Abstracts

Rationale: Risk adjustment in outcomes studies is necessary to account for patient characteristics and disease severity. Our group recently developed an epilepsy specific comorbidity risk adjustment index to be used for mortality outcomes research. The objective of the current study was to validate our epilepsy specific comorbidity index in a new population within the same health region.Methods: Data were extracted from four linked administrative health databases in Calgary, Canada from April 1, 2002 to March 31, 2010. These included a hospitalization database, an emergency visits database, a physician claims database and the provincial population health registry. Epilepsy patients were defined using a validated ICD-9-CM and ICD-10-CA based case definition (identified with a 3-year washout period to increase likelihood of identifying new cases). Total prognostic scores were calculated and compared for all subjects using the epilepsy specific comorbidity index (14 comorbidities) and the Charlson index (17 comorbidities). Comorbidities unique to the new epilepsy specific index included hypertension, cardiac arrhythmias, solid tumour without metastasis, paraplegia and hemiplegia (Charlson only includes hemiplegia), brain tumor, pulmonary circulation disorders, aspiration pneumonia and anoxic brain injury. Crude mortality and survival curves of both indices were compared. Survival curves were compared using the Kolmogorov-Smirnov test.Results: We identified 13,040 persons who met our case definition for epilepsy from a population of 1,408,647 people. The mean age of participants with epilepsy was 40.0 years (range 0-103.3) and 49.3% were male. Over the study period, 5.4% of those with epilepsy died. The most prevalent comorbidities amongst persons with epilepsy were depression (27.0%), hypertension (22.9%), chronic obstructive pulmonary disease (17.6%), cerebrovascular disease (13.1%), fractures (11.6%), alcohol abuse (8.6%), arrhythmia (7.1%) and head/brain injury (7.1%). The epilepsy specific comorbidity index appeared to demonstrate better discrimination between prognostic scores and survival when examining the survival curves plots visually but the difference was not statistically significant. Conclusions: The new epilepsy specific comorbidity index appears to perform similarly to the Charlson index for predicting mortality. However our index included comorbidities related to epilepsy and needs to be evaluated in various outcome measures such as readmission. Further validation in different health care settings and health regions is recommended.