Does considering severity of illness improve interpretation of patient satisfaction data?

Abstract

With the emergence of new, relatively low-cost code-based severity indexes, this question arises: Do complex descriptions of patient population in terms of severity yield a clearer picture of patients' opinions about hospital care and service? Consumers and third-party payers of healthcare are using patient satisfaction data with increasing frequency to evaluate the quality of care that hospitals provide. Insurers also use satisfaction data, when they are available, for contracting and ensuring provider accountability. The study described here examines whether the all patient refined-diagnosis related groups (APR-DRG) severity-of-illness rating system, in particular, can explain the variability in inpatient satisfaction ratings independently of patient demographics and clinical events. Multiple logistic regression was used on a data set of 3,720 patient records from one tertiary care facility, and model terms were fitted on the basis of reason for admission, year, gender, length of stay, age, and severity. The findings were that age and reason for admission were consistent predictors of high satisfaction on 14 survey items. APR-DRG severity was not a significant factor. Length of stay made a small but significant contribution on three items related to clinical quality.