Use of subsequent anticoagulants to increase the predictive value of Medicaid deep venous thromboembolism diagnoses.

Linked data bases that derive their information from health care administrative sources are increasingly being used to conduct pharmacoepidemiologic research. Computerized case ascertainment using these data would be highly advantageous in terms of time and cost considerations. For a study of oral-contraceptive-associated deep venous thromboembolism, we evaluated the utility of using anticoagulant treatment codes to validate diagnostic codes suggestive of deep venous thrombosis and pulmonary embolism. By requiring evidence of outpatient anticoagulant use within six months of hospitalization, the predictive value of case ascertainment increased from 42% to 65% for "probable" deep venous thromboembolism and from 70% to 97% for "possible" deep venous thromboembolism. In addition, use of anticoagulant treatment codes as a second marker of disease resulted in nondifferential outcome misclassification when the study base was restricted to current oral-contraceptive users. Use of confirmatory treatment claims may provide a rapid, cost-effective alternative to medical-record-based case ascertainment for pharmacoepidemiologic studies of selected outcomes conducted in Medicaid and other linked universal health care coverage populations.