Validation studies of claims data in the Asia-Pacific region: A comprehensive review

Nana Koram, Megan Delgado, James H. Stark, Soko Setoguchi Iwata, Cynthia de Luise

Published Year: 02/01/2019

Purpose: To describe published validation studies of administrative health care claims data in the Asia-Pacific region. Methods: A comprehensive literature search was conducted in PubMed for English language articles published through 31-Oct-2017 in humans from 10 Asian-Pacific countries or regions (Japan, Australia, New Zealand, China, Hong Kong, India, Singapore, South Korea, Taiwan, and Thailand) that validated claims-based diagnoses with a gold standard data source. Search terms included the: validation, validity, accuracy, sensitivity, agreement, specificity, positive predictive value, kappa, kappa coefficient, and Cohen’s kappa. Results: Forty-three studies across six countries were identified: Australia (21); Japan (6); South Korea (6); Taiwan (7); Singapore (2); and New Zealand (1). Gold standard diagnoses were obtained from: medical records (18); registry data (11); self-reported questionnaires (5); and other data sources (9). Validity measures used included sensitivity, specificity, positive and negative predictive values (12); sensitivity, specificity, and positive predictive value (4); sensitivity and specificity (4); sensitivity and positive predictive value (4); and combinations of other measures (19). Validated outcomes included medical conditions (28); disease-specific comorbidities (8); death, smoking, and other (ie, injury, hospital outcome measures) (5); medication/transfusion (2). Approximately 72% of the studies were published within the last 5 years. Conclusions: Validation studies of claims data published in the English language in the Asia-Pacific region are very limited. Given the increased reliance on administrative health care databases for pharmacoepidemiology and the need for ensuring the credibility of results from such data, additional support for the conduct of validation research of claims data in the Asia-Pacific region is needed.