Three essays on the data-driven analysis and modeling of public policy
Abstract/Contents
- Abstract
- The goal of the research presented in this dissertation was to apply mathematical modeling and optimization techniques to the evaluation and improvement of public policies. This data-driven approach was applied to three studies. The first study examined the problem of child obesity. We built a stochastic model to describe the evolution of US children's Body Mass Index (BMI) under a treatment plan targeting children with a BMI exceeding a certain threshold. We solved the optimization problem of assigning thresholds at different ages to minimize the projected obesity-related disease prevalence at age $40$ with a constraint on the total screening cost. Using National Longitudinal Surveys data on US children's BMI development, we found that our optimal screening policy results in reduced projected obesity-related disease prevalences compared to the United States Preventive Services Task Force (USPSTF) recommendations. Furthermore, our results suggest that the optimal screening policy would be to focus on older adolescents, instead of screening children uniformly throughout earlier childhood. The second study evaluated the ballistic imaging systems that help solve crimes by comparing newly acquired images of cartridge casings or bullets to a database of images obtained from past crime scenes using data from the Israeli police force. Specifically, we incorporated time and spatial location of new images into our matching decisions. Inclusion of such extraneous information allows the matching to favor pairs of images closer together in space and time. Our results show an increase in the detection probability compared to the traditional approach. Moreover, we extend our model to the case of the US and show the potential of expanding regional ballistic matching programs to a national program. The third study focused on strategies for allocating ready-to-use therapeutic foods and ready-to-use supplementary foods to children in developing countries. Using a longitudinal data set from the Bwamanda region in DR Congo, we built a stochastic model to track each individual child's height-for-age z score (HAZ) and weight-for-height z score (WHZ) throughout the first five years of life. We identified a simple strategy of allocating the entire budget to therapeutic food for prioritized children based on a linear combination of one's HAZ, WHZ and age. The results reduced the mean number of disability-adjusted life years (DALYs) per child during 6 -- 60 months of age compared to a class of benchmark policies. The results of these three studies demonstrate that formulating public policy tradeoffs as constrained optimization problems can indeed yield improvements in effectiveness or reduction of costs. In particular, the application of such techniques in healthcare and law enforcement can lead to better understanding of the interplaying factors behind the policy in question and help policy makers to make more informed decisions.
Description
Type of resource | text |
---|---|
Form | electronic; electronic resource; remote |
Extent | 1 online resource. |
Publication date | 2012 |
Issuance | monographic |
Language | English |
Creators/Contributors
Associated with | Yang, Yan |
---|---|
Associated with | Stanford University, Institute for Computational and Mathematical Engineering. |
Primary advisor | Wein, Lawrence |
Thesis advisor | Wein, Lawrence |
Thesis advisor | Goldhaber-Fiebert, Jeremy D |
Thesis advisor | Murray, Walter |
Advisor | Goldhaber-Fiebert, Jeremy D |
Advisor | Murray, Walter |
Subjects
Genre | Theses |
---|
Bibliographic information
Statement of responsibility | Yan Yang. |
---|---|
Note | Submitted to the Institute for Computational and Mathematical Engineering. |
Thesis | Thesis (Ph.D.)--Stanford University, 2012. |
Location | electronic resource |
Access conditions
- Copyright
- © 2012 by Yan Yang
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...