PhD, University of Cincinnati, 2021, Medicine: Biomedical Informatics

Last-minute surgery cancellation, also known as day-of-surgery cancellation (DoSC), represents a substantial wastage of hospital resources and can cause significant emotional and economic implications for patients and their families. However, only few existing studies attempted to predict risk of cancellation for individual surgical cases, hampering the development of efficient interventions in clinical settings. Also, we currently lack knowledge of actionable factors underlying DoSC and barriers experienced by families (e.g., poor transportation access). The objectives of this dissertation are to 1) identify key predictors and develop machine learning models to predict cancellation for individual surgery schedules, and 2) understand potential underlying contributors to pediatric surgery cancellation at geographic level. In Aim 1, five-year data sets were extracted from the electronic health record (EHR) at Cincinnati Children's Hospital Medical Center (CCHMC). By leveraging patient-specific information and contextual data, a representative set of machine learning classifiers were developed to predict cancellations. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC) using ten-fold cross-validation. The best performance for predicting all-cause cancellation was generated by gradient-boosted logistic regression models, with AUC 0.793 (95% CI: [0.778, 0.808]) and 0.741 (95% CI: [0.725, 0.757]) for the two campuses. Of the four most frequent individual cancellation causes, no show and NPO violation were predicted better than patient illness or patient/family refusal. Models showed good cross-campus generalizability (AUC: 0.725/0.735, when training on one site and testing on the other). Feature importance techniques were applied to identify key predictors. An online tool for predictive modeling was developed using R Shiny package. In Aim 2, a five-year geocoded data set was extracted from the CCHMC EHR and an equiv (open full item for complete abstract)

Committee: Surya Prasath Ph.D. (Committee Chair); Richard Brokamp Ph.D. (Committee Member); Danny T. Y. Wu (Committee Member); Jayant Pratap (Committee Member); Yizhao Ni Ph.D. (Committee Member) Subjects: Health Sciences

Doctor of Philosophy, The Ohio State University, 2010, City and Regional Planning

The objectives of this dissertation are: (1) to develop a commuting spatial interaction model incorporating various spatial structure variables, such as Competing Destinations (CD) and Intervening Opportunities (IO) factors with Tobit regression; (2) to develop optimization planning applications with both linear and non-linear programming utilizing the empirically estimated Tobit models with the goal of minimizing total commuting costs, and (3) to assess multiple regional scenarios combining alternative land development strategies and zonal population density constraints. All the models are applied to the Fredericksburg Area (FAMPO Region) combining the counties of Stafford, Spotsylvania, King George, and Caroline, and the city of Fredericksburg in the State of Virginia. Census Transportation Planning Package (CTPP) 2000 data and Auditor's property data are used. The empirical results support considering the spatial structure of the region in the modeling process. The independent variable of this research, commuting flow, is a censored variable and seems appropriate to be handled with Tobit model. Four different spatial interaction commuting models have been estimated using Tobit regression: (1) Model 1 with basic variables of the gravity model; (2) Model 2 with two more additional variables, which represent spatial structure. Those are intervening opportunities (IO) and competing destinations (CD) variables; (3) Model 3 with further additional socio-economic variables; and (4) Model 4 extends Model 3 by adding quadratic terms. These four models are estimated by Tobit regression, and the results point to improvements by adding spatial structure variables and additional socio-economic variables and the quadratic functional form demonstrates the best goodness-of-fit statistics. To measure the impacts of independent variables on the dependent variable, two further analyses are conducted: (1) elasticity; (2) marginal effect analysis. It is shown that marginal effect of (open full item for complete abstract)

Committee: Jean-Michel Guldmann PhD (Advisor); Philip Viton PhD (Committee Member); Steven Gordon PhD (Committee Member); Gulsah Akar PhD (Committee Member) Subjects: Urban Planning