Diagnostic characteristics of questionnaires for identification of obstructive lung disease

Background: COPD is undiagnosed in almost two thirds of COPD patients. A two-step process including a questionnaire followed by a spirometry has been proposed. Diagnostic characteristics of questionnaires are therefore crucial. Aims and Objectives: to compare performance of three questionnaires: GOLD (5 items), IPAG (8 items), Platino (8 items) (Draft Evidence Review: Chronic Obstructive Pulmonary Disease: Screening. U.S. Preventive Services Task Force. August 2015). Methods: items shared by the three questionnaires were prospectively collected in 409 patients referred for lung function test in two labs. ROC curves were analyzed using FEV1/FVC either Results: The study population comprised 234 men and 175 women, median age was 62 years (IC95% 60 – 63), median tobacco use 10 Pack-Year (IC95% 3-15) and median BMI 26.4 kg/m² (IC95% 25.6-26.7). COPD prevalence was 41% and 30% using 0.7 and LLN threshold, respectively. Surface areas under the curve (SAU) were not significantly different between the three questionnaires irrespective of the used threshold : Table 1. Setting a sensitivity at 80%, the specificity was at best 49%, 31% and 29% for GOLD, IPAG and Platino questionnaires, respectively. Moreover, with decreasing diagnostic FEV1 (80, 70, 60 and 50% of theoretical values) there was a significant but slight increase in SAU. Conclusion: the diagnostic performance of questionnaires is low. Using them as a screening tool with high sensitivity, the specificity is low leading to spirometry in a large number of patients without COPD.