Analisis Data Kategorik dengan Crosstabulation dan Correspondence Analysis (Studi Retrospektif Data Puskesmas Pringgarata Triwulan I)

Abstract

The use of categorical data in health sciences requires meticulous analytical methodology, such as correspondence analysis for exploring relationships and associations between variables. The objective of this study was to apply and evaluate categorical data analysis techniques, specifically Contingency Tables and Correspondence Analysis, to the Electronic Medical Record (EMR) data from Pringgarata Public Health Center during the First Quarter. This study is expected to provide insights into the distribution patterns of disease in the region. The method employed is a retrospective study of disease types using a Categorical Data Analysis approach. Cross-tabulation was performed for Disease Type with Month and Disease Type with Village (location). The Chi-square test was applied to assess the significance of the relationships, followed by Correspondence Analysis for visualizing patterns in a low-dimensional space. The results indicated that Fever (50.1%) was the most frequent case, with 54% of total cases originating from Pringgarata and Murbaya Villages. A significant association was found between Disease Type and Month (p-value=0.006), and between Disease Type and Village (p-value=0.017). The correspondence plot visualized strong associations: ISPA with March (indicating a seasonal trend), Gastritis with February (potentially linked to lifestyle patterns), ISPA strongly associated with Arjangka Village, and Gastritis strongly associated with Murbaya Village. The application of Correspondence Analysis successfully visualized specific relationship patterns, which can serve as a basis for planning resource allocation and targeting public health programs.