Principal Component-Based Logistical Regression Algorithms to Predict Health care Accessibility for Texas Medicaid Gap

Abstract

Background12 states without expanded Medicaid caused 2 million people who were under the poverty line across the U.S to be in Medicaid limbo and not eligible for subsidized health plans on the Affordable Care Act insurance exchanges. In order to amplify geographic equity, this paper aims to explore the health access for Medicaid gaps in Texas. MethodsPrincipal Component-based logistical regression algorithms (PCA-LA) is provided data visualization and comparison in between unadjusted and adjusted Medicaid programs. Initially, Principal Component Analysis (PCA) eliminated well-known multiplicity problems between explanatory variables in the application of epidemiology. Optimized the traditional logistical Regression (LR), the PCA-LA method, is considered health status (HS) as a dependent variable with 0 (“poor” health) and 1 (“good” health), fourteen social-economic indexes as independent variables. ResultsAfter Principal Component Analysis (PCA), four composite components incorporated health conditions (i.e., “no regular source of care” (NRC), “Last check up more than a year ago” (LCT)), demographic impacts (i.e., four categorized adults (AS)), education (ED), and marital status (MS). Compared to the unadjusted LA, direct adjusted LA, and PCA-unadjusted LA three methods, the PCA-LA approach exhibited objective and reasonable outcomes in presenting an Odd Ratio (OR). They included that health condition is positively significant to HS due to beyond 1 OR, and negatively significant to ED, AS, and MS due to less than 1 OR. ConclusionsThis paper provided quantitative evidence for the Medicaid gap in Texas to extend Medicaid, exposed healthcare geographical inequity, offered a sight for the Centers for Disease Control and Prevention (CDC) to raise researchable direction of the Medicaid program and make a timely scientific judgment of Texas healthcare accessibility.