In-hospital death and adverse clinical events in elderly patients according to disease clustering: the REPOSI study

Rejuvenation Res. 2010 Aug;13(4):469-77. doi: 10.1089/rej.2009.1002.

Abstract

Objective: The aim of the study was to recognize clusters of diseases among hospitalized elderly and to identify groups of patients at risk of in-hospital death and adverse clinical events according to disease clustering.

Method: This was a cross-sectional study conducted in 38 internal medicine and geriatric wards in Italy participating in the Registro Politerapie SIMI (REPOSI) study during 2008. The subjects were 1,332 inpatients aged 65 years or older. Clusters of diseases (i.e., two or more co-occurrent diseases) were identified using the odds ratios (OR) for the associations between pairs of conditions, followed by cluster analysis. Logistic regression models were used to evaluate the effect of disease clusters on in-hospital death and adverse clinical events.

Results: A total of 86.7% of the patients were discharged, 8.3% were transferred to another hospital unit, and 5.0% died during hospitalization; 36.4% of the patients had at least one adverse clinical event. Patients affected by the clusters, including heart failure (HF) and either chronic renal failure (CRF) or chronic obstructive pulmonary disease, had a significant association with in-hospital death (OR, 4.3;95% confidence interval [CI], 1.6-11.5; OR, 2.9; 95% CI, 1.1-8.3, respectively), as well as patients affected by CRF and anemia (OR, 6.1; 95% CI, 2.3-16.2). The cluster including HF and CRF was also associated with adverse clinical events (OR, 3.5; 95% CI, 1.5-7.8). The effect of both HF and CRF and anemia and CRF on in-hospital death was additive.

Conclusion: Several groups of older patients at risk of in-hospital death and adverse clinical events were identified according to disease clustering. Knowledge of the relationship among co-occurring diseases may help developing strategies to improve clinical practice and preventative interventions.

MeSH terms

  • Aged
  • Aged, 80 and over
  • Cluster Analysis
  • Female
  • Hospital Mortality*
  • Humans
  • Italy / epidemiology
  • Male
  • Medical Errors / statistics & numerical data*