Specific Aims and Hypotheses
The three aims are: 1) to evaluate the effectiveness of implementation of a multicomponent decentralized care model in a public primary care system for hypertension and diabetes compared with usual care and digital health intervention alone; 2) to examine implementation processes, including factors that explain how the interventions influence hypertension and diabetes management, and the barriers/facilitators to deliver and sustain interventions; and 3) to undertake an economic evaluation.
We hypothesize that compared with usual care, multicomponent decentralized primary care will improve all steps of the hypertension and diabetes care continuum. Additionally, we hypothesize that the digital health intervention alone (Simple App) will improve BP and glycemic control compared with usual care but will have limited impact on other steps of the care continuum. We also hypothesize that the multicomponent decentralized care will lead to a higher treatment success rate relative to the digital health intervention alone.
Study Design
This is a type 2 effectiveness-implementation hybrid study with a dual focus on testing intervention effectiveness and implementation strategies.34 A three-arm mixed-methods quasi-experimental design will be used to achieve study objectives. Two intervention subdistricts and one comparison subdistrict from Dinajpur district, Rangpur division in Northern Bangladesh (Fig. 1), will be selected to implement the study. The key outcomes will be evaluated with community-based repeated cross-sectional surveys with independent samples. We will also make use of facility-based prospective cohort data to supplement community-based surveys. A mixed methods approach will be used to generate data on both the primary and secondary outcomes, as well as quantitative and qualitative data for process evaluation. The duration of the study is 36 months, including an intervention period of 24 months.
Study setting and populations
Bangladesh is a lower-middle income country in south Asia. Accompanying a strong economic growth in recent decades, Bangladesh is experiencing rapid demographic and epidemiologic transitions like most other LMICs. In 2019, NCDs accounted for an estimated 70% of total deaths in Bangladesh.35 The prevalence of hypertension and diabetes among Bangladeshi adults were estimated to be 27.4% and 9.8%, respectively.36,37 According to the 2017–2018 Bangladesh Demographic and Health Survey, only 36.7% of hypertensive Bangladeshi adults were aware of their condition, only 31.1% were on treatment, and only 12.7% had blood pressure controlled.37 The awareness, treatment, and control rates among those with diabetes were estimated to be 30.9%, 28.2%, and 7.5%, respectively.36 Tobacco use, insufficient fruit/vegetable intake, and overweight/obesity, are highly prevalent.38
Over the past decade, the Government of Bangladesh has promoted improved NCD care through national multisectoral actions, most notably by the establishment of dedicated NCD care delivery points (“NCD Corners”) at subdistrict hospitals since 2011.39 These NCD Corners were designed to provide preventive and clinical care for common NCDs. Despite the initiative, NCD corners remained nascent due to gaps in specific guidelines, a trained workforce, appropriate medicines and supplies, adequate laboratory facilities, and consistent record-keeping and reporting.12,13 Although originally planned to involve lower-level primary care facilities in the delivery of NCD care, village-level primary care facilities are not equipped to provide NCD care.40 Thus, access to public NCD services remains poor in rural Bangladesh.40
Recently, the NCD control program of the government of Bangladesh launched an initiative to integrate a digital system for the management of hypertension and diabetes. To streamline the digital NCD care initiatives taken by the partner organizations, the NCD control program has encouraged the uptake of Simple App.
In rural Bangladesh, a subdistrict (Upazila) is the second lowest tier of administration, while union is the lowest rural administrative unit. Each union is made up of nine wards. The primary healthcare system in rural areas consists of a subdistrict level hospital (Upazila Health Complexes), union health and family welfare centers, and village level community clinics. A subdistrict hospital lies at the top of the primary care system. It is typically a 50-bed hospital, serving a population of 2–4 hundred thousand.
Community clinics are at the lowest level of the healthcare facility hierarchy. With two rooms and drinking-water and lavatory facilities under a covered waiting area, community clinics were designed to be accessible to more than 80% of the rural population within a walking distance of less than 30 minutes.39 Community clinics are staffed by a Community Health Care Provider (CHCP), who usually receives 3 months of basic medical training.38
The study will be conducted in Dinajpur district, Rangpur Division, in northern Bangladesh. Rangpur is among the poorest divisions in Bangladesh. The multicomponent interventions will be implemented in the Chirirbandar subdistrict, while the digital health only intervention will be carried out in the Parbatipur subdistrict. The Biral subdistrict will serve as a reference subdistrict with no additional intervention. The basic demographic information and government health facility statistics are presented in Table 1.
Table 1
Population and health facility statistics in study areas
Characteristics § | Study Sites |
Chirirbandar (Multicomponent) | Parbatipur (digital health only) | Biral (Usual care) |
Area, sq km | 312.7 | 395.0 | 353.4 |
Population | 292,500 | 365,103 | 257,925 |
% Adults aged 40+ | 25.6 | 25.4 | 25.3 |
% Elderly aged 60+ | 7.3 | 7.2 | 7.3 |
% Muslim | 76.3 | 85.5 | 72.0 |
% urban | 3.0 | 11.0 | 3.5 |
Unions | 12 | 10 | 10 |
Villages | 141 | 229 | 238 |
Health facilities | | | |
Subdistrict hospital | 1 | 1 | 1 |
Community clinic | 35 | 40 | 34 |
Literacy rate | 52.9 | 53.9 | 47.3 |
§ Data source: 2011 Bangladesh Census; Bangladesh Ministry of Health and Family Welfare facility registry |
Interventions
Building upon an assessment of barriers to hypertension and diabetes management at the patient, provider, and health system levels,40–43 and previous interventional studies,29–31 a multicomponent intervention package has been designed to increase access to primary care, and to improve care quality and patient retention. The intervention package includes digital health, decentralization with task sharing, community-based care, and supportive monitoring visits. The effectiveness of the multicomponent interventions will be compared with both digital health only intervention and with usual care. An outline of intervention components of three arms in the three subdistricts is presented in Table 2. The program theory for the multicomponent intervention is described in Table 3. Details of the intervention components are discussed below.
Table 2
Level of intervention | Multicomponent decentralized care (Chirirbandar) | Digital health only (Parbatipur) | Usual care (Biral) |
NCD corner (subdistrict level) | 1. Training and support on Simple app for hypertension/diabetes management 1.1 Conduct centralized and onsite training for orientation to Simple app 1.2 Provide basic equipment and supply to support use of Simple app (e.g., tablet computers, power bank, patient booklets with barcode) 1.3 Continuous IT support to solve any technical issues 2. Training NCD corner clinicians on clinical guidelines and establish team-based care 2.1 Provide training on national hypertension/diabetes guidelines and team-based care; refresher training every 6 months 2.2 Provide necessary equipment and supply (e.g., digital BP machine, glucometer & consumables, weight/height scale) 2.3 Train district public health manager to monitor NCD corner performance and facilitate supportive supervision 2.4 Organize quarterly meeting to facilitate dialogue between healthcare providers at different level to improve collaboration and communication | 1. Training and support on Simple app for hypertension/diabetes patient management 1.1 Conduct centralized & onsite training for orientation to Simple app 1.2 Provide basic equipment and supply to support use of Simple app (e.g., tablet computers, power bank, patient booklets with barcode) 1.3 Continuous IT support to solve any technical issues 2. Training NCD corner clinicians on clinical guidelines and establish team-based care 2.1 Provide training on national HTN/DM guidelines and team-based care; refresher training every 6 months 2.2 Provide necessary equipment and supply (e.g., digital BP machine, glucometer & consumables, weight/height scale) 2.3 Train district public health manager to monitor NCD corner performance and facilitate supportive supervision | N/A |
Community clinics (union level) | 3. CHCPs conduct BP/BG screening, monitoring, lifestyle counseling, dispensing of medications, follow-up overdue patients at CCs 3.1 Provide training on hypertension/diabetes clinical guidelines; refresher training every 6 months 3.2 Provide equipment and supplies, including job aid and IEC materials 3.3 Facilitate supportive supervision by experienced NCD corner staff & subdistrict health officials (HI), monitoring CHCPs’ performance on designated tasks and communication with CHWs and NCD corner clinicians for patient follow-up and dose titration, respectively 4. CHCP care coordination supported by Simple app 4.1 Conduct centralized & onsite training for orientation to Simple app 4.2 Provide basic equipment and supply to support implementation of Simple app (e.g., tablet computers, power bank, patient booklets with barcode) | N/A | N/A |
CHW (village level) | 5. CHWs assist in screening, counseling, referral, and follow-up. 5.1 Provide training on knowledge and skills pertinent to the tasks; refresher training will every 6 months 5.2 Provide job aid and IEC materials 5.3 Support CHWs in dissemination NCD knowledge, encourage universal screening; home visit to existing patients providing counseling; follow up with overdue patients to improve retention 5.4 Facilitate supportive supervision by experienced NCD corner staff & subdistrict health officials (HI), monitoring CHWs’ performance on designated tasks | N/A | N/A |
Table 3
Programme theory of multicomponent decentralized NCD care
Interventions/activities § → | Output → | Outcome → | Impact |
NCD corner (subdistrict level) 1. Training and support on Simple app for hypertension/diabetes management 2. Training NCD corner clinicians on clinical guidelines and establish team-based care Community clinics (union level) 3. CHCPs conduct BP/BG screening, monitoring, lifestyle counseling, dispensing of medications, follow-up overdue patients at CCs 4. CHCP care coordination supported by Simple app Community health worker (village level) 5. CHWs assist in screening, counseling, referral, and follow-up | NCD corner (subdistrict level) 1. New and existing hypertension/diabetes patients registered in Simple 2. Clinicians use Simple app to monitor treatment progress and tracking defaulted patients 3. SMS reminders are sent to patients for follow-up visit 4. Patients are referred to CC for drug refill and BP/BG monitoring 5. Clinicians monitor drug stock supported by Simple app 6. Essential equipment is calibrated regularly and in adequate working condition 7. Essential drugs supply sufficient for 2 + months 8. Supportive supervision by higher level health officials informed by Simple app facility performance metrices Community clinics (union level) 9. CHCPs using Simple app to monitor patient Tx progress, follow-up overdue patients, and perform other designated tasks 10. Hypertension/diabetes patients take BP/BG measure, receive drug refill and counseling for adherence and lifestyle modification at CCs 11. Patients with complications are referred to NCD corner or higher-level health facilities 12. Equipment properly maintained and drug supply sufficient for 2 + months 13. Timely audit and feedback by NCD corner clinicians to CHCP, ensuring quality of care 14. CHCPs conduct screening for hypertension and diabetes Community health worker (village level) 15. CHWs conduct home visit or other community-based activities, motivating people at risk to get screening and address risk factors 16. CHWs conduct home visit to follow up overdue patients 17. IEC materials distributed; health promotion posters prominently displayed | 1. NCD care accessibility and quality improved 2. Clinicians follow protocol, and intensify treatment timely when needed 3. Clinician consultation time increased 4. Waiting time at NCD corner reduced 5. Financial burden associated with travel, forgone productive time, and medications reduced for patients 6. Patient retention rate improved 7. Patient medication adherence improved 8. Increased proportion of adults at risk screened for hypertension/diabetes 9. Increased proportion of adults with hypertension/diabetes are enrolled in care 10. Improved control rate among treated 11. Health promotion behaviors are more prevalent | 1. CVD incidence, and premature CVD mortality reduced 2. Socioeconomic disparities reduced in care accessibility, and treatment outcome 3. Community NCD awareness improved |
§ Intervention and activity details are available in Table 2. |
Multicomponent decentralized care
Digital health for patient management. In line with the Bangladesh government’s recommendation, Simple App will be used by primary care facilities for coordinated NCD management. Using unique patient IDs, the app allows seamless transfer of electronic health records at primary care facilities and thus ensures continuity of the process. Other functions include 1) personalized short message service/WhatsApp reminders automatically sent to patients, reminding them to return for visits, 2) flagging patients who are overdue or miss a follow-up visit and prompting actions to contact patients, 3) progress reports showing facility performance on key performance indicators including the number of patients enrolled, and proportion of hypertensive patients have control status among others. The accompanying patient component, the BP passport, will not be included in the intervention package due to limited penetration of smartphones in the study area.
Decentralization with task sharing. Decentralization will include shifting screening, routine follow-up, and dispensing of medication refills from subdistrict NCD clinics to community clinics, the most peripheral primary care facilities. To implement restructured facility-based care, we will 1) design and implement a training programme customized to the anticipated activities of different roles of health care providers across levels of primary care facilities; 2) equip community clinics to carry out screening, routine follow-up, and dispensing of medications; and 3) facilitate regular supportive supervision of primary care facilities by health officials and managers. The training programme is designed based on existing international guidelines and training manuals, and the national guidelines of the Government of Bangladesh. CHCPs at community clinics will also be trained to carry out their new roles in hypertension and diabetes screening, counseling, and care coordination. Refresher training will be done every 6 months throughout the intervention period. In addition to training, community clinics will be equipped with basic medical devices for routine patient monitoring, and Android tablets for keeping patient health records. The list of devices that will be supplied is shown in Table 4. Medications for hypertension and diabetes will be available in study community clinics.
Table 4
List of equipment and supplies to be provided
Measurement/device | Type | Number | |
1. NCD Corner (one set per NCD corner) | |
| Tablet | 7" Fusion5 Android Tablet or equivalent | 2 | |
| Power bank | Imuto 20000mAh Compactable Charger or equivalent | 2 | |
| Patient booklet | Patient booklet with QR code and NCD self-management guideline | 2000 | |
| Weight | Tanita UM70 or equivalent | 1 | |
| Height | Infant Child Adult ShorrBoard | 1 | |
| Blood pressure | Omron M3 IT or equivalent | 1 | |
| Blood pressure | Paramed Aneroid mercury sphygmomanometer or equivalent (for calibration) | 1 | |
| Blood glucose | Accu-Chek Active or equivalent | 1 | |
| Glucometer consumables | Strips, lancet, battery, calibration solution | As needed | |
| Waist circumference | Tape measure | 10 | |
| Pamphlets | IEC materials patients can take home freely | 2 | |
| Posters | How to measure BP and blood glucose; hypertension protocol; diabetes protocol; lifestyle advice | 2 | |
| Flipcharts | To conduct health education | 2 | |
2. Community clinics in multicomponent subdistrict (one set per CC) | |
| Tablet | 7" Fusion5 Android Tablet or equivalent | 1 | |
| Power bank | Imuto 20000mAh Compactable Charger or equivalent | 1 | |
| Patient booklet | Patient booklet with QR code and NCD self-management guideline | 500 | |
| Weight | Tanita UM70 or equivalent | 1 | |
| Height | Infant Child Adult ShorrBoard | 1 | |
| Blood pressure | Omron M3 IT or equivalent | 1 | |
| Blood glucose | Accu-Chek Active or equivalent | 1 | |
| Glucometer consumables | Strips, lancet, battery, calibration solution | As needed | |
| Waist circumference | Tape measure | 10 | |
| Pamphlets | IEC materials patients can take home freely | 1000 | |
| Flipcharts | To conduct health education | 2 | |
| Posters | How to measure BP and blood glucose; hypertension protocol; diabetes protocol; lifestyle advice | 2 | |
3. Community health workers in multicomponent subdistrict (one set per CHW) | |
| Pamphlets | IEC materials patients can take home freely | 1000 | |
| Flipcharts | To conduct health education | 2 | |
| Posters | For raising NCD awareness in community, lifestyle advice | 2 | |
NCD, noncommunicable disease; CC, community clinic; CHW, community health worker; IEC, information education and communication; HTN, hypertension; DM, diabetes |
Engaging community health workers in NCD care. Integrating hypertension and diabetes care into the community by training community health workers (CHWs) to provide screening, counseling, and follow-up has been shown to improve hypertension and diabetes care management in many LMICs, including countries from South Asia.29–31,44−47 In line with previous studies, we will enlist CHWs in case finding, counseling, referral, and follow-up. Existing CHWs will be trained and provided with necessary equipment and health education materials and tools to carry out these tasks. The initial training will cover topics including the basics of hypertension and diabetes, lifestyle risk factor counseling, patient self-management, and collaboration with CHCPs. Refresher training will be conducted every 6 months.
Supportive monitoring and supervision. Quarterly supervision by higher level health administrators and medical professionals will be organized at subdistrict NCD corner to help solve issues with patient management, medication supplies, and other related issues. Visits may be informed by performance summaries made available by the Simple App dashboard. Similarly, quarterly supervision visits to CCs will be done by subdistrict NCD corner medical staff and subdistrict health officials; quarterly supervision of CHWs will be conducted by CHCPs. Supervision in the first quarter immediately after initiation of the intervention will be monthly. Supportive visits by the study team and Simple app technicians will be organized to help NCD corners and CCs solve technical issues with Simple App. Continuous technological support will be made available throughout the intervention period.
Patient pathway within the multicomponent decentralized care (Fig. 2). CHWs will be tasked with raising awareness of NCD in communities and encouraging all adults aged 40 and above, following national protocols, to be screened for hypertension and diabetes at community clinics. Adults aged 40 and above without known hypertension or diabetes will be referred to subdistrict level NCD Corner for confirmation and care plan initiation (if confirmed). At the subdistrict level NCD clinic, nurses will perform additional examinations and update patient data. Doctors at subdistrict NCD Corner will decide whether a patient can be referred back to community clinics for follow-up and routine management based on specific medical conditions. Participant dashboards will be updated continuously, enabling healthcare providers at community clinics to follow up with routine examination, medication refill, and counseling. At the follow-up visit, CHCPs at community clinics will conduct basic physical examinations, ask for new symptoms, assess CVD risk, and counsel on medication adherence and healthy lifestyle. Patients will be back referred to the subdistrict NCD Corner for medication titration and further evaluation if their treatment target is not met or any urgent symptoms or conditions are recorded; otherwise, patients will receive drug refills at CCs, and the next follow-up visit will be scheduled.
Digital health only
In the subdistrict with the digital health intervention only, training and support to use the Simple App for hypertension and diabetes management will be provided to the subdistrict NCD corner in the subdistrict hospital. To be consistent with standard practice in Bangladesh, training on the national protocols for hypertension and diabetes management will also be provided, as well as training of health officials on conducting supportive supervision. Community clinics and CHWs will have essentially no involvement in NCD care provision. Patient pathways will remain the same as usual care.
Usual care
Existing usual care will be provided by the government primary care system, including screening, treatment initiation, drug refill, and routine follow-up, at the subdistrict NCD corner. Community clinics and CHWs will have essentially no involvement in NCD care provision.
Outcomes
The primary outcome will be the proportion of patients treated for NCDs who achieve or maintain disease-specific control status, according to national or the World Health Organization (WHO) Package of essential noncommunicable disease interventions protocols. Hypertension will be considered controlled if systolic blood pressure < 140 mmHg and diastolic blood pressure < 90. The goal for glycemic control is fasting plasma glucose 4.4–7.2 mmol/L or random plasma glucose < 11.1 mmol/L. Effectiveness will be assessed using data from the repeated independent community-based surveys.
Secondary outcomes include hypertension and diabetes care cascade (i.e. % of patients aware of their condition, % of diagnosed patients on treatment), smoking cessation among hypertension and diabetes patients, and meeting recommended weekly physical activity level. Operational definitions, data sources, instruments, and timelines for these outcome variables are listed in Table 5.
Table 5
Data collection plan for evaluation indicators
RE-AIM Domains/ Indicators | Operational definition | Data source | Instrument | Timeline, at month |
Effectiveness | | | | |
Primary outcomes | | | | |
% of HTN patients receiving treatment achieved BP control | BP < 140/90 mmHg | Community surveys | Omron HEM 7120 | 0, 12, 24 |
% of DM patients receiving treatment achieved glycemic control | Fasting plasma glucose < 7.0 mmol/L or random plasma glucose < 11.1 mmol/L | Community surveys | Accu-Chek Active | 0, 12, 24 |
Secondary outcomes | | | | |
% adults with HTN aware of their condition | Ever been told by a healthcare provider that they have raised BP or hypertension | Community surveys | Questionnaire | 0, 12, 24 |
% adults with DM aware of their condition | Ever been told by a healthcare provider that they have diabetes | Community surveys | Questionnaire | 0, 12, 24 |
% adults with diagnosed HTN on treatment | Currently taking medication for HTN, not including herbal/traditional remedy | Community surveys | Questionnaire | 0, 12, 24 |
% adults with diagnosed DM on treatment | Taking medication for DM, not including herbal/traditional remedy | Community surveys | Questionnaire | 0, 12, 24 |
% of HTN/DM patients quitted smoking | Quitted smoking is defined as never smoked, even one or two puffs, during past 6 months | Community surveys | Tobacco use Questionnaire | 0, 12, 24 |
% of HTN/DM patients met recommended PA level | ≥ 150 mins MVPA per week | Community surveys | WHO STEPS | 0, 12, 24 |
% of HTN/DM patients achieve 80% adherence | Self-reported medication adherence | Community surveys | VAS | 0, 12, 24 |
Reach | | | | |
% adults with HTN screened for HTN | Had a BP measurement in past 12 months | Community surveys | Questionnaire | 0, 12, 24 |
% adults with DM screened for DM | Had a BG measurement in past 12 months | Community surveys | Questionnaire | 0, 12, 24 |
% HTN/DM patient receiving treatment from public PHC facilites | Currently receiving care from government PHC facilities | Community surveys | Questionnaire | 0, 12, 24 |
Adoption | | | | |
% HTN/DM patients registered in Simple | Information recorded in Simple, separated for CC & NCD corner | Simple app | Simple app | Quarterly |
% CCs conducted NCD screening | Screening 50 + adults/month | Simple app | Simple app | Quarterly |
% CCs perform routine care for NCD patients | Following up 25 + patients/month | Simple app | Simple app | Quarterly |
% CHWs conducted home visit | Home visit 5 + patients /month for lifestyle counseling and adherence support | CHW report | CHW report | Quarterly |
Implementation | | | | |
% of providers participated training and refresher | Participated in training on NCD management | Training report | Training registry | Biannually |
% of CCs with functional essential equipment and supply | BP/BG machines are calibrated, strips and IEC materials available | Facility records | Facility report | Quarterly |
% of planned supportive supervision conducted | 8 planned supervisions, separated by levels of PHC and CHW | Supervision report | Checklist | Quarterly |
Maintenance | | | | |
% of newly enrolled patients retained in care | LTFU: ≥ 3 months late for the last scheduled visit, 6/12 months | Patient registry | Patient registry | 12, 24, 30 |
CCs perform routine care for NCD patients | Following up 25 + patients/month | Patient registry | Patient registry | 30 |
NCD corner using simple for patient management | 80% + patient encounters recorded in Simple app | Patient registry | Patient registry | 30 |
HTN, hypertension; DM, diabetes mellitus; BP, blood pressure; BG, blood glucose; STEPS, STEPwise approach to NCD risk factor surveillance; IDI, in-depth Interview; FGD, focus group discussion; LTFU, lost to follow-up; PHC, primary health care. |
Safety-related measures will include serious adverse events, including death, hospitalization, and other conditions that result in persistent or significant disability. Medical judgment will be exercised in deciding whether an adverse event is serious. Important adverse events that are not immediately life-threatening or do not result in death or hospitalization but may jeopardize the subject or may require intervention to prevent one of the other outcomes listed above, may also be considered serious.
Specific Aim 2 will be evaluated with a RE-AIM framework utilizing patient assessments and stakeholder interviews, informed by previous studies.23,48 Implementation fidelity and process evaluation will be guided by the UK Medical Research Council guideline on process evaluation of complex intervention,49 and the WHO’s Noncommunicable Disease Facility-Based Monitoring Guidance.50 Data sources will include training reports, prescription practice captured by Simple app, facility records, and patient registry. Details on the measurement, data source, instrument, and timeline are provided in Table 5. In addition, qualitative data will be collected from patients, healthcare providers, CHWs, and public health managers to inform the evaluation of barriers and enablers of implementing the interventions in primary care system for improved NCD outcomes.
For Specific Aim 3, primary cost-effectiveness measures will be incremental cost per 1 percentage point increase in the proportion of participants achieved control status over the 24-month intervention period and incremental cost per projected disability-adjusted life years (DALY) averted.
Sampling strategy and power calculation
Three community-based surveys will be conducted at baseline, 12 months, and 24 months (at the end of the study). For each evaluation survey, an independent random sample of adults aged 40 and above who are community residents in three study subdistricts (5 unions each) will be randomly selected through a multistage cluster sampling approach (Fig. 3). A buffer zone between Chirirbandar and Parbatipur was created to mitigate potential spillover effects. Given the available resources and time limitation, we will select a total of 15 wards from each subdistrict. The wards will be divided into segments of ~ 250 households without disrupting the boundary of villages/paras. One segment per ward will then be randomly selected. Subsequently, a list of adults aged 40 years of age within each village/para of the selected segment will be obtained. From each segment (cluster), an equal number of women and men ≥ 40 years of age will be randomly selected by using systematic random sampling with the condition that not more than one adult male and one adult female is included from the same household.
The planned sample size is 6750 subjects for each community survey, with an equal number of clusters per subdistrict (15 clusters per subdistrict, or 45 clusters total) and similar cluster sizes (150 subjects,75 females and 75 males per cluster). Assuming a prevalence of diagnosed hypertension of 40%, and 28% for diabetes among adults aged 40 and above based on South Asia Biobank51 data (unpublished findings), we expect to collect information from 2700 individuals with hypertension, and 1890 individuals with diabetes. Assuming a conservative intraclass correlation coefficient of 0.02,23,52 and a two-sided type I error rate of 5%, the survey will have 90% power to detect a difference as small as 10 percentage points between the intervention and reference groups for blood pressure control, and an 11 percentage point difference for glycemic control. The power calculation was done using Stata 17 (Stata Inc.) power analysis for a clustered two-sample proportions test.53 The planned sample size for binary outcomes will ensure sufficient power for continuous measurements.23,52,54
All health care providers in primary care facilities who are directly involved in NCD care will be approached for qualitative data. This includes approximately 20 doctors and nurses from three subdistrict NCD corners, 15 CHCPs from community clinics and 10 CHWs. Moreover, 6 health officials (2 from each subdistrict) will be approached for an in-depth interview (IDI). Thus, the target sample for qualitative data collection include 51 healthcare providers or managers. In addition, 6 focus group discussions (FGD) with patients living with hypertension and/or diabetes (3 male FGDs and 3 female FGDs) will be conducted in each subdistrict. We will include 8–12 patients for each FGD, resulting in approximately 180 patient participants in total. A purposeful selection of the participants will be done to maximize the diversity of the sample on sociodemographic characteristics (e.g., age, sex, religion, socioeconomic status, geographic distance to health facilities), and NCD history (new vs experienced).
Data Collection
Community surveys will include interviewer-administered questionnaires and physical examinations. Questionnaires will include domains of sociodemographic characteristics, comorbidity, medication use, and health behavior, and adherence to scheduled visits and prescribed medications. Questionnaires will be translated into the local language (Bengali) and piloted in the field to ensure the cultural appropriateness and clarity of the questions. Then, we will digitalize the questionnaire using Kobo Toolbox, a widely used open online data system for data collection and management. Physical examinations will be conducted to diagnose hypertension and diabetes, according to national protocols. Physical examinations will include taking anthropometric measures, assessing blood pressure, and testing blood glucose to identify patients with hypertension and diabetes.
Facility-based data collection will be conducted via Simple app by onsite research staff. Basic demographic and NCD history (age, sex, heart attack, stroke, chronic kidney disease, diabetes mellitus), medication for hypertension, registration date, BP/BG at each facility visit, and treatment dosage at each visit will be recorded by Simple app. In the usual care subdistrict, the same set of data will be collected by onsite research staff through reviewing the facility patient registry. In addition, we will collection facility-level data on availability of essential medications, functionality of blood pressure devices, glucometer, and treatment success rates for process evaluation.
A questionnaire on the feasibility, acceptability, and utility of the Simple app for hypertension and diabetes management will be administered to healthcare providers in both intervention subdistricts at months 3 and 12 after the implementation of the project. All healthcare providers engaged with hypertension and/or diabetes care at NCD corners and community clinics will be approached to participate in this survey.
Qualitative data will be collected through observations, IDI, and FGD to understand patients’ experiences with different models of service delivery, healthcare providers’ perspectives on restructured health services, and the digital tool, among other aspects of the interventions. The mixed methods approach will help explain why the interventions may or may not have been effective and particularly component(s) that may have been most important.
Patient FGDs will be conducted in all three subdistricts (6 FGDs per subdistrict). Data will help inform barriers and facilitators in accessing care and perception of care quality. Additional data, where pertinent, on experience with intervention components (e.g., CC drug refill, CHW home visit, CHCP counseling and follow-up) will be collected.
All participating health administrators, clinicians, and CHWs in the intervention subdistricts will provide qualitative information on 1) barriers and facilitators to delivering NCD care; experiences with Simple app, 3) views of study participation and impact of the intervention; and 4) opinions on restructured workflow, including teamwork with community clinics and CHWs (for multicomponent decentralized care only). Health officials directly involved in the project will be approached for IDI, for barriers/facilitators to implement and sustain the interventions.
FGD/IDI guides will be developed in advance, guided by WHO’s seven-domain framework of healthcare delivery55, chronic care model,8 and prespecified program theory. Interviews will be audio-recorded and are expected to each take 30–45 minutes to complete. Qualitative data collection will be performed at baseline, month 12, and month 24.
Treatment-related costs for patients will be collected for the economic evaluation. These will include direct and indirect patient costs to access care (e.g., transportation, food, childcare, lost worktime). We will collect health service delivery costs (e.g., staff, transportation, laboratory and clinical, training and workshop, utilities, overhead and supervision). We will apply micro-costing to document the time put into implementing different activities prospectively.56 The cost of the proposed interventions will take into account the setting, target population, human resources, building space and equipment, consumables such as medical supplies and medications, and overhead costs such as electricity, water, and maintenance.
Statistical analysis plan
Baseline characteristics of participants, including sociodemographic characteristics, medical history, anthropometrics, and lifestyle factors will be compared between the two intervention and one control subdistricts using one-way ANOVA or Rao-Scott ꭓ2 tests. Analyses will be stratified by hypertension and diabetes conditions. A difference-in-difference (DiD) estimate for hypertension and diabetes will be implemented with multivariate logistic regression analyses, which take the following form.
Log(p/1-p) = β0 + β1*Time + β2*Interventions + β3*[Time*Intervention] + β4*Covariates + ε
The β3 coefficient captures the difference in change over time. Clustered standard errors at the ward segment level will be specified to allow for intragroup correlation, relaxing the usual requirement that the observations within a cluster are independent.57 Given that the evaluation follows a multistage process whereby ward segment clusters are randomly sampled from each subdistrict, and individuals are sampled randomly from the ward segment clusters, a clustering adjustment for standard error is necessary to avoid inflating the precision of the estimated intervention effect.58,59 Marginal effects will be calculated to illustrate how the predicted probability (i.e., the proportion of patients with controlled conditions) changes over time among the three groups. Potential heterogeneity of effects will be explored in subgroup analyses, defined by age, sex, religion, and socioeconomic status.
Several sensitivity analyses will be performed to check the robustness of the findings. First, the validity of our difference-in-differences approach predicates on the “parallel trends” assumption - the differences in outcomes between groups move in tandem over time in the absence of intervention. Ideally, the “parallel” assumption can be tested with repeated measures before intervention; however, this is infeasible given time limitations and our budget restriction. Instead, we will assess the assumption with a negative control in treatment and/or outcome.60 In addition, we will assess and evaluate the impact of other programs on NCD care running in the study subdistricts (and nearby subdistricts) during the implementation period. Second, potential ecological fallacies will be assessed with hierarchical linear and nonlinear models.61,62 The nested nature of the data structure and over time will be modeled as random effects. Third, while missing data will be minimized by careful training of the survey team, computer-assisted personal interviews, and regular data quality control during the implementation of the study, it is likely unavoidable. Missing data patterns and potential mechanisms (i.e., whether missing at random can be assumed) will be assessed. Sample characteristics between participants with and without missing data will be compared to evaluate potential biases. If missing data only affect a small proportion (i.e. <5%) of the sample, listwise deletion will be performed. Otherwise, missing data will be handled with multiple imputation with a chained equation if the missing-at-random assumption is reasonable. The results will be compared with and without multiple imputation and interpreted in light of the potential limitations due to missing data.
The facility-based cohort data collected longitudinally will capture patients’ treatment trajectory over the study periods. Changes in the proportion of patients with a controlled condition will be analyzed with generalized estimating equation Poisson regression with robust variance.63,64 Continuous changes in systolic blood pressure and diastolic blood pressure from baseline will be assessed with linear mixed effects models.
For qualitative data, dual Bengali-English language speakers on the study team will transcribe and translate the audio-recorded in-depth interviews and FGDs into English. The English language version of the transcripts will then be coded and analyzed thematically using NVivo, a qualitative data software program developed by Lumivero. Qualitative data will allow the identification of specific barriers to and facilitators of interventions by exploring the experiences of patients and clinicians engaged in different models of NCD care delivery. Qualitative data analysis will be guided by prespecified program theory, the RE-AIM framework, and standard ground theory to identify themes, build and apply codebooks, and describe thematic characteristics, patterns, and relationships.65,66
For the economic evaluation, we will compare the costs and effects of intervention with usual care group both from financial and economic perspectives.67 Financial cost per unit outcome will be calculated by dividing the total cost by the quantifiable unit of outcome (screening, treatment, retention in care, control for each of three group. Incremental cost-effectiveness ratios (ICERs) will be computed to compare additional costs and effects of each intervention with the usual care. A nonparametric bootstrap with 1000 replications will be used to estimate 95% CIs around point estimate of ICERs. In addition to primary and secondary outcomes, Disability-adjusted life-year (DALY) metric will be used to facilitate comparison with other programs for competing priorities. All local currency (Bangladesh Taka) costs will be converted to US dollars using the prevailing exchange rates, and will be adjusted for inflation rates, discounted at 3%, and expressed in the 2025 US dollar present value terms.
Sensitivity analysis will explore the robustness of the results to alternative probability distributions, time horizon, and uncertainty of key variables. In addition, we will conduct univariate sensitivity analyses to explore the impact on the costs and cost effectiveness of the interventions of varying key input parameters, including different discount rates of costs and values of effectiveness in blood pressure and glycaemic control. A 1% annual reduction of declining persistence of effect will be evaluated.68