Measuring severity of illness: six severity systems and their ability to explain cost variations

This paper presents results from a study that used a common set of patient records to compared how well different severity measurement systems are able to explain the variations in estimated costs among hospital patients. The systems examined were: APACHE II, MedisGroups, Computerized Severity Index (CSI), Disease Staging, Patient Management Categories (PMCs), and Acuity Index Method. In regressions on costs, all of the measures were found to improve upon DRGs for some types of cases but to offer little or no improvement for others. Indicators of maximum severity, especially Max CSI, explained greater proportions of cost variation than measures of admission severity and measures based on discharge abstracts. In most of the analyses, PMCs and Disease Staging yielded somewhat higher R2 values than the measures of admission severity.