The relationship between a severity of illness indicator and mortality and length-of-stay

One important determinant of the cost and outcome of medical care is the severity of illness; there is no routinely available severity measure used in the United Kingdom. An American clinical information system, Medisgroups, was used to allocate 2,279 general medical patients to one of five admission severity groups in an English teaching hospital. The results showed a highly significant association between increasing severity and both length-of-stay and mortality. However, a further analysis of patients in the ten most common Diagnosis Related Groups showed that diagnostic group alone accounted for about twice the amount of variation explained by severity. The addition of severity to Diagnosis Related Groups only gave a very small rise in the overall explanatory power. These results are broadly consistent with those from larger American studies, which suggest that severity of illness at admission may not be as important as widely supposed in explaining variations in length-of-stay and mortality.