WEDNESDAY, March 18 (HealthDay News) -- A widely used hospital quality measure that compares mortality rates and takes into account the mix of cases is inaccurate and may increase the bias that case mix adjustment is intended to decrease, according to research published March 18 in BMJ Online First.
Mohammed A. Mohammed, Ph.D., from the University of Birmingham in the United Kingdom, and colleagues retrospectively compared observed hospital deaths with deaths predicted by the Dr. Foster Intelligence case mix method, which is used to derive standardized mortality ratios for hospitals in England and has been adopted internationally to assess hospital quality. The researchers used data from 286,282 hospital admissions.
The investigators found systemic differences in the associations between hospital mortality rates and several of the case mix adjustment variables, particularly the Charlson (comorbidity) index and emergency admission. The interaction effects of the two variables could be explained by differences in clinical coding and admission practices across hospitals.
"These findings undermine the credibility of standardized mortality ratios and indicate that their role in labeling hospitals as good or bad is unjustified," write the authors of an accompanying editorial.