An Innovative Mobile Technology Intervention for Maternal and Child Health Care in Cambodia

Innovative Mobile Technology for Maternal and Child Health Care in Cambodia (i-MoMCARE): A Cluster Randomized Controlled Trial

The Innovative Mobile Technology for Maternal and Child Health Care (i-MoMCARE) project is a digital health intervention aiming to increase the coverage of and access to maternal and child health (MCH) services for pregnant women and their babies in rural communities in Cambodia. We will fully adopt an innovative model (ImTeCHO) developed, successfully implemented, and evaluated by Society for Education, Welfare and Action-Rural (SEWA Rural) in Gujarat, India. This study will take advantage of Cambodia's well-structured healthcare system, increased internet coverage, and the sharp rise in smartphone adoption to implement innovative mobile technology intervention.

i-MoMCARE will provide village health support groups (VHSGs) access and training in mobile-based job aid (mobile apps) to facilitate their work with pregnant women and mothers in the villages. Using mobile apps, VHSGs can register pregnant women and mothers, (re)schedule health care appointments, develop a digital record of the medical history of pregnant women and mothers, show educational videos to pregnant women and mothers, and alert health center staff on high-risk cases. i-MoMCARE will also offer health center staff access to the web interface where they can do longitudinal tracking of VHSGs' log-in rate, track the medical history of patients, and screen for complications and referrals.

In Cambodia, long distances to health facilities and inadequate education for women of reproductive age are among the determinants of poor health outcomes, predominantly in rural communities, where the demand for MCH services is high. VHSGs could help overcome the long distance to health facilities with mobile technology. i-MoMCARE could also improve women's understanding of the importance of regular antenatal and postnatal utilization and vaccination and, by extension, reduce high-risk cases and maternal and child deaths. This study is the first to be implemented in Cambodia, contributing to the use of digital health in MCH interventions, which remain in their infancy.

Goal and objectives

To test, implement, and evaluate the applicability and efficacy of the i-MoMCARE to support VHSGs and health center staff in rural communities to increase the coverage of and access to MCH services in Cambodia. Specific objectives include:

1. To capture changes in MCH coverage, maternal mortality ratio (MMR), the infant mortality rate (IMR), and other key indicators that are of national program's priorities.
2. To identify gaps in MCH care through conducting a literature review, consulting with local partners, and qualitative need assessments to understand the barriers to MCH services provided or facilitated by VHSGs.
3. To identify technology components that provide solutions to fill the gaps in MCH care.
4. To pilot and qualitatively evaluate the operational delivery of the intervention.
5. To deliver the intervention among VHSGs and health center staff.
6. To evaluate the acceptability, feasibility, and usefulness of the components of the i-MoMCARE application.
7. To conduct a cost-effectiveness analysis to adapt and scale up the intervention in Cambodia.

Methods This study is a two-arm stratified multi-stage cluster randomized controlled trial (RCT). The digital health technology will be deployed in five provinces. The five provinces have been selected based on the composite index of the MCH indicators and socio-economic status indicators and consultation with the National Maternal and Child Health Center (NMCHC) of Cambodia's Ministry of Health. A literature review and gap analysis using in-depth interviews (IDIs), key informant interviews (KIIs), and focus group discussions (FGDs) will be conducted to acquire information necessary for adapting to the digital health technology, successfully deployed in Gujarat in India to the Cambodian setting.

The randomization will be done at the operational district (OD) level (cluster), the third tier of the health system in Cambodia. To address the issue of contamination that may arise and maintain comparability between the two arms at baseline, we will perform a matched-pair selection of 10 provinces (five for each arm) out of the 25 provinces in Cambodia based on similar levels of MCH-SES index and recommendation from NMCHC. We will use the principal component analysis approach to construct the provinces' MCH-SES index from a basket of indicators of ANC, delivery, and PNC from the Cambodia Demographic and Health Survey (CDHS 2021) and indicators of socioeconomic status from the national population census in 2019 (National Institute of Statistics, 2020a) and Cambodia Socio-economic Survey (CSES 2019. The indicators from CDHS 2021 include the proportion of at least four antenatal care (ANC) visits among mothers, the proportion of women delivered in a health facility, the proportion of women having a postnatal visit in the first two days after delivery, the proportion of fully vaccinated (basic antigen) children aged 12-23 months, the proportion of fully vaccinated (according to the national schedule) among children aged 12-23 months, and proportion of no vaccinations among children aged 12-23 months. The indicators from the population census 2019 include population density (number per square kilometer), literacy rate of the population aged at least 7 years, the proportion of the population aged at least 15 years with primary education without completion, the proportion of households with city electricity, and proportion of households with piped water into a dwelling while median monthly income per capita in thousand Khmer riel is from the CSES 2019.

One OD will be randomly selected from each province, and purposeful selection will be conducted for the province with only one OD (Figure 2). The random selection will be performed after visiting study sites and excluding ODs located in the provincial town (urban area) and those without phone and internet coverage. Within each pair of the 10 selected provinces, we randomly assign one OD to an intervention arm and the other to a control arm. We will then randomly select two health centers from each OD. This random selection will be conducted after excluding health centers without cell phones and internet coverage. Currently, health centers provide a minimum package of activities, such as preventive and primary curative services, each can serve approximately 10,000 to 20,000 people in villages under its administration. The number of staff at health centers may vary between eight and 11. The staff members usually consist of medical doctors (uncommon), primary or secondary midwives, primary or secondary nurses, and others.

This trial will include three primary groups of participants. The first group will be the beneficiaries of the trial, including pregnant women and mothers of infants and their newborn babies or children aged 0-23 months enrolled in the intervention. The second group will be the i-MoMCARE facilitators (level 1 and level 2 tech support personnel), who will assist users of digital health technology. The third group will include the implementers or mobile application and web interface users, including VHSGs and health center staff.

Intervention arm Phase 1: Single-arm pilot of i-MoMCARE A single-arm pilot study will be conducted in the last quarter of year 1 to ascertain the reliability and success of its deployment. In the initial stage, we will adapt three steps from a previously successful application of digital health technology in Gujarat, India. The three steps will include (1) identifying gaps in MCH standard care, (2) identifying components of digital health that address the gaps in MCH standard care, and (3) piloting the operational delivery of the intervention.

Phase 2: Cluster randomized controlled trial An open cluster randomized controlled trial will be conducted to test the mobile application utilization and the intervention's effectiveness in improving MCH service coverage aided by the mobile application. We have estimated that approximately 200 VHSGs will be included in the intervention arm across five ODs in five provinces. We will select one OD from each selected province. Two health centers, each covering approximately 10 villages, will be randomly selected from each OD, while two health center staff will be purposively selected from each health center in the intervention arm. The designated staff will receive a web interface that will allow them to offer job aid for supervision, including real-time information about VHSGs' performance in the form of process indicators and coverage of MCH services. The tool will also support health center staff in their daily tasks, such as high-risk patient tracking, low-supply-inventory alerts, supply chain management, electronic health records, vital events tracking, and automatic calculation of performance-based incentives and motivation for VHSGs. The mobile application will be offered to two female VHSGs from each of the 200 villages connected to their respective health centers in the intervention arm. The mobile application will provide job aid to VHSGs for scheduling home visit tasks, including ANC, home-based newborn care, reporting outcomes of pregnancies, and follow-up visits of complicated cases. The application will also help with behavioral change communication (e.g., brief educational videos about danger signs in pregnancies and newborns to be shown to pregnant women), diagnosis, and patient management, such as identification and management of complications using an electronic checklist to remind VHSGs to perform, and record recommended examinations linked to an algorithm). At endline, we will conduct a qualitative acceptability and feasibility assessment of the intervention by conducting KIIs with 11 representatives from NMCHC, the five provincial health departments in the intervention provinces, and the five intervention ODs, IDIs with 10 VHSGs and 10 health center staff purposively selected from five of the 10 intervention health centers, and five FGDs with six to eight mothers of under-two children in the five health centers selected for the IDIs.

Control arm As discussed in the study design, health centers in the control arm will be randomly selected from their respective ODs, similar to those in the intervention arm. All villages under the health center's catchment areas in the control arm will continue to receive the MCH standard care provided by the government and other local and international non-governmental organizations (NGOs) and facilitated by VHSGs. VHSGs in both intervention and control arms will receive refresher training on maternal, neonatal, and childcare to avoid ethical issues that may arise and to ensure that change in outcomes in the intervention arm is due only to i-MoMCARE, but not the VHSGs MCH training.

Data collection and management Data collection, entry, quality assurance, and security maintenance will be supervised and coordinated by the Public Health Department at the university of health sciences (UHS), the lead implementer of the i-MoMCARE program. Technical support will be provided by the Saw Swee Hock School of Public Health (SSHSPH) of the National University of Singapore.

Mobile and web-based applications The process indicators will be collected from the web interface of the i-MoMCARE program during the intervention. Data from the mobile apps will be linked to the web interface and stored in an encrypted cloud storage hosted by the University of Health Sciences. Data will be backed up regularly into encrypted computers. Only research team members will have access to the data.

Quantitative household survey Household surveys will be conducted at baseline and end line to measure the primary and secondary outcomes of interest in both control and intervention arms. The household survey tool will be based on the tool used in the Cambodia Demographic Health Survey (CDHS) 2021. Mothers of children aged 6-24 months and pregnant women aged 18-49 will be eligible for the survey. The survey collects household demographic information, durable assets, access to antenatal care (ANC), delivery, and postnatal care (PNC), health insurance coverage, social health protection, health expenditure, and child anthropometric information. An online platform will be used to store the data, while an electronic tablet installed with the data collection application and equipped with internet access will be used to key in the data while interviewing. Data collection teams will be recruited and trained to conduct a pre-test of the questionnaire and interview the mothers. The village-level household population list from the population census in 2019 will be used as a sampling frame for household selection.

Qualitative data We will conduct key informant interviews (KIIs), in-depth interviews (IDIs), and focus group discussions (FGDs) at three study stages, including gap analysis, pilot, and intervention. Each qualitative interview should not be continued when data saturation is reached. Interviews, except KIIs, will be recorded using an audio recorder and transcribed for thematic analysis. In the gap analysis, we will conduct seven KIIs (a presentative from the national program and one representative from the provincial health departments and ODs of three purposefully selected intervention provinces), 12 IDIs (two VHSGs and two health center staff from three intervention provinces), and three FGDs with 6-8 mothers of children aged 6-24 (one from each of the three intervention provinces). At the pilot, to evaluate its feasibility and acceptability, we will conduct three KIIs (one representative from NMCHC, one from the provincial health department, and one from the OD of the pilot province), ten IDIs (6 VHSGs purposefully selected from the 50 VHSGs and two health center staff from each of the two pilot health centers), and two FGDs with 6-8 mothers of children aged 6-24 months. Last, at the end line, to assess the feasibility and acceptability of the mobile application, we will conduct 11 KIIs (one from NMCHC, one from each of the five provincial health departments in the intervention provinces, and one from each of the five intervention ODs), 20 IDIs (10 VHSGs and 10 health center staff purposively selected from five of the 10 intervention health centers), and five FGDs with 6-8 mothers of under-two children (one FGD in each of the five health centers selected for the IDIs). The audio-recorded will be uploaded in a secured computer folder by the study team. The records will be transcribed into Khmer language and store in another folder of the same secured computer. The transcriptions will then be translated into English and kept them in another separate secured folder. The study team will assign unique identification number instead of participants' personal identifiers to protect their privacy and confidentiality.

Analysis plan Since the randomization will be performed at the cluster level, it is implying that all pregnant women and infants in the intervention arm will be eligible for the program. Hence, the analysis of the i-MoMCARE's impact on outcomes of interest will be the intention to treat (ITT) analysis. Additional analyses will be per the study protocol. We will test (i.e., t-test and Chi-square test) the differences in means or proportions of the maternal and child characteristics in the intervention and control arms at the beneficiary and the cluster levels to observe balances in the study arms. In case of an imbalance in certain specific characteristics, a generalized estimating equation (GEE) approach will be applied to account for this imbalance. A difference-in-difference method using the GEE approach will be employed for all outcomes of mothers, pregnant women, and infants collected at baseline and end line. We will cluster standard errors at the village level to account for intra-cluster correlation in the characteristics of mothers. The data such as weight, height, gender, and date of birth will be entered into the software ENA2008 to compare the weight-for-age, height-for-age, and weight-for-height data to the international WHO reference standards to calculate the nutritional status of eligible children.

Qualitative data will be analyzed thematically using both inductive and deductive coding . Main and sub-themes will be developed using NVivo software according to the pre-define themes. Emerging themes will be included in the results as appropriate.

The economic evaluation will be performed to align with the principles, methodological specifications and reporting standards recommended by the International Decision Support Initiative Reference Case for Economic Evaluation (iDSI Reference Case). Health benefits and costs will be measured at the societal perspective to consider the impact at the programme, health system and patient levels. Using a bottom-up costing approach, we will estimate the recurrent and fixed cost, including manpower (e.g. staff salaries), capital (e.g. equipment), other operating cost (e.g. consumables) and overheads. Capital cost will be annualised over the expected lifespan of the product, taking into account the depreciation and opportunity cost.

Inclusion Criteria:

1. Mothers of children aged 6-24 months
2. Pregnant women aged 18-49