Purpose: This study aimed to assess the usefulness of Danish patient registers for epidemiological studies of subdural hematoma (SDH) and to describe clinical characteristics of validated cases.
Methods: Using a patient register covering a geographically defined area in Denmark, we retrieved hospital contacts recorded under SDH International Classification of Diseases version 10 codes S065 and I620 in 2000-2012. Neurosurgeons reviewed medical records of all potential cases. Based on brain scan results, verified cases were classified by SDH type (chronic SDH (cSDH) or acute SDH (aSDH)). Thirty-day mortality and preadmission antithrombotic drug use were established through linkage to population-based registers. We calculated the positive predictive value of the SDH code and compared mortality and preadmission antithrombotic drug use of cSDH with those of aSDH (age-adjusted and sex-adjusted odds ratio (OR), 95% confidence interval (95%CI)).
Results: We verified the diagnosis in 936 of 1185 identified patients. The positive predictive value was highest for hospital contacts with principal discharge diagnosis code S065 (96%) but was low for other contact types under code S065 (25-54%), and only moderate for patients recorded under code I620 (62%). cSDH represented 57% of verified cases, and aSDH the remaining 43%. cSDH differed markedly from aSDH with regard to a number of clinical characteristics, including a much lower mortality (OR 0.2, 95%CI 0.1-0.3). However, preadmission antithrombotic drug use did not vary by SDH type (OR 0.9, 95%CI 0.6-1.2).
Conclusions: Danish patient registers are a useful resource for SDH studies. However, choice of International Classification of Diseases code markedly influences diagnostic validity. Distinction between cSDH and aSDH is not possible based on SDH diagnosis codes only. Copyright © 2016 John Wiley & Sons, Ltd.
Keywords: antithrombotics; pharmacoepidemiology; subdural hematoma; validity.
Copyright © 2016 John Wiley & Sons, Ltd.