This dissertation studies the integration of analytical modeling and optimization, machine learning, and Bayesian learning to optimize cost, access, and quality of healthcare delivery. Chapter 1 models computer-aided triage (patient prioritization) as feature-based priority queuing where types (diseases) are not perfectly observed but are inferred from observed features using a classifier. We propose a novel approach that optimizes the classifier to predict the priority queue directly from features. We present an analytic model to study how the optimal number of priority queues and the assignment of features to queues change with the classifier accuracy. Our numerical study using state-of-the-art image classifiers on an actual data set of medical images shows that our direct approach can improve delay costs relative to the practically appealing approach that combines an off-the-shelf type classifier with type-based priority queueing by up to 54%. Chapter 2 models the multi-period differential diagnosis problem as a Markov Decision Process. The qualitative insights from the model help investigate the optimality (or lack thereof) of rules of thumb used by physicians, such as confirmation bias (only considering the test for the most probable disease) or the more-sophisticated restricted rule-out (only considering the tests for the most probable or the disease with the greatest reward from correct diagnosis). Our analytical results help construct a tailored heuristic that outperforms all other heuristics. We use information relaxation and functional approximation to formulate a tight and tractable upper bound. Chapter 3 studies the integration of large medical claims and spatial road network databases to extract rural and urban demand, supply, and access (travel time) patterns for ancillary services across different provider types. We also present a data-driven location-allocation model to minimize payer cost and patient travel time. We find that the optimal solution shifts episodes from regulated (hospital outpatient departments) to unregulated (private physician offices and independent laboratories) sites---leading to significant savings but widening the extant travel time gap between rural and urban patients. Notably, imposing an extreme constraint of no increase in travel time retains about half the savings.

Creator

• Singh, Simrita

DOI

• https://doi.org/10.21985/n2-d2c3-jw42

Subject

• Operations research

Language

• en

Alternate Identifier

• etdadmin_upload_918485
• http://dissertations.umi.com/northwestern:16166

Date created

• 2022-01-01

Resource type

• Dissertation

Rights statement

• In Copyright