Objective: To assess the validity of hospital discharge abstract databases (DAD) in regards to patient comorbidities, with a focus on the trend of validity from 2002 to 2013.
Method: We compared patient comorbidity coding in DAD, against a reference standard of the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH) database. We calculated sensitivity, specificity, positive predictive value and negative predictive value for nine patient comorbidities from 2002 to 2013.
Results: 63 483 patients were linked between DAD and APPROACH. DAD coding validity varied depending on condition. From 2002 to 2013, hypertension and diabetes had consistent levels of validity, whereas hyperlipidemia, peripheral vascular disease, cerebrovascular disease and pulmonary disease had declining levels of validity. Validity trends for malignancy, heart failure and liver disease were inconclusive.
Conclusion: The level of DAD coding validity and the trend of validity over time is strongly dependent on the condition tested.