Overview
This research introduces an optimization framework designed to plan Community Health Worker (CHW) visit schedules for individuals with diabetes. The framework aims to maximize glycemic control at a community level, explicitly modeling the trade-off between screening new patients and providing management visits to individuals already enrolled in treatment. It incorporates patients' motivational states, which influence their decisions to enroll in or drop out of treatment programs.
Research Context
Diabetes is identified as a global health priority, particularly in low- and middle-income countries (LMICs). In these regions, more than 50% of premature deaths are attributed to high blood glucose. CHW programs are recognized as a means to deliver affordable and culturally appropriate solutions for the early detection and management of diabetes.
Many existing CHW programs utilize optimization techniques to allocate capacity across various services, such as tuberculosis, maternal and child health, and HIV. However, these programs often do not account for individuals' decisions regarding enrollment or adherence to care. Previous work on diabetes care planning has focused on identifying high-risk patients for screening or optimizing overall resource allocation across different diseases, often overlooking the dynamic nature of patient engagement. The motivation for patients to enroll or drop out of treatment programs, and the impact of intervention plans on these decisions, are important factors in diabetes care within LMICs. This framework seeks to address this gap by considering patient motivational states and their influence on treatment outcomes.
Approach
The core of the research is an optimization framework that determines personalized CHW visit plans. This framework builds visit schedules by estimating patients' health and motivational states. By doing so, it accounts for a patient's trade-offs when making decisions about treatment enrollment.
- The framework explicitly models the balance between two primary CHW activities: screening new patients to identify cases of diabetes and providing ongoing management visits to individuals already registered for treatment.
- Patient motivational states are integrated into the model. These states affect an individual's decision to enroll in or drop out of treatment, which, in turn, influences the effectiveness of the CHW intervention.
- The model's design is intended to reduce dropout rates and improve resource allocation by tailoring visit plans to these estimated patient states.
The developed approach was applied and evaluated using operational data collected from urban slums in India.
Findings
- The application of the optimization framework in urban slums in India demonstrated that it can achieve a reduction in fasting blood glucose levels by up to 25%. This was observed when compared to the best baseline method, while operating with the same CHW capacity.
- Experiments indicated that the approach maintained effective performance even under conditions of imperfect information.
- The model's consideration of patients' motivational states and their impact on treatment enrollment decisions is suggested to lead to reduced dropout rates and improved resource allocation.
Why This Matters
The documented reduction in fasting blood glucose levels and the demonstrated ability of the framework to reduce dropout rates and improve resource allocation suggest a potential to enhance diabetes care delivery in LMICs. By considering patient decision-making and motivational states, the approach offers a method for planning CHW interventions that may lead to more effective health outcomes in communities.