Essays in Applied Economics Linking Policy, Obesity, and Health Economics

Chapter 1: Improved Health Insurance Coverage Rates for Adults with Low-Income and Chronic Disease Conditions: Evidence from 2014 Medicaid Expansion. This study examined health insurance coverage rates among low-income adults with chronic disease conditions before and after Medicaid expansion under the Patient Protection and Affordable Care Act (ACA). Data for this study was from the Behavioral Risk Factor Surveillance System (BRFSS) (2011 – 2018). The study sample includes nonpregnant, noninstitutionalized adults 18 to 64 years of age with household income below $15,000 per year. A difference-in-difference (DID) design with logit regression was used to examine the likelihood of insurance coverage before and after Medicaid expansion for adults living with at least one chronic condition. Among the sample living with a chronic disease, the increase in coverage in Medicaid expansion states was 10.96% (p-value <0.001). Based on race and ethnicity, the increase was highest for the White subgroup (14.50%; p-value <0.001) compared to other races and ethnicities. The results from this study suggest that Medicaid expansion improved healthcare access for low-income adults who have at least one chronic disease condition. Therefore, policymakers can usefully target low-income earning adults living with chronic health conditions to improve access. Chapter 2: Synthesizing Reviews for a Holistic Understanding of Real Causes and Potential Effect Sizes of Studies on Child and Adult Obesity: A Systematic Review Obesity is a growing epidemic in the United States, and interventions developed to date have not been sufficient to solve it. The influence of obesity worldwide has been on the rise with several impacts on the health and economic condition of affected individuals. This study conducts a systematic review and applies the socio-ecological model (SEM) to do a narrative synthesis of the impacts of the different levels of determinants (and their interaction) on adult and child obesity. This study paper seeks to add more evidence and uncover missed research areas by synthesizing existing evidence in systematic reviews and meta-analyses within the SEM and identifying levels that have not been studied extensively. Four electronic databases were searched for relevant articles until August 2020. 84 articles were included in this review for the narrative synthesis. The analysis showed that 41.67% of studies addressed the individual level of the determinant, 20.24% the interpersonal level, 11.90% the community level, and 2.38% addressed aspects from the societal & policy level. It was also found in this study that interactions between the individual (genetics) and interpersonal (relationships) levels are viable pointers as to why there is a difference in the development of obesity among individuals living in the same environment. This paper further highlights that incorporating multiple levels of the SEM has a greater impact on designing interventions to reduce the impact of obesity risk factors. Chapter 3: Analyzing Food Insecurity Coping Strategies in Malawi: A Machine Learning Application with Panel Data Famines and food shortages are becoming more common in Malawi, where smallholder farmers make up more than 80% of the population. Malawi is currently experiencing a high level of food insecurity for a variety of causes, such as high food costs, extreme climate and weather conditions, and land degradation. The main goal of this study is to investigate and evaluate how well comprehensive panel data and machine learning models can identify the characteristics of food-secure households in Malawi. This study used the World Bank IHPS data on household, climatic, and geographic factors obtained in Malawi in 2013, 2016, and 2019 to conduct this analysis. In this study, two food security outcomes—the reduced Coping Strategy Index (rCSI) and the availability of food for 7 days—are measured. A pooled and fixed effect logistic regression and machine learning algorithm namely random forest and extreme gradient boosting technique were estimated to find the most important predictors of food insecurity. Findings suggest that livestock ownership, the age of household heads, the education status of the household head, and coping strategies are key predictors of food security in Malawian households. We also find that climatic and geographic factors including annual precipitation, elevation, mean annual temperature, droughts experienced, the distance of the household to market, percent agriculture, the distance of the household to a major road, workability of the household farm plot contribute to predicting and understanding the food security status of households. Machine learning models predicted the food security status of households with an accuracy level as high as 82%. These results imply that ongoing investments in rural development through government assistance programs and education may contribute to an increase in household food security. Our Models combined with trained and validated machine learning methods alongside a rich variety of data can be replicated and deployed to study and estimate the food security status of many countries.

Author dc:creator

• Uche, Samuel

Contributors dc:contributor

• Amin, Modhurima
• Badruddoza, Syed
• Lacombe, Donald
• Lyford, Conrad

• Obesity
• Socio-ecological Framework
• Medicaid
• Causal Inference
• Machine Learning
• Food Security

Statement dc:rights

• Access is not restricted.

Language dc:language

en

Handle dc:identifier.uri

https://hdl.handle.net/2346/96937

OAI identifier oai:identifier

oai:tdl-ir.tdl.org:2346/96937