Predictive performance of the PUMA questionnaire as an opportunistic COPD case-finding tool in Singapore primary care

Background The PUMA scale has shown good discrimination in identifying people with COPD in primary care. We evaluated the predictive performance of PUMA for opportunistic case-finding and assessed the prevalence of COPD in at-risk primary care patients in Singapore. Methods This is a multicentre cross-sectional study of participants aged ≥40 years and current/former smokers. Participants completed the PUMA scale and spirometry. Predictive performance of PUMA was assessed using AUC-ROC; optimal cutoff was determined by Youden’s index. Results 359 participants were included in final analysis; 12.5% had COPD confirmed on spirometry. PUMA showed acceptable discrimination with AUC-ROC of 0.75 (95% CI:0.67–0.83). Optimal cutoff maximising sensitivity and specificity was ≥5 (Se 62.2%, Sp 79.3%; PPV 30.1%, NPV 93.6%); cutoff of ≥4 increased sensitivity to 80.0% (Sp of 56.7%; PPV 20.9%, NPV 95.2%.) Conclusion The PUMA scale demonstrated acceptable predictive performance for opportunistic COPD case-finding in Singapore's primary care setting. A cutoff of ≥4 enhanced case identification.