The present study developed and validated a comorbidity index specifically for prediction of in-hospital mortality among diabetic inpatients in Canada. The analysis was based on data from the Hospital Person-Oriented Information Database (HPOI) for the study period from 1995/96 through 2000/01. The study included all the hospitalizations with a primary or secondary diagnosis of diabetes (ICD-9 code: 250.x) in acute care hospitals for patients aged 45 years or older with a length of stay of 90 days or less in ten provinces. All episodes of hospitalization for each patient were linked using a unique patient identifier, and one was randomly selected for the analyses. The study population of 578,222 diabetic inpatients was randomly divided into two parts, which were used either to develop or to validate the index. Multiple logistic regression models were used to develop and validate the index. A total of 22 diabetic comorbidities including 14 coexistent general medical conditions and 8 diabetic complications were included in the study-derived index, which had a better predictive performance as compared with D'Hoore-Charlson index and the simple count of comorbidities. The study-derived index can be used to control for potential confounding caused by comorbidity in the exploratory data analysis of diabetes research, to assist in creating more effective diabetes management system and to evaluate the prognosis of diabetic inpatients for health care provider.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/27100 |
| Date | January 2005 |
| Creators | Zhao, Wenxia (Helen) |
| Publisher | University of Ottawa (Canada) |
| Source Sets | Université d’Ottawa |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | 113 p. |
Page generated in 0.0017 seconds