Background: there is a lack of outcome information with respect to older health service users. The purpose of this study was to examine 30-day in-hospital mortality and its predictors in all elderly patients admitted as a medical emergency to our hospital.
Methods: all patients admitted between 2002 and 2008 were studied, linking anonymised clinical, administrative, laboratory and mortality data. Significant univariate predictors of outcome, including co-morbidity and illness severity score, were entered into a multivariate logistic regression model, adjusting the univariate estimates of the effect of age on in-hospital mortality.
Results: we admitted 23,114 consecutive acute medical admissions between 2002 and 2008; 30-day in-hospital mortality was 20.7% in the over 75 age category versus 4.5% in those younger. The unadjusted OR for a 30-day in-hospital mortality in the over 75 category of 5.21 (95% CI 4.73, 5.73) fell to 4.69 (95% CI 4.04, 5.44) when adjusted for outcome predictors excluding acute illness severity and 2.93 (95% CI 2.50, 3.42) when acute illness severity was added as a covariate. When the interaction between age and co-morbidity is examined, the odds ratio adjusts to 3.22 (95% CI 2.63, 3.6).
Conclusion: acute illness severity is more important than co-morbidity in explaining the outcome in older patients admitted as medical emergencies. Service planning for acute elderly care should be based on effective disease management programmes but recognise the contribution of acute illness severity to outcome when conditions deteriorate.