Published on in Vol 8, No 3 (2020): March

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/15702, first published .
Preliminary Effects of a Mobile Interactive Supervised Therapy Intervention on People Living With HIV: Pilot Randomized Controlled Trial

Preliminary Effects of a Mobile Interactive Supervised Therapy Intervention on People Living With HIV: Pilot Randomized Controlled Trial

Preliminary Effects of a Mobile Interactive Supervised Therapy Intervention on People Living With HIV: Pilot Randomized Controlled Trial

Original Paper

1Alice Lee Centre for Nursing Studies, National University of Singapore, Singapore, Singapore

2University Medicine Cluster, National University Health System, Singapore, Singapore

3Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore

4School of Computing, National University of Singapore, Singapore, Singapore

Corresponding Author:

Yan Pang, MSc

Alice Lee Centre for Nursing Studies

National University of Singapore

NUHS Tower Block Level 8, 1E Kent Ridge Road

Singapore, 119228

Singapore

Phone: 65 60015375

Email: e0011744@u.nus.edu


Background: As people living with HIV infection require lifelong treatment, nonadherence to medication will reduce their chance of maintaining viral suppression and increase the risk of developing drug resistance and HIV transmission.

Objective: This study aimed to evaluate the efficacy of a mobile app, Mobile Interactive Supervised Therapy (MIST), for improving adherence to oral HIV medications among HIV-infected adults in Singapore.

Methods: We conducted a two-group pilot randomized controlled trial (RCT) with a process evaluation, in which 40 HIV-infected participants with once-daily medication regimes were recruited from a public tertiary hospital in Singapore and randomly assigned equally to either the intervention (receiving MIST and routine care) or control (receiving routine care only) groups. The intervention lasted for 2 months. The outcome of antiretroviral therapy (ART) adherence was measured by a 7-day recall self-report (SR), pill count (PC), an electronic medical device—Medication Event Monitoring System (MEMS)—and a mobile app—MIST (for the intervention group only). In total, 20 participants from the intervention group were interviewed at the end of the intervention to assess the acceptability of MIST. Data were collected at baseline and at 1-month and 2-month postintervention.

Results: All participants had excellent medication adherence at baseline (median 100, IQR 100-100). The use of MIST did not result in a significant improvement in ART adherence when measured by the SR, PC, and MEMS, as compared with the control group at 1-month (P values >.99, .86, and .74, respectively) and 2-month (P values=.80, .84, and .82, respectively) postintervention. ART adherence also did not improve in each group over the same period. MIST was perceived to be a beneficial tool based on the process evaluation results.

Conclusions: Although MIST did not enhance medication adherence to HIV treatments, mainly owing to the ceiling effect, it was perceived to be beneficial among the participants of this study. Our process evaluation provided useful data to further develop MIST for bigger and long-term mobile phone app–assisted intervention RCTs in the future.

Trial Registration: ClinicalTrials.gov NCT03794648; https://clinicaltrials.gov/ct2/show/NCT03794648

JMIR Mhealth Uhealth 2020;8(3):e15702

doi:10.2196/15702

Keywords



Background

HIV infection is a global public health issue. An estimated 36.9 million people were living with HIV in 2017 [1]. In the same year, 7982 residents of Singapore were living with HIV, with 434 being new cases [2]. In people living with HIV, who require lifelong treatment, nonadherence to medication not only reduces their chance of maintaining viral suppression but also increases the risk of developing drug resistance and HIV transmission [3-6]. To achieve viral suppression, a patient needs to maintain an adherence level of at least 90% throughout the treatment period [6-8]. However, there are challenges to meeting this target. A meta-analysis showed that only 62% of patients of a pooled study aged older than 18 years taking prescribed highly active antiretroviral therapy reported an adherence rate of ≥90% [9]. There are no available antiretroviral therapy (ART) adherence data in Singapore.

With phone technology evolving, mobile phones and smartphones are now able to incorporate many creative features such as recording and transmitting high-quality videos and telemedicine capabilities. There have been a growing number of mobile app reminders supporting patients’ adherence to medication in the market; however, there is a lack of robust evidence to establish their efficacy. To date, the quality of these supporting systems has been variable [10,11]. In addition, current evidence supporting the effectiveness of mobile phone app–based interventions in enhancing adherence among people living with HIV is unclear because of limited evidence and the heterogeneity of the study populations, treatments, and ART adherence measurements [12-17].

Aims

This study aimed to examine the effects of a mobile phone app, Mobile Interactive Supervised Therapy (MIST), for improving adherence to oral HIV medications among HIV-infected adult patients in Singapore and to gather user experiences to develop the intervention for large scale randomized controlled trials (RCTs) in the future.

Hypotheses

The hypotheses were as follows:

  1. When compared with the control group, patients in the MIST intervention group will have better HIV medication adherence rates at the 1-month and 2-month follow-up time points.
  2. When compared with baseline, patients in the MIST intervention group will have better HIV medication adherence rates at the end of the study.

Design

A two-group pretest and posttest RCT design was used. Adult patients infected with HIV (n=40) were recruited from a public hospital in Singapore. Recruited participants were randomly assigned into either of the two groups: the intervention group or the control group.

Participants

All adults infected with HIV who attended their routine clinic appointments in a public hospital in Singapore from March to June 2018 were approached to participate in this study. The inclusion criteria included those who were (1) aged ≥21 years, (2) taking a once daily regimen of HIV medications, (3) able to take pills orally, (4) willing and able to give informed consent, and (5) able to read and speak English or Chinese. Participants were excluded if they (1) were unable to operate a mobile phone or had an active tuberculosis infection that required directly observed therapy during the study period, (2) had substance use such as methamphetamine use, (3) had visual, speech, or hearing impairment despite the use of aids, (4) had a known medical history of psychiatric disorder(s) and were seeking any form of psychiatric treatment, (5) had a known medical history of cognitive impairment, (6) were fully dependent on a caregiver for taking medications, (7) were pregnant at the time of recruitment and data collection, (8) had a terminal illness such as cancer or late stage cardiovascular disease, or (9) had experienced bereavement within the past 6 months.

Sample Size Determination

On the basis of pragmatic considerations of the pilot study, a sample of 40 participants, with 20 in each group, was decided a priori [18-20]. All 20 participants in the intervention group were invited for the final process evaluation. There was no dropout during the follow-up period.

Randomization

After obtaining written consent, baseline sociodemographic (age, gender, ethnicity, and education level) and clinical data (ART regimen and duration), phone operation technical skills, and a 7-day recall of HIV medication adherence data were collected, and participants were randomly assigned to two groups: the intervention group (receiving the MIST intervention plus routine care) or the control group (receiving routine care only) in a 1:1 allocation ratio. Block randomization with length 4 was used [21] to ensure a balanced representation in each group [22]. Participants were each asked to open an opaque, sealed envelope with a piece of cardboard inside indicating their randomly allocated groups and were assigned in a successive order according to their enrollment sequence. Through these processes, randomization and allocation concealments were ensured [23-25]. The detailed workflow is described using the Consolidated Standards of Reporting Trial (CONSORT) flowchart (Figure 1).

Figure 1. The consolidated standards of reporting trial chart. MEMS: Medical Event Monitoring System; MIST: Mobile Interactive Supervised Therapy; PC: pill count; SR: 7-day recall self-report.
View this figure

Mobile Interactive Supervised Therapy Development and Intervention

MIST was developed based on the Theory of Planned Behavior. The MIST system consisted of three components: electronic reminder notifications from the health care system to the patient, transmissions of a daily video of pill-taking from the patient to the health care system, and a color-coded calendar tracker to allow the patient to compare his or her adherence data on any given day with the past week, month, and overall. Screenshots of the MIST features can be found in Figure 2. The previous prototype was developed for both Android and iOS operating systems. Our past pilot study using the earlier prototype showed that MIST was useful in measuring adherence and was acceptable to 42 healthy volunteers [26]. However, as MIST was still a pilot system focusing on producing a functional system, some bugs were identified during pilot testing. Further work was needed to improve the design and usability of the system to ensure the system was robust across a range of users. Since then, MIST has been further developed including fixing bugs such as video upload failure and app instability and was tested among a wide spectrum of phone models during 2016 to 2017. In addition, a reward system was incorporated to incentivize adherence. Before use in this study, the new version was tested in 10 healthy volunteers to ensure technical feasibility.

Participants in the intervention group were asked to use MIST, which comprised integrated Web- and mobile phone–based components. The Web component of the system allowed the administrator to set the time of pill taking each day and to review the submitted videos. The time of the reminder was tagged to the participants’ existing daily ART dosing schedules. Once a daily pill dose time was set, the Web system sent pill reminders (via an SMS or a push notification) to the participant’s mobile phone 15 min before the pill time, right on the pill time, and half an hour after the pill time until a video was received. If no response was received within half an hour after the pill time, the notifications would stop, and the video would be logged as missing. When the participants logged their medications on time (between 15 min before and after pill set timing), the frequency of reminders decreased. The Web system stored and uploaded date- and time-stamped videos onto a secure cloud server linked only to a subject number. Successfully uploaded videos were automatically deleted from the phone. Participants who were randomized to the intervention group received one Xiaomi 4A phone (using the Android system) with the MIST app already installed. The standardized phone allowed the app supporter to rapidly resolve any technical issues experienced by participants.

A research assistant (RA) reviewed the uploaded videos daily to ensure that these were of acceptable quality and confirmed pill ingestion. Participants who failed to send a video on time or sent a video that did not permit a verification of pill ingestion were required to provide a reason in their next log-in. If no reason was provided, the RA would contact the participant the next day and ask for an explanation.

The calendar cell colors were synchronized with color codes: a green-colored cell indicated medication adherence, a red-colored cell indicated that the medication was not taken, and a yellow-colored cell indicated that the video was pending the reviewer’s grading. The participants could track their compliance by tapping on each cell of the calendar matrix. There was a built-in reward system that served to reward positive behavior. A reward of Singapore $5 (approximately US $4) was given for every five consecutive successful videos that verified pill intake. The balance could be accumulated and stored in the virtual wallet module of the MIST app and paid to the participants at the final study visit. A maximal reward of Singapore $60 (approximately US $43) would be paid to the MIST users upon completion of the 2-month study.

The participants in the intervention group received a step-by-step guided tour of all the MIST app features via face-to-face demonstrations. All participants verbally indicated that they were confident in using MIST after the demonstrations.

Figure 2. Screenshots of Mobile Interactive Supervised Therapy features.
View this figure

Outcome Measures and Instruments

The HIV medication adherence rates during the 2-month study period were measured with a combination of objective (pill count [PC], Medication Event Monitoring System [MEMS], and MIST) and subjective (7-day recall self-report [SR]) measures as suggested by the literature [27,28]. We considered taking at least 95% of the prescribed medication as a cut-off to define good adherence in this study as suggested by the literature [29,30].

7-Day Recall Self-Report

Participants were asked about the number of HIV medications they missed in the preceding 7 days at the baseline visit and during the two follow-up visits (1-month visit and 2-month visit). The self-reported medication adherence rate was calculated by subtracting the number of prescribed doses and the number of self-reported missed doses and dividing it by the number of prescribed doses in the past 7 days.

Pill Count

The HIV adherence rate was calculated by subtracting the number of prescribed doses in the past 30 days and the number of missed doses in the past 30 days and dividing it by the number of prescribed doses in the past 30 days [31].

Medication Event Monitor System

MEMSCap Medication Event Monitoring System, developed by Aardex Group, has a pill bottle with a special cap that contains integrated microcircuits to record the date and time whenever a patient opens a bottle, which wirelessly transfers dosing data when used in conjunction with an MEMSCap reader. A secure, Web-based data platform, AARDEX, processes the transferred adherence data using validated algorithms and presents the information in easy-to-interpret graphs and tables [32-34]. The medication adherence rate is calculated based on the number of times the MEMSCap has been opened divided by the number of times the MEMSCap is scheduled to be opened.

Mobile Interactive Supervised Therapy App

Our previous work has demonstrated that MIST is able to accurately measure medication adherence among healthy volunteers in a small pilot study [26]. The ART adherence rate measured by MIST is calculated based on the number of videos received confirming pill intake plus the number of videos not received due to technical problems divided by the number of videos expected during the 2 months.

Process Evaluation

Acceptability and perceptions on the use of the MIST app were assessed during the final study visit. The RA, who was bilingual in English and Chinese and trained in qualitative research, conducted the semistructured interviews. The interview questions were developed based on the research aims in addition to literature review. Two experts were invited to review the interview guide’s contents upon its development, and minor revisions were made thereafter. The interview was pilot-tested on 3 healthy volunteers to check the appropriateness of the questions as well as the interview process. Further revisions were made to make the questionnaire understandable. The interview mainly addressed five areas: perceptions of the overall MIST app user experience, perceived usefulness of each feature, strengths and weaknesses of MIST, interest in continuing to use MIST in the future, and recommendations for further improvements. A total of 20 individual face-to-face interviews were conducted, lasting for 3 to 13 min.

Study Procedure

Data collection commenced after ethical approval had been obtained. The primary physicians of the potential recruits were responsible for introducing the participants to the RA. The RA passed a short list of potential recruits to their primary physicians during their scheduled clinic appointments. The primary physicians would briefly introduce the study to their patients. Those who showed interest were then introduced to the RA who would further explain the study to them in a separate private room. Only after obtaining a written consent would the potential participant be recruited into the study. This was followed by a collection of demographic and clinical data and baseline data via a self-administered questionnaire before randomization took place. Outcomes were measured at the following time points for all participants from the two groups: (1) 1 month after the intervention (posttest 1) and (2) 2 months after the intervention (posttest 2).

The RA set up a unique account that was password protected for each participant in the intervention group to log into the MIST app at the baseline visit. The MIST app was activated in the provided Xiaomi 4A mobile phones, and the participants were provided with detailed face-to-face instructions on how to use it. The same RA was responsible for the collection of all data, including process evaluation interviews.

Data Analysis

Quantitative analyses were performed using the Statistical Package for the Social Sciences (IBM, Version 21.0 [35]).P<.05 was considered statistically significant. Descriptive statistics were used to describe demographic data, clinical data, and baseline self-reported medication adherence rates for both groups. Either the chi-square test, Fisher exact test, or the Mann-Whitney U test was used to assure comparability between the intervention and control groups.

To answer hypothesis 1, the Mann-Whitney U test was used to examine differences in ART adherence rates as measured by the 7-day recall SR, PC, and MEMS between the two groups at the 1-month and 2-month visits.

To answer hypothesis 2, the Wilcoxon signed-rank test was used to compare ART adherence rates (measured by PC, MEMS, and with or without MIST) at different time points (1-month and 2-month visits) for each group, and the Friedman test was used to compare the ART adherence rates (measured by SR) at baseline and the two follow-up visits.

Qualitative data from the interviews were analyzed using content analysis. The audio-taped interview data were transcribed verbatim by the RA to capture nonverbal nuances. A range of highlighter colors were used to highlight segments of data portraying similar ideas and meaning. Subsequently, data extracts that have been coded with the same color were collated and organized into categories. The refined potentially repeated patterns across the entire data were, then, transferred to a separated word document. The initial codes were reviewed again by the RA and another investigator, and related codes were collated to form subthemes [36]. Different opinions were discussed until a mutual agreement was reached between the two investigators [37]. Rigor was ensured through credibility, transferability, dependability, and confirmability in this study process [38].

Ethical Considerations

Ethics approval for the study was received from the institutional review board before commencing the study (Ref: 2017/00150). All participants provided written consent for audio-recordings of the interviews. Confidentiality of the data and voluntary participation were explained to all participants.


Participants’ Demographics and Clinical Data and Group Comparison

Among the 40 participants who participated in the study, 38 (95%) were male and 25 (63%) were Chinese, 8 (20%) were Malay, 3 (8%) were Indian, and 4 (10%) were from other ethnic groups. The median age was 37.45 (IQR 30.07-44.95) years. In total, 21 of 40 (53%) participants were doing professional work. Nearly half of them (19/40, 48%) completed university or higher education. Most were single (32/40, 80%), lived with family (29/40, 73%), lived in a Housing and Development Board flat or a studio apartment (36/40, 90%), and had home Wi-Fi (29/40, 73%). The median ART durations at the baseline were 25.15 (IQR 12.23-48.03) months and 22.35 (IQR 8.93-43.28) months for the control and intervention groups, respectively. Taking two types of tablets of HIV medication was the most commonly reported regime (16/40, 40%) among the participants, followed by one type of tablet (11/40, 28%), three types of tablets (8/40, 20%), and at least four types of tablets (5/40, 13%). The vast majority of the participants were very confident in operating smartphones (34/40, 85%). As shown in Multimedia Appendix 1, baseline characteristics were similar between the control and interventional groups except for types of ART medication (P=.01).

Comparison of Outcomes

As shown in Table 1, there were no significant differences in the medication adherence rates between the intervention and control groups at the follow-up time points. In addition, neither of the measures reported significant differences in ART adherence rates between the two groups across the 2-month period (Table 2).

Table 1. Comparison of median antiretroviral therapy adherence rates between the control group and intervention group at the 1-month and 2-month follow-up visits (N=40).
Follow-up visit and ARTa adherence rate (%) measurementControl group, median (IQR)Intervention group, median (IQR)P value
1 month

Self-report100 (100-100)100 (100-100)>.99

Pill count100 (100-100)100 (100-100).86

MEMSb median100 (93-100)100 (93-100).74
2 months

Self-report100 (100-100)100 (100-100).80

Pill count100 (100-100)100 (100-100).84

MEMS median97 (93-100)97.5 (92-100).82

aART: antiretroviral therapy.

bMEMS: Medication Event Monitoring System.

Table 2. Comparison of median antiretroviral therapy adherence rates at different time points for each group (N=40).
ARTa adherence rate (%) measurementBaseline, median (IQR)1-month visit, median (IQR)2-month visit, median (IQR)P value
Self-report.22b

Control group100 (100-100)100 (100-100)100 (100-100)

MISTc intervention group100 (100-100)100 (100-100)100 (100-100)
Pill count.07d

Control groupN/Ae100 (100-100)100 (100-100)

MIST intervention groupN/A100 (100-100)100 (100-100)
MEMSf.63d

Control groupN/A100 (93-100)97 (93-100)

MIST intervention groupN/A100 (93-100)97.5 (92-100)
MIST

MIST intervention groupN/A96 (87-100)94 (87-99).25d

aART: antiretroviral therapy.

bFriedman test.

cMIST: Mobile Interactive Supervised Therapy.

dWilcoxon signed-rank test.

eN/A: Not applicable.

fMEMS: Medication Event Monitoring System.

Process Evaluation

All 20 participants from the intervention group completed the semistructured interviews. Among 20 participants who were interviewed, the majority of them were extremely or very confident (18/20, 90%) in operating a mobile phone. Participants were aged between 27 and 52 years. Most were males (n=19) and Chinese (n=13). Half of them were doing professional work, 2 were doing skilled work, while 7 were unemployed, and 1 was a student. Slightly over half of them completed university or higher education, 35% (7/20) completed a college, and the remaining 10% (2/20) received a secondary school education. The majority of them were single (16/20, 80%) and lived with family or friends. Findings related to the MIST app user experiences were grouped into six categories: (1) perceived ease of use, (2) benefits of MIST, (3) MIST app features preferences, (4) MIST app dislikes, (5) future willingness to use MIST, and (6) suggestions for MIST improvement.

Category 1: Perceived Ease of Use

When asked about their overall experience with the MIST app, half of the participants reported that the app was user-friendly.

Category 2: Benefits of Mobile Interactive Supervised Therapy

Most of the participants considered MIST to be beneficial. It not only reminded and motivated them to take their ART medications punctually and consistently every day but also helped in situations when they were tired and forgot to take their medications. They also agreed the color-coded calendar helped them to track their medication-taking histories. In all, 2 participants verbalized that receiving cash helped to lighten their financial burdens.

Category 3: Mobile Interactive Supervised Therapy App Feature Preferences

When asked for their preferred app features, the participants voted almost in equal numbers for each feature, that is, SMS reminders, color-coded calendars, and the reward system were voted by 7, 6, and 5 participants, respectively.

Categories 4 and 5: Mobile Interactive Supervised Therapy App Dislikes and Willingness for Future Utilization

Some participants revealed that there were certain aspects of the MIST app that they disliked, and these included the need for an internet connection, app glitches and instability, limitation to mobile phones using the Android platform, extra burdens (such as recording themselves taking their medications), and privacy concerns. Due to these concerns, among the 20 participants who were interviewed, only 6 (30%) participants expressed their willingness to use MIST in the future.

As there was stigma associated HIV infections, the participants tried to minimize receiving SMS reminders or taking their medications in public to prevent attracting attention. For instance, 1 participant switched off the provided phone as he was unsure what the contents of the incoming SMS reminders would look like. Another participant was concerned that recording himself taking the medication would trigger curious onlookers. A third participant suggested that MIST should have an option that allowed the users to change the app icon image as the default icon might inadvertently reveal their HIV status.

Category 6: Suggestion for Mobile Interactive Supervised Therapy Improvement

Most of the participants suggested making the app features more user-friendly. The MIST app should be made available for all types of phones. They also suggested alternative reward methods and methods to track the medications taken. Some thought that the app should incorporate a fun element to engage future users.


Principal Findings

The pilot trial on the use of MIST, which followed patients infected with HIV for 2 months, did not show any significant improvements in ART adherence rates. However, this finding needs to be interpreted with caution as this was just a pilot study and the sample recruited were those who had good medication adherence. On the basis of the process evaluation, MIST was perceived to be beneficial among our participants. Our process evaluation provided useful data to further develop MIST for bigger and long-term mobile phone app–assisted intervention RCTs in the future. The reason for the lack of improvement was the ceiling effect. Before the study, both groups of participants were already adhering to their medications. For example, the median 7-day self-recalled ART adherence at the baseline was 100% in both groups, leaving no room for them to improve their adherence further. Given a longer enrollment time, we would have liked to target subjects who were at high risk of nonadherence. However, our subject recruitment was limited by the small number of newly diagnosed patients annually. In addition, the recruitment was hampered by perceived stigma in people living with HIV and concern that the intervention might compromise confidentiality and privacy. After a slow start of having only 3 patients recruited, we expanded our eligibility criteria to include those who had no adherence problems as well. However, the rejection rate was still nearly 48%. The study ended with most of the participants already having good ART adherence. The selection bias was likely to contribute to our observation that patients with poor ART adherence might be also those who were less interested to participate in the trial. In support of our assertion of the ceiling effect, one study [15] did not have the ceiling effect as it only enrolled participants who reported less than 95% adherence and demonstrated adherence improvements following their mobile phone app use. The second possible reason for the lack of improvement in our study is that most of our participants were already on ART for about 1 year, as evidenced by the median ART duration (IQR) for the control group of 25 months (IQR 12-48 months) and the intervention group of 22 months (IQR 8-43 months) before the study enrollment. The need for lifelong daily ART adherence would have developed among them a well-established daily medication-taking routine [39]. In all, two previous phone technology–assisted adherence RCTs [40,41], which targeted only patients newly initiated onto ART, demonstrated the effectiveness of SMS or SMS combined with counseling in improving medication adherence. The results highlighted the importance of initiating the intervention at the early stage before the participants establish medication-taking habits [40,41].

Despite the absence of an effect of MIST on ART adherence among the participants who had 100% adherence rates at baseline, our qualitative findings demonstrated that the MIST app was beneficial to most participants. They perceived MIST to be user-friendly. It is interesting to note that although most of our participants did not forget to take their medications, the reminders were especially helpful for others who did shift work or who were busy with other things at the time their medications were due. This finding is consistent with a previous RCT [39], in which some participants who did not benefit from reminders would nevertheless recommend it to someone who already had a fixed HIV medication–taking routine [39]. Some perceived the color-coded calendar tracking system to be more useful compared with a conventional tracking tool, such as a diary or pillbox, or even no tracking tool. The visual calendar medication-taking tracking feature helped our participants to save time and, more importantly, motivated them to continue their good medication-taking behaviors. Our study findings are consistent with previous research, which suggested that an app with health-related behaviors and goal tracking features, such as medication-taking behaviors, was perceived as valuable [42].

We believe the following reasons contributed to the reluctance of most of the participants (14/20) in continuing to use MIST in the future. First, the availability of the latest prototype version was limited to the Android operating system as it was still being optimized across all Android operating system phones during the study period. Owing to this limitation, we had to provide each participant with a Xiaomi 4A model phone that contained the MIST app. Failure to carry the phone, turn it on, or have the battery charged prevented the reminders from serving their desired purpose. Second, occasional app glitches and instability were barriers that discouraged participants from continuing to use the MIST app. Some participants claimed that they would rather have taken their medications than wait for the app to start working. Third, some participants found video recording while taking their medications burdensome, which might not be sustainable as HIV requires lifelong medication adherence. However, in spite of this, all 20 participants maintained their commitments and continued using MIST until the end of the study. On the contrary, a trial [13] observed a 25% reduction in the use of their Heart2HAART mobile phone app during a 3-month trial period. In their study, participants in the Heart2HAART group were expected to respond to randomly generated daily medication prompts, including questions about HIV medication side effects and substance cravings during the study period. They concluded that user fatigue might have contributed to app usage reduction [13]. Our study finding stood in contrast to the previous study findings [13], possibly because of the cash reward system that was built into MIST, whereas there was no incentive system in the previous study [13]. As having cash incentives might not be a practical measure in the long term, future studies should consider alternative incentive methods.

Although the MIST app is password protected, some of the participants expressed concerns about possible disclosures of their HIV status during the interviews. Compounding the privacy concern was the bulky electronic monitor device—MEMS—which might have triggered onlookers’ curiosity and a loss of privacy thereafter. Our findings are consistent with findings from other studies. For example, a global qualitative review study revealed that the acceptability of mobile technology–assisted interventions that aimed to improve ART adherence among people living with HIV could be affected by privacy and confidentiality issues [43]. In view of the findings of this study, we will optimize MIST to make sure that it is more secure. We will modify the MIST prototype based on other suggestions of our participants. For instance, the app icon should be made customizable. MIST should be optimized to market-level quality, meaning that it should be made more sensitive and less glitchy and it should be downloadable for use in other operating systems. We will also develop a backup plan in case the MIST app fails to work. For example, verbal reports or SMS as an alternative to report adherence.

Limitations of the Study

This study had several limitations. First, owing to pragmatic considerations, the small sample size of 40 participants might have failed to detect a significant difference in medication adherence (ie, type 2 error). Second, we did not collect biological data such as CD4 count and viral load, which are good indicators for treatment success. Third, although this study was conducted in a tertiary hospital, due to the intervention itself, some older people were excluded from this study as they were not confident in using the technology. Blinding was not possible for both participants and researchers due to the nature of the app. Hence, this study might have introduced some response bias. MIST participants might have responded differently owing to social expectation pressure. Fourth, the interview duration was short. However, this was a process evaluation as opposed to an in-depth qualitative study, and all interview questions were straightforward. Furthermore, the majority of potential participants declined to be interviewed owing to stigma. Stigma could also have limited the information that enrolled participants were willing to share during interviews. Finally, because of time constraints, the majority of our participants had good baseline ART adherence resulting in the ceiling effect. The intervention was limited to a 2-month follow-up, and this short follow-up period does not reflect the real-world situation, as HIV requires lifelong commitment and adherence.

Conclusions

In conclusion, this pilot study showed that MIST was perceived as beneficial by our participants. Future RCTs with a bigger sample size and the use of an optimized and more acceptable version of MIST should be conducted to evaluate the effectiveness of MIST in enhancing medication adherence among people living with HIV, especially those who have poor adherence and those who have just started on ART.

Acknowledgments

The authors appreciate the National Health Innovation Centre (NHIC) and the National University Hospital Singapore who provided funding for this study. The authors thank all participants and clinic staff members for their support and contributions. They also thank the National University Health System Medical Publications Support Unit for assistance in language editing of this manuscript. This study was supported by the NHIC Innovation to Develop Grant (grant number: NHIC-I2D-1509082) and the National University Health System Allied Health and Nursing Grant (grant number: NUHRO/2018/043/AHN/01).

Conflicts of Interest

None declared.

Multimedia Appendix 1

Comparison of the participants’ sociodemographic, clinical, and technical skills data (N=40).

DOCX File , 20 KB

Multimedia Appendix 2

CONSORT-EHEALTH checklist V 1.6.1.

PDF File (Adobe PDF File), 2344 KB

References

  1. UNAIDS. Joint United Nations Programme on HIV/AIDS. 2017. UNAIDS Data 2018   URL: https://www.unaids.org/sites/default/files/media_asset/unaids-data-2018_en.pdf
  2. Ministry of Health Singapore. Update on the HIV/AIDS situation in Singapore 2017 (JUNE 2018)   URL: https:/​/www.​moh.gov.sg/​resources-statistics/​infectious-disease-statistics/​hiv-stats/​update-on-the-hiv-aids-situation-in-singapore-2017-(june-2018) [accessed 2019-01-20]
  3. Mills EJ, Lester R, Thorlund K, Lorenzi M, Muldoon K, Kanters S, et al. Interventions to promote adherence to antiretroviral therapy in Africa: a network meta-analysis. Lancet HIV 2014 Dec;1(3):e104-e111 [FREE Full text] [CrossRef] [Medline]
  4. Braitstein P, Brinkhof MW, Dabis F, Schechter M, Boulle A, Miotti P, Antiretroviral Therapy in Lower Income Countries (ART-LINC) Collaboration, ART Cohort Collaboration (ART-CC) groups. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet 2006 Mar 11;367(9513):817-824. [CrossRef] [Medline]
  5. Gill CJ, Hamer DH, Simon JL, Thea DM, Sabin LL. No room for complacency about adherence to antiretroviral therapy in sub-Saharan Africa. AIDS 2005 Aug 12;19(12):1243-1249 [FREE Full text] [CrossRef] [Medline]
  6. de Olalla Garcia P, Knobel H, Carmona A, Guelar A, López-Colomés JL, Caylà JA. Impact of adherence and highly active antiretroviral therapy on survival in HIV-infected patients. J Acquir Immune Defic Syndr 2002 May 1;30(1):105-110. [CrossRef] [Medline]
  7. Bangsberg DR, Deeks SG. Is average adherence to HIV antiretroviral therapy enough? J Gen Intern Med 2002 Oct;17(10):812-813 [FREE Full text] [CrossRef] [Medline]
  8. Bartlett JA. Addressing the challenges of adherence. J Acquir Immune Defic Syndr 2002 Feb 1;29(Suppl 1):S2-10. [CrossRef] [Medline]
  9. Ortego C, Huedo-Medina TB, Llorca J, Sevilla L, Santos P, Rodríguez E, et al. Adherence to highly active antiretroviral therapy (HAART): a meta-analysis. AIDS Behav 2011 Oct;15(7):1381-1396. [CrossRef] [Medline]
  10. Haase J, Farris KB, Dorsch MP. Mobile applications to improve medication adherence. Telemed J E Health 2017 Feb;23(2):75-79. [CrossRef] [Medline]
  11. Santo K, Richtering SS, Chalmers J, Thiagalingam A, Chow CK, Redfern J. Mobile phone apps to improve medication adherence: a systematic stepwise process to identify high-quality apps. JMIR Mhealth Uhealth 2016 Dec 2;4(4):e132 [FREE Full text] [CrossRef] [Medline]
  12. 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 Nov;28(11):579-586 [FREE Full text] [CrossRef] [Medline]
  13. Himelhoch S, Kreyenbuhl J, Palmer-Bacon J, Chu M, Brown C, Potts W. Pilot feasibility study of Heart2HAART: a smartphone application to assist with adherence among substance users living with HIV. AIDS Care 2017 Jul;29(7):898-904. [CrossRef] [Medline]
  14. Hightow-Weidman L, Muessig K, Knudtson K, Srivatsa M, Lawrence E, LeGrand S, et al. A gamified smartphone app to support engagement in care and medication adherence for HIV-positive young men who have sex with men (AllyQuest): development and pilot study. JMIR Public Health Surveill 2018 Apr 30;4(2):e34 [FREE Full text] [CrossRef] [Medline]
  15. Martin CA, Upvall MJ. A mobile phone HIV medication adherence intervention: acceptability and feasibility study. J Assoc Nurses AIDS Care 2016;27(6):804-816. [CrossRef] [Medline]
  16. Dworkin M, Chakraborty A, Lee S, Monahan C, Hightow-Weidman L, Garofalo R, et al. A realistic talking human embodied agent mobile phone intervention to promote HIV medication adherence and retention in care in young HIV-positive African American men who have sex with men: qualitative study. JMIR Mhealth Uhealth 2018 Jul 31;6(7):e10211 [FREE Full text] [CrossRef] [Medline]
  17. Westergaard RP, Genz A, Panico K, Surkan PJ, Keruly J, Hutton HE, et al. Acceptability of a mobile health intervention to enhance HIV care coordination for patients with substance use disorders. Addict Sci Clin Pract 2017 Apr 26;12(1):11 [FREE Full text] [CrossRef] [Medline]
  18. Julious S, Machin D, Tan SB. An Introduction to Statistics in Early Phase Trials. Hoboken, New Jersey, United States: Wiley; 2010.
  19. Billingham SA, Whitehead AL, Julious SA. An audit of sample sizes for pilot and feasibility trials being undertaken in the United Kingdom registered in the United Kingdom Clinical Research Network database. BMC Med Res Methodol 2013 Aug 20;13:104 [FREE Full text] [CrossRef] [Medline]
  20. Kieser M, Wassmer G. On the use of the upper confidence limit for the variance from a pilot sample for sample size determination. Biomed J 1996;38(8):941-949. [CrossRef]
  21. Matthews JN. Introduction to Randomized Controlled Clinical Trials. Milton Park, Abingdon, Oxfordshire: Routledge; 2006.
  22. Research Randomizer.   URL: https://www.randomizer.org/ [accessed 2018-03-15]
  23. Doig GS, Simpson F. Randomization and allocation concealment: a practical guide for researchers. J Crit Care 2005 Jun;20(2):187-91; discussion 191. [CrossRef] [Medline]
  24. Schulz KF, Grimes DA. Allocation concealment in randomised trials: defending against deciphering. Lancet 2002 Feb 16;359(9306):614-618. [CrossRef] [Medline]
  25. Lancaster GA, Dodd S, Williamson PR. Design and analysis of pilot studies: recommendations for good practice. J Eval Clin Pract 2004 May;10(2):307-312. [CrossRef] [Medline]
  26. Molton JS, Pang Y, Wang ZC, Qiu BQ, Wu P, Rahman-Shepherd A, et al. Prospective single-arm interventional pilot study to assess a smartphone-based system for measuring and supporting adherence to medication. BMJ Open 2016 Dec 20;6(12):e014194 [FREE Full text] [CrossRef] [Medline]
  27. Farmer KC. Methods for measuring and monitoring medication regimen adherence in clinical trials and clinical practice. Clin Ther 1999 Jun;21(6):1074-90; discussion 1073. [CrossRef] [Medline]
  28. Lam WY, Fresco P. Medication adherence measures: an overview. Biomed Res Int 2015;2015:217047 [FREE Full text] [CrossRef] [Medline]
  29. Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000 Jul 4;133(1):21-30. [CrossRef] [Medline]
  30. Bangsberg DR, Hecht FM, Charlebois ED, Zolopa AR, Holodniy M, Sheiner L, et al. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS 2000 Mar 10;14(4):357-366. [CrossRef] [Medline]
  31. Castillo-Mancilla JR, Haberer JE. Adherence measurements in HIV: new advancements in pharmacologic methods and real-time monitoring. Curr HIV/AIDS Rep 2018 Feb;15(1):49-59 [FREE Full text] [CrossRef] [Medline]
  32. Shuter J, Sarlo JA, Kanmaz TJ, Rode RA, Zingman BS. HIV-infected patients receiving lopinavir/ritonavir-based antiretroviral therapy achieve high rates of virologic suppression despite adherence rates less than 95%. J Acquir Immune Defic Syndr 2007 May 1;45(1):4-8. [CrossRef] [Medline]
  33. Gross R, Bellamy SL, Chapman J, Han X, O'Duor J, Palmer SC, et al. Managed problem solving for antiretroviral therapy adherence: a randomized trial. JAMA Intern Med 2013 Feb 25;173(4):300-306 [FREE Full text] [CrossRef] [Medline]
  34. Hardy H, Kumar V, Doros G, Farmer E, Drainoni M, Rybin D, et al. Randomized controlled trial of a personalized cellular phone reminder system to enhance adherence to antiretroviral therapy. AIDS Patient Care STDS 2011 Mar;25(3):153-161 [FREE Full text] [CrossRef] [Medline]
  35. IBM Corp. IBM. Armonk, NY: IBM Corp; 2012. IBM SPSS Statistics 21 Documentation   URL: https://www.ibm.com/support/pages/ibm-spss-statistics-21-documentation [accessed 2019-02-26]
  36. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol 2006;3(2):77-101. [CrossRef]
  37. Krefting L. Rigor in qualitative research: the assessment of trustworthiness. Am J Occup Ther 1991 Mar;45(3):214-222. [CrossRef] [Medline]
  38. Guba EG, Lincoln YS. Competing paradigms in qualitative research. In: Handbook of Qualitative Research. Thousand Oaks, California, United States: Sage Publications; 1994:105-117.
  39. Mao LM, Buchanan A, Wong HTH, Persson A. Beyond mere pill taking: SMS reminders for HIV treatment adherence delivered to mobile phones of clients in a community support network in Australia. Health Soc Care Community 2018 Jul;26(4):486-494. [CrossRef] [Medline]
  40. Lester RT, Ritvo P, Mills EJ, Kariri A, Karanja S, Chung MH, et al. Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): a randomised trial. Lancet 2010 Nov 27;376(9755):1838-1845. [CrossRef] [Medline]
  41. Abdulrahman SA, Rampal L, Ibrahim F, Radhakrishnan AP, Shahar HK, Othman N. Mobile phone reminders and peer counseling improve adherence and treatment outcomes of patients on ART in Malaysia: A randomized clinical trial. PLoS One 2017;12(5):e0177698 [FREE Full text] [CrossRef] [Medline]
  42. Dennison L, Morrison L, Conway G, Yardley L. Opportunities and challenges for smartphone applications in supporting health behavior change: qualitative study. J Med Internet Res 2013 Apr 18;15(4):e86 [FREE Full text] [CrossRef] [Medline]
  43. Ma QY, Tso LS, Rich ZC, Hall BJ, Beanland R, Li H, et al. Barriers and facilitators of interventions for improving antiretroviral therapy adherence: a systematic review of global qualitative evidence. J Int AIDS Soc 2016;19(1):21166 [FREE Full text] [CrossRef] [Medline]


ART: antiretroviral therapy
MEMS: Medication Event Monitoring System
MIST: Mobile Interactive Supervised Therapy
NHIC: National Health Innovation Centre
PC: pill count
RA: research assistant
RCT: randomized controlled trial
SR: self-report


Edited by G Eysenbach; submitted 31.07.19; peer-reviewed by J Zhu, C Sun; comments to author 04.10.19; revised version received 29.11.19; accepted 16.12.19; published 27.03.20

Copyright

©Yan Pang, James Steven Molton, Wei Tsang Ooi, Nicholas Iain Paton, Hong-Gu He. Originally published in JMIR mHealth and uHealth (http://mhealth.jmir.org), 27.03.2020.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.