Co-morbidity in an elderly inpatient population: are the risks of co-existing conditions differentially distributed?

This study examines the distribution of co-morbidity among 167,738 inpatients, aged 65 or more, who experienced an episode of hospitalization during 1999 in short-term institutions that are located in Oklahoma. The analysis was conducted in two phases. In the first, logistic regression analysis was used to examine covariates that separate inpatients who presented at least one secondary diagnosis from those who were not co-morbid. Limited to those whose condition was complicated by at least one secondary diagnosis, Probit analysis was used to assess the covariates of the number of co-existing conditions. The covariates examined were the individual's racial status, demographic attributes, primary diagnosis, source of admission and discharge destination. Results indicate that the oldest of the elderly, African Americans and, to a lesser extent, Native Americans presented more complex cases than other members of the study population. The paper concludes with the policy implications that are associated with the differential distribution of co-morbidity.