive MSM in the United States. The intervention also includes a focus on HIV-status disclosure. For 12 months, participants received either a monthly internet-based sexual behavior survey, or the same monthly survey and tailored SMS focused on personalized (algorithm-based) risk-reduction. Increase in disclosure and decrease in sexual risk was found in both study arms without significant differences. This study reports on in-depth qualitative interviews that were conducted with healthcare providers regarding acceptability and barriers to use of healthcare apps and patient dashboards as a tool of patient self-management and care. Physician concerns focused on the privacy of patient information, ease of access/use, lack of coordination between dashboards and electronic medical records, and time restrictions both during and between patient visits.

57.
& Kalichman SC, Kalichman MO, Cherry C. Forget about forgetting: structural barriers and severe nonadherence to antiretroviral therapy. AIDS Care 2017; 29:418-422. This study analyses data from 556 people living with HIV with lower than 95% ART adherence. Analysis then classified participants as 'severely nonadherent' (less than 75% ART) and 'moderately nonadherent' (75-95% ART) and compared reasons for nonadherence. Severe nonadherence was significantly associated with substance use and other structural barriers (e.g., cost of medication). The authors conclude that reminders/memory aids alone will not be enough to overcome severe nonadherence in similar populations as that studied. 58.