Validated Questionnaire for the Identification of Previous Personal or Familial Venous Thromboembolism

Abstract

The recognition of previous episodes of venous thromboembolism is of primary importance for the diagnosis of familial thrombophilia. Since there is no validated method to diagnose past venous thromboembolism for epidemiologic purposes, the authors administered a questionnaire to a group of subjects who had had venous thromboembolism in the past and to a group of normal controls. Both cases and controls were less than age 65 years and were identified by the discharge diagnosis International Classification of Diseases codes of records of the Vicenza general hospital, Italy, from 1975 to 1992. Cases were further validated by critical review of pertinent medical records. For each key question, the sensitivity and specificity were computed; questions were then aggregated into sets of maximal sensitivity and specificity. The best diagnostic efficiency (98.2%) was obtained by a set of questions about past symptoms and anticoagulant treatment; this set showed a very high specificity (99.4%), but a low sensitivity (37.1%). A higher sensitivity (71.3%) was obtained without loss of specificity (98.9%) by retaining as positive those subjects who gave a history of venous thromboembolism and anticoagulant treatment or those who had a history of venous thromboembolism and venous reflux at Doppler ultrasound scanning but had no treatment. The combined use of the questionnaire and Doppler scan is suggested for epidemiologic surveys in population. Am J Epidemiol 1996; 143: 1257–65.