The Bangladesh D-Magic Trial. Diabetes Mellitus: Action Through Groups or Information for Better Control?

As survival improves in low-income countries the problem of non-communicable diseases, including diabetes, increases. Diabetes currently affects between 20-30% of the population of Bangladesh either as pre-diabetes or fully expressed diabetes mellitus. The problem of diabetes and its complications is a growing concern among communities, health service providers and policymakers, yet raw data on the magnitude and underlying causes of the problem are scarce, especially in rural areas. Underlying the increasing prevalence in Bangladesh are patterns of dietary change, changing lifestyles and increases in other risk factors like lack of exercise and smoking. Our study will generate prevalence data on pre-diabetes, diabetes, and associated risk factors in rural villages in Bangladesh and explore the impacts of two interventions to improve the detection, management and prevention of diabetes in these deprived populations. The first intervention will use mobile phone technology to send health promotion information and targeted diabetes prevention and management messages to the population whilst also implementing simple messages and quiz 'apps' to facilitate health provider training and decision support in relation to diabetes care. The second intervention will focus on participatory community groups, engaging communities in a series of meetings in which the population themselves identify and address factors that may increase their risks of diabetes or prevent the effective management of the disease. It is hypothesised that collective community messaging and engagement in this way will enhance awareness and trigger behaviour change leading to lower burdens of risk factors and disease indicators, measured through large community surveys. The simultaneous evaluation of the two interventions in different areas will allow a comparison of the benefits of each compared to a 'control' area that only receives standard diabetes prevention and care services. A strong focus on implementation and process evaluation will identify the likely scalability and sustainability of these interventions.

Aim: To evaluate and understand the impact and cost-effectiveness of mHealth and a community mobilisation interventions to raise awareness, change behaviour and reduce the prevalence of diabetes and pre-diabetes and NCD risk-factors in rural Bangladesh.

Design: Three-arm cluster randomised control trial with 60 clusters. 20 clusters will receive the mHealth intervention, 20 will receive the participatory community mobilisation intervention and 20 will be designated control clusters, receiving only standard services. All study clusters will benefit from health systems strengthening initiatives.

Participants: Adults aged 20 years and above residing in the 60 study clusters (impact evaluation); health service providers, policy-makers and community groups (process and economic evaluation).

Methods: Sample surveys of 6000 participants will take place before and after intervention implementation. Diabetes prevalence will be estimated from random capillary blood glucose levels. Anthropometric and blood pressure measures and survey data on participant demographic and socio-economic characteristics, lifestyle and behavioural risk factors, diabetes awareness indicators and health seeking behaviour will be gathered. Process and cost data will be gathered using mixed methods.

Primary outcome: Prevalence of diabetes and pre-diabetes in individuals aged 20 years or more between intervention and control arms after the intervention

Selected secondary outcomes: self-awareness of diabetic status; smoking prevalence; physical inactivity; mean population diastolic and systolic blood pressure; mean population body mass index; mean population waist and hip circumference ratio; awareness of diabetes symptoms; utilisation of diabetic services.

Process evaluation will address the aim of understanding the operational factors affecting intervention delivery and impact. An economic evaluation will address the aim of understanding the interventions' cost-effectiveness.

This research has the potential to impact three main non-academic beneficiaries:

a) Individuals and communities in the study areas
Our study builds on increasing recognition of the potential role of mHealth and participatory interventions in facilitating enhanced awareness of health and its determinants among individuals and communities. Through active community engagement throughout the project, we anticipate that the study will increase individual and collective awareness of diabetes and its determinants, which will result in interrelated effects of improved understandings of the prevention and treatment of the disease, behaviour change, service use and effective management of community resources for healthier lives and improved diabetes management. In addition, should the interventions being evaluated prove effective, our project will have brought about direct quantifiable improvements in health of the study population. Should the interventions prove effective, they will be rolled out to the study control areas as a matter of priority.

b) Service providers & policy makers
One of the greatest challenges to improving health services in resource-poor settings is the absence of reliable data needed to understand populations, disease burdens and changing patterns of lifestyle and behavioural factors associated with increased risk of disease. Our project will generate locally relevant data needed to inform responsive health systems and service provision. At the health facility level, for example, better quality data regarding their catchment population and improved awareness will better inform the management of facility resources for diabetes prevention and management, leading to better quality services. Traditional data flows to higher levels of health service management will also strengthen opportunities for well-informed changes to practice to be implemented from within the hierarchy of the service supply side. At the district level, this may include evidence-based resource allocation and, assuming evidence of cost-effective intervention impact, inclusion of the intervention strategies in local or national implementation plans. At the same time, a better informed population, with increased individual and population awareness of diabetes, its determinants and consequences, will have enhanced capacity to hold service providers to account - an essential part strengthening health services.

c) Advocacy and implementation groups
The strong implementation research focus of our project and our commitment to share protocols, data and results throughout the study will inform diabetes and non-communicable disease advocacy and implementation groups throughout Bangladesh and globally. Lessons learned from our project and broad engagement with a range of stakeholders will enable others to build on the work we do share key lessons throughout their own networks.