This is a randomized controlled clinical trial to test an intervention comprising faster data capture plus targeted voice messages compared with PE monthly monitoring in a community-based sample in Cambodia for improvement of hypertension and diabetes outcomes. This study received Institutional Review Board approval from the University of Washington Division of Human Subjects and the National Ethics Committee for Health Research in Cambodia. All PEs, regardless of study allocation, and participants assigned to the telephone message intervention provided written informed consent. Informed consent from community participants who did not receive mHealth messages were waived by the Institutional Review Boards as routine monitoring provided study outcome data for these groups. This study was determined to not qualify as an “Applicable Clinical Trial” according to the National Institutes of Health definition [16] at the time it was initiated and, thus, was not registered in clinicaltrials.gov. As the risk of harm is minimal in this study, no criteria for discontinuation were developed. An Advisory Committee comprising Cambodian Ministry of Health officials, health care providers, and technology experts was convened to provide oversight and guidance to the study.

In Cambodia, where greater life expectancy is causing rapid increases in NCDs [17], nongovernmental organizations have stepped in when government programs have not been able to address disease locally. MoPoTsyo Patient Information Center is a Cambodian nongovernmental organization for people with chronic disease in Cambodia [18]. It was established in 2004 to provide an institutional and practical response to the information and care needed by patients with hypertension and diabetes. Their model recruits and involves patients as volunteers and trains them to provide counseling on NCDs and monitor key health indicators over time. PEs see patients on a monthly basis to reinforce training and monitor key metrics, including blood pressure, glucose, weight, and adherence to medications. Patients keep paper logs of their health information that are updated at each PE visit. Currently, logs summaries are transferred to the MoPoTsyo Patient Information Center for manual entry and to update patient histories. Since 2005, MoPoTsyo has trained >200 PEs who have registered >31,000 patients in 7 rural provinces and poor urban slum areas of Phnom Penh. Patient and pharmacy records have been used to document the success of the program showing dramatic improvement in chronic disease management [19].

Seven Operational Districts (ODs) representing rural geographic regions or urban slums were selected for inclusion in the project; these included 4 ODs in Kampong Speu province, 2 ODs in Kampong Thom province, 1 OD in Kampong Cham province, and the municipality of Phnom Penh. This area includes 87 PEs who were included for simple randomization (by computer-generated random numbers) into the 3 arms of the study. All patients registered into the MoPoTsyo Patient Information Center system at the time of PE randomization were included as the community sample to receive mHealth messages with no exclusion criteria. Owing to the nature of this intervention, it was not possible to blind participants to study allocation. However, as patients of PEs were clustered geographically, there was a minimum chance of patients interacting across clusters to discuss the study.

In this study, we are collaborating with MoPoTsyo Patient Information Center to enhance their Peer Educator Network model through the application of mHealth-tailored phone messaging and improved eHealth communication throughout the system. A focus on hypertension, justified by the extremely low medication adherence of hypertensive patients compared with diabetic patients, shifts the intervention to an area of great need. Two interventions will be tested, one at the PE level and one at the patient level. The first includes providing electronic tablets to PEs for data collection and transfer to the MoPoTsyo database in Phnom Penh. The tablets are intended to speed up data collection and entry, reduce paper, increase accuracy, and eliminate lengthy distances that must be traveled to bring paperwork to MoPoTsyo. In addition, the tablets will allow PEs to scan a patient’s record “handbook” (log kept with the patient) to be verified by MoPoTsyo quality control staff should data appear suspicious. At the patient level, voice messages developed to address specific patient problems (eg, uncontrolled blood pressure or glucose, medications not picked up at the pharmacy, weight gain, etc) will be sent to patients based on the data received from the monthly visits to the PE. This study has randomly selected 75 PEs for assignment to 1 of 3 intervention groups—tablet+messaging, tablet only, or control (no tablet or voice messages). Figure 1 shows the study design. As we estimate that each of the 25 PEs in each group monitors about 60 patients, each arm of the study will include about 1500 participants for a total community sample of about 4500.

Fifty PEs (25 in the tablet+messaging group and 25 in the tablet only group) received a 12-inch Chuwi Hi12 Tablet with touchscreen compatible for Windows 10-Android 5.1 to record patient information each month. Data are entered online or offline during the patient visit with backup recorded in the patient handbook (log). Data are transferred to the MoPoTsyo database stored on a Dell PowerEdge T630 Server, RAM 32GB (Gigabytes), HDD 4TB, (hard disk 4 Terabytes), speed 2.4 GHz, Cache 20MB. The operating system is Windows Server 2012R2 with database application Microsoft SQL server 2014. Initial training has been provided by MoPoTsyo staff to assure PEs are comfortable with using the tablets for their monthly monitoring visits. It has become clear, perhaps because of their older age and lack of experience, that many of the PEs are challenged by the technical aspects of this new tool. We have begun weekly Web-conferences (broadcast on the tablets) to help address problems and increase PE skills in tablet use. In addition, pharmacies have received tablets to input prescriptions and provide information to MoPoTsyo regarding invoices when medications are purchased. We do not anticipate significant dropout of PEs based on historical experience. PEs seldom leave MoPoTsyo unless illness or death intervenes. Retention efforts for community participants include the free health monitoring provided by PEs and the social interactions that result in this popular program. PEs in all groups are coached to be vigilant in responding to patient questions and needs as well as to contact the MoPoTsyo Center with any issues that arise.

We used an exploratory qualitative study design to hear patient and PE perspectives on NCD management and mHealth. We used the 32-item checklist Consolidated Criteria for Reporting Qualitative Research [20] to guide our study design, data collection, analysis, and reporting of our research study. The Information-Motivation-Behavior theoretical framework was used to guide mHealth message development [21] based on providing accurate information, personal motivation, and social motivation to impact behavioral change in chronic disease self-management. We conducted focus groups (N=59) in Khmer with MoPoTsyo patients at 6 ODs within the study territory (Kampong Speu, Chamkar Leu, Baray-Santuk, Stoong, Kong Pisey, and Phnom Penh) to better understand the content of messages that would be most effective and provide guidance on logistics for sending them. Focus groups were held at community Health Centers and lasted for approximately 1.5 hours each. Focus group discussions were audiorecorded with participant permission. For the message content, the focus groups discussed current activities for managing NCDs (eg, medication use, doctor’s visits, lab tests and monitoring, PE groups, and lifestyle changes in diet, physical activity, smoking, and alcohol use), as well as facilitators and barriers to these activities. For the message format, questions included the frequency and duration of messages, preference for text or voicemail format, access to cell phones, the best time of day to send, and ability to respond to a text. At each district, 2 PEs (convenience sample) were also interviewed to evaluate features of the tablet for tracking their patients, to provide outreach and assure usability.

The taxonomy of behavioral change communication deemed most appropriate for our intervention can be best described by the Behaviour Change Wheel [22] using the following: sources of behavior of psychological capability, reflective motivation, and automatic motivation; intervention functions of education, persuasion, incentivization, and enablement; and policies driven by communication, marketing, and service provision.

We are contracting with experts at Innovative Support to Emergencies, Diseases and Disasters (InSTEDD) Southeast Asia to develop and implement the mHealth app for sending targeted phone messages to MoPoTsyo patients. The app “Verboice” is being used to access the information available in the MoPoTsyo database to tailor messages according to the specific characteristics of patients’ disease. Verboice is a free and open-source tool developed by InSTEDD that runs apps through voice, allowing users to listen and record messages in their own language. At the time of informed consent, patients provided 2 current cell phone numbers, their own plus a backup belonging to a family member. Following the development of specific content for message themes identified in the focus groups, electronic capture of data provided by the PEs will allow us to match the specific issues and needs related to hypertension and diabetes management and send relevant messages to individual patients.

Process outcomes will follow guidelines of the RE-AIM Model [23] to include Reach (how representative is the sample to the targeted population?), Effectiveness (clinical and intermediate outcomes), Adoption (uptake, utilization, and initial implementation), Implementation (fidelity including adherence and delivery), Maintenance (sustainability), as well as acceptability (satisfaction with the intervention). Key clinical outcomes will include control of blood pressure and glucose, medication adherence, use of medical services (laboratory and physician office visits), and improvement in lifestyle factors such as smoking, body mass index, diet, and exercise. Comparisons will be made by the study intervention group at the end of the trial and by changes in clinical outcomes pre- and postintervention period. Furthermore, comparisons will be made for message effectiveness to determine whether certain message topics led to better clinical outcomes.

Data collected by PEs during the pilot study will be entered into the tablet at their homes and will be transferred electronically to the newly enhanced Patient Information System database at MoPoTsyo headquarters. The collected data will be stored on a relational database management system and will be securely backed up to an offsite data center affiliated with MoPoTsyo. All data, however, will become a part of the patients’ medical record for longitudinal tracking and access to clinical staff. These data will be available for inclusion in individual, as well as combined, data reports that are a part of the proposed information system.

All data will be monitored by a trained information techinology staff funded under the program dedicated to overseeing data. He will check for missingness and consistency using preprogrammed algorithms to identify problems. After cleaning, data to be analyzed for outcomes will be deidentified and formatted into analytic files for use by coinvestigators and analysts for evaluation.