Improving Tuberculosis Treatment Adherence Support: The Case for Targeted Behavioral Interventions

Problem definition: Lack of patient adherence to treatment protocols is a main barrier reducing the global disease burden tuberculosis (TB). We study operational design support (TAS) platform that requires patients verify their on daily basis. Academic/practical relevance: Experimental results effectiveness TAS programs have been mixed; and rigorous research needed how structure these motivational programs, particularly in resource-limited settings. Our analysis establishes engagement can be increased by personal sponsor outreach behavior data used identify at-risk for targeted outreach. Methodology: partner with TB provider use from completed randomized controlled trial. administrative variation timing peer evaluate impact messages subsequent verification behavior. then develop rolling-horizon machine learning (ML) framework generate dynamic risk predictions enrolled platform. Results: find that, average, increases odds ratio next-day 35%. Furthermore, patients’ prior accurately predict short-term (treatment verification) long-term (successful completion) outcomes. These allow target implement behavioral interventions patients. Managerial implications: indicate compared benchmark policy, could reach same number 6%–40% less capacity, or 2%–20% more using various ML-based prioritization policies leverage data. Personal all likely very costly, so may substantially improve cost-effectiveness programs. History: This paper has accepted Manufacturing & Service Operations Management Special Section Responsible Research Management. Funding: The authors are grateful financial MIT Sloan Health Systems Initiative. Supplemental Material: e-companion available at https://doi.org/10.1287/msom.2021.1046 .