Scale-up of a primary care intervention for cardiovascular risk management in Malang, Indonesia

SMART (Systematic Medical Appraisal, Referral and Treatment) health is a technology-enabled ecosystem that aims to improve the delivery of consistent high-quality essential primary healthcare to communities. SMARThealth cardiovascular disease (CVD) supports the provision of preventive CVD care at the community and household level by strengthening existing health systems. In Indonesia, the SMARThealth CVD program is a complex intervention with multiple components, each of which includes a component of digital support using mobile devices: • Raising community awareness through strengthening existing programs: Through a government-funded chronic disease management program at village and neighbourhood-level primacy health care centres, community health workers (kaders) and nurses raise community awareness of CVD and its risk factors. Activities occur on an approximately monthly basis for 24 months and include role play, traditional theatre and doctor-led group education sessions of 20 minutes duration. A guided 1 hour physical activity demonstration, led by Kaders and nurses, occurs once a week for 24 months in all villages. These activities will be strengthened by the training of health care providers to improve knowledge about CVD and associated risk factors (see ‘Training and performance management’ below), and by individualised patient counselling provided by kaders during monthly household visits using a risk communication tool with pre-recorded animations on the SMARThealth application (see below). • Training and performance management of health care providers: Kaders participate in one intensive 5-day face-to-face training programme led by study researchers together with the Malang District Health Authority’s (DHA) SMARThealth implementation team, with subsequent ongoing remote or in-person support from district-level field supervisors. The training session consists of modules to improve knowledge about CVD and associated risk factors, as well as the technical use of the SMARThealth platform (mobile tablet, SMARThealth application and basic medical equipment) for the identification, referral and follow-up of patients at high predicted CVD risk. In one 3-day face-to-face workshop (again led by study researchers together with Malang DHA SMARThealth implementation team), primary care doctors and nurses are provided guidance in the use of the electronic data transmitted by the kader, interpretation of the decision support output from the SMARThealth application for disease and risk management, and use of audit and feedback capabilities. Regular monthly meetings of kaders at the village level are used for problem resolution dring the 24 months intervention period. Kaders, nurses and doctors all receive two automated pre-recorded voice messages by mobile phone each month for 24 months reinforcing SMARThealth procedures. • CVD risk assessment with clinical decision support: As part of their routine duties, kaders perform household visits and invite all household members aged 40 years and above to participate. Those who agree undergo CVD risk assessment through a clinical decision support system on a 7-inch Android tablet device using an Android 4.1 operating system. This application prompts the kader to collect basic sociodemographic information, as well as a relevant personal and family health history including medication use. The kaders also use standardized equipment to measure height and weight, and an automated sphygmomanometer to record blood pressure (BP; Omron HEM7130). Three BP measurements are recorded, with the average of the last two considered by the clinical decision support system. Random capillary blood glucose levels are also measured using a Freestyle Optium Neo blood glucose monitoring system, with a value of greater than or equal to 200 mg/dL (11.1 mmol/L) considered by the clinical decision support system to be consistent with diabetes in those without a prior diagnosis. The clinical decision support system then identifies individuals considered at high predicted CVD risk, defined by the presence of any of the following: (1) a past history of CVD confirmed by a doctor diagnosis; or (2) an extreme BP elevation (SBP greater than 160 mmHg or DBP greater than 100 mmHg); or (3) a 10-year predicted CVD risk greater than or equal to 30%; or (4) a 10-year predicted CVD risk of 20-29% and a SBP greater than 140 mmHg. In the absence of Indonesian guidelines, the 10-year risk of fatal or major non-fatal major CVD event (myocardial infarction or stroke) is automatically estimated using algorithms based on the World Health Organization/International Society of Hypertension “low information” risk charts tailored to the South-East Asian Region-B, which recommends screening individuals aged 40 years and above and uses age, sex, blood pressure, smoking and diabetes status. Based on the clinical decision support system output, kaders are prompted to provide individualised lifestyle advice and refer all high-risk individuals to nurses or doctors at the primary health care centre for consideration of preventive medication prescription. The clinical decision support system for doctors and nurses is similar to that provided to the kaders, but also provides recommendations for medication use. Unless contraindicated, prescription of a BP lowering drug, a statin and aspirin is recommended for patients with a past history of doctor-diagnosed CVD, while a combination of a BP lowering drug and a statin is recommended for all other high-risk individuals. High-risk individuals are automatically referred back to the kader for follow-up in the community to support lifestyle and medication adherence. The decision support algorithm prioritizes individuals for follow-up depending on the estimated absolute risk (monthly for patients with CVD and/or estimated absolute risk greater than or equal to 30%; every year for those with estimated risk 20-30%). The prioritization algorithm also considers other factors, including whether or not the high-risk individual has seen a doctor following referral; has been prescribed medications; has achieved target BP; or is a current smoker. Priority listings for follow-up are provided to kaders. At each encounter with a kader, nurse or doctor, additional decision support is generated for those not achieving target BP or with high random blood glucose levels. Routine cholesterol screening is not available in this context. On average, an initial screening requires approximately 30 minutes, with 10 minutes for kader follow-up visits. All data collected by kaders, nurses and doctors through the SMARThealth application are uploaded into a shared electronic medical record (OpenMRS) via the Sana Mobile Dispatch Server and stored on a central server. This allows doctors and nurses to view data acquired by kaders and for kaders to view the treatment recommendations made by doctors and nurses. All aspects of this intervention component occur for the duration of the 24 month intervention period. • Patient engagement: High risk patients who have been prescribed medications are provided with a medication calendar each month for the 24 month intervention period to record and track daily medication use over a one-month period. During their monthly follow-up visit, kaders will assess medication adherence through observation of medication packets and review and replace the medication calendar. In addition to the monthly follow-up visits by the kaders, automated pre-recorded personalized voice messages are sent to the mobile phone of high-risk patients to promote lifestyle changes, medication adherence and medical follow-up. Two messages are sent every week, with one conveying the advantages of healthy lifestyle changes and the other targeting patient-specific issues such as reminders for a doctor visit, or for adherence to a specific medication.