Optimal Screening, Monitoring, and Prevention Strategies in Infectious Disease Management

Abstract

Mathematical models of infectious disease such as Markov models, dynamic compartmentalmodels have been increasingly utilized in medical decision making. Most studies primarily focus on assessing the effectiveness, cost-effectiveness of policies, interventions by balancing costs and direct health benets (often in qualify-adjusted life-years gained, or disability-adjusted life-years averted). There are challenges with this classical approach. First, it may overlook the future impact of the current decision. For example, in treating bacterial infections, antibiotic over-prescription is an increasingly urgent healthcare issue to be addressed. Second, previous works focus less on incorporating individual response and heterogeneity effect into an infectious disease control policy optimization setting. In Chapter 2, we address the antibiotic over-prescription in febrile illness management,by formulating the problem of minimizing the weighted average of antibiotic underuse and overuse to inform the optimal diagnostic test and antibiotic treatment options for given occurrence probabilities of several bacterial and viral infections. The model accounted for multiple infections simultaneously and incorporated test, treatment, and other direct and indirect costs, as well as the effect of delays in seeking care and test turnaround times. We used the Markov models to numerically estimate disability-adjusted life years (DALYs), pre-penalty costs, and the likelihood of antibiotics overuse per patient for fifteen different strategies in Thailand settings (a typical viral and bacterial endemic setting). In Chapter 3, we formulate a Markov decision process to address patient adherenceheterogeneity by optimizing viral load monitoring strategies for HIV-infected patients. In Chapter 4, we provide a framework to optimize public health control policies inresponding to an infectious disease outbreak like the COVID-19 pandemic. We use a multinomial discrete choice model to characterize an individual activity level and integrate it into a repeated game-theoretical model with a SIR disease transmission dynamics. We derive a few insightful structural properties from these models and conduct numerical studies based on representative data for COVID-19 in Minnesota. We conclude with a discussion of the work and directions for future research inChapter 5.