Identification of patients with systemic lupus erythematosus in administrative healthcare databases

Abstract

Objective

Our aim was to validate and compare decision rules for the identification of patients with systemic lupus erythematosus (SLE) in administrative healthcare databases.

Methods

A retrospective cohort study was performed using administrative health care data from a population of 1 million people with access to universal healthcare. Information was available on hospital discharges and physician billings over a 10-year period. Each SLE case was matched 4:1 by age and gender to randomly selected controls. Seven case definitions were applied to identify SLE cases and their performance compared with the diagnosis by a rheumatologist.

Results

We identified 373 SLE cases and 1492 non-SLE controls, all of whom had been reviewed by a rheumatologist. The overall accuracy of the case definitions for SLE cases varied between 88.2–95.6% with a kappa statistic between 0.53–0.86. The sensitivity varied from 41.0–86.6% and the specificity between 92.4–99.9%. In a total reference population of 1 million the mean estimated annual incidence of SLE was between 29–255 and the mean estimated annual prevalence was between 172–920.

Conclusion

The accuracy of case definitions for the identification of SLE patients in administrative healthcare databases is variable and this should be considered when comparing results across studies. This variability may also be used to advantage in different study designs depending on the relative importance of sensitivity and specificity for identifying the population of interest to the research question.