Exploratory data analysis of racial/ethnic disparities in county hospitalization rates for New York State from 2007 to 2009
*Nathan T. Donnelly, University at Albany
Keywords: hospitalizations, racial/ethnic, health disparities, exploratory data analysis, factor analysis, New York State, diabetes, heart disease, stroke, unintentional injury
This study inspects differences in county age-adjusted hospitalization rates between the majority and minority racial/ethnic groups for diagnoses of Cerebrovascular Disease, Diseases of the Heart, Diabetes, and Unintentional Injuries. The New York State Health Department aggregated the data from hospital inpatient discharges by ICD-9 codes and posted the age-adjusted rates on their website. The dependent sample paradigm employed a pair of rates for each county. Difference scores were examined in R through graphics via the granova.ds function and image factor analysis. Counties with increased difference scores saliently emerged in visualization and ordering further revealed patterns across diagnoses, particularly in these upper tails of the sampling distributions. The sample generated high intercorrelations and the four sets of difference scores yielded a parsimonious one factor model fit (RMSE=.01) linked to large loadings for all diagnoses. The image factor methodology computed the following correlations with the singular latency: Diseases of the Heart (.88), Unintentional Injuries (.87), Diabetes (.84), Cerebrovascular Disease (.80). The results indicated a communality among quantities of racial/ethnic disparities in county age-adjusted hospitalization rates for these diagnoses. The hospitalization rate differences conditioned on race/ethnicity may derive from a single exposure, such as access to health care or low socioeconomic status. The inclusion of unintentional injuries limited a list of plausible latencies.
Important Dates & Deadlines
- October 9 - 11, 2013