An analysis of practice-level mortality data to inform a health needs assessment.

BACKGROUND The utility of practice death registers has been indicated but, in the wake of the recent Harold Shipman case in the United Kingdom, the value of individual practice-level analysis has been questioned. AIM To assess the value of analysing practice-level mortality data to inform health needs assessment. DESIGN OF STUDY Comparative analyses of mortality. SETTING Two large practices, an inner-city study practice, and a reference practice in a medium-sized town. METHOD All premature deaths (aged one to 74 years) during 1994-1998 at the study practice (n = 170), and reference practice (n = 340), were identified. Cause-specific standardised mortality ratios (SMRs) were calculated using national reference data. The proportions of the total number of years of life lost (YLL) up to age 75 years associated with alcoholism, drug dependency, and severe mental illness were calculated and a comparison between practices was made, using standardised proportional mortality methods. RESULTS Significantly raised SMRs for the study practice were lung cancer (SMR = 234), digestive system diseases (SMR = 362), and injuries and poisonings (SMR = 180). Having standardised for age, there were nearly four times as many YLLs in the study practice population associated with a history of alcoholism, and over three times as many associated with drug dependency, compared with the reference practice. CONCLUSION Mortality analyses can provide useful insights for informing needs assessment in an individual practice. Small number problems may occur with smaller practice populations, but collation of data at PCG/T level also has potential utility. The study reinforces the argument that practices need to set up and maintain complete and accurate death registers.