Understanding quality in data from hospital diagnosis and procedures: A statistical characterization

This work describes a characterization of hospital episodes through the study of an empirical distribution of data from diagnosis (n = 3.506.341) and procedures (n = 4.180.291), that occurred in public hospitals from the north region of Portugal, during the years of 2007 and 2008. The main purpose of this study was to propose the creation of a model for classifying this data and to identify outlier cases, using graph theory. With this work, we aim to contribute to a better quality in data within hospital information systems, including in electronic health record, in the automatic correction of data bases, or through the use of this data in statistical analysis and organizational reports.