Funding: This study was undertaken without additional funding by staff of the Northern Territory Government Department of Health and Families and the Menzies School of Health Research.
Performance of comorbidity indices in measuring outcomes after acute myocardial infarction in Australian indigenous and non-indigenous patients
Article first published online: 18 JUL 2012
© 2011 The Authors. Internal Medicine Journal © 2011 Royal Australasian College of Physicians
Internal Medicine Journal
Volume 42, Issue 7, pages e165–e173, July 2012
How to Cite
Condon, J. R., You, J. and McDonnell, J. (2012), Performance of comorbidity indices in measuring outcomes after acute myocardial infarction in Australian indigenous and non-indigenous patients. Internal Medicine Journal, 42: e165–e173. doi: 10.1111/j.1445-5994.2011.02539.x
Conflict of interest: None.
- Issue published online: 18 JUL 2012
- Article first published online: 18 JUL 2012
- Accepted manuscript online: 1 JUN 2011 01:38AM EST
- Received 22 November 2010; accepted 28 April 2011.
- Aboriginal Australian;
- acute myocardial infarction;
Background: Indigenous Australians have higher prevalence of chronic diseases and worse acute care outcomes than other Australians. The extent to which higher chronic disease comorbidity levels are responsible for their worse outcomes is not clear, and the performance of comorbidity indices has not been assessed for this population with very high comorbidity levels.
Methods: Using hospital separations data, the Charlson and Elixhauser comorbidity indices were used to measure chronic disease prevalence in 2035 indigenous and non-indigenous patients hospitalised after their first acute myocardial infarction (AMI) in the Northern Territory of Australia between 1992 and 2004, and to adjust for comorbidity in multivariate analysis of mortality outcomes (in-hospital and long-term deaths from coronary heart disease and all causes). Index performance was assessed by the difference between C statistic, Akaike information criterion statistic and estimate of excess indigenous mortality in models with and without comorbidity adjustment.
Results: Comorbidity index scores were higher for indigenous than non-indigenous patients and increased considerably over time, at least partly because of information bias. Indigenous patients' higher risk of in-hospital all-cause death was almost fully explained by their higher comorbidity levels. Their higher risk of long-term coronary heart disease and all-cause death was partially explained by higher comorbidity levels. Charlson and Elixhauser indices performed satisfactorily and similarly in this population.
Conclusion: Comorbidity indices performed well in a population with very high chronic disease prevalence. After adjusting for comorbidity, short-term outcomes were similar for indigenous and non-indigenous AMI patients, but comorbidity at the time of the acute episode only partly explained the worse long-term outcomes for indigenous patients.