Validation of the DECAF score to predict hospital mortality in acute exacerbations of COPD

Background\ud Hospitalisation due to acute\ud exacerbations of COPD (AECOPD) is common, and\ud subsequent mortality high. The DECAF score was derived\ud for accurate prediction of mortality and risk strati\ud fi\ud cation\ud to inform patient care. We aimed to validate the DECAF\ud score, internally and externally, and to compare its\ud performance to other predictive tools.\ud Methods\ud The study took place in the two hospitals\ud within the derivation study (internal validation) and in\ud four additional hospitals (external validation) between\ud January 2012 and May 2014. Consecutive admissions\ud were identi\ud fi\ud ed by screening admissions and searching\ud coding records. Admission clinical data, including DECAF\ud indices, and mortality were recorded. The prognostic\ud value of DECAF and other scores were assessed by the\ud area under the receiver operator characteristic (AUROC)\ud curve.\ud Results\ud In the internal and external validation cohorts,\ud 880 and 845 patients were recruited. Mean age was\ud 73.1 (SD 10.3) years, 54.3% were female, and mean\ud (SD) FEV\ud 1\ud 45.5 (18.3) per cent predicted. Overall\ud mortality was 7.7%. The DECAF AUROC curve for\ud inhospital mortality was 0.83 (95% CI 0.78 to 0.87) in\ud the internal cohort and 0.82 (95% CI 0.77 to 0.87) in\ud the external cohort, and was superior to other\ud prognostic scores for inhospital or 30-day mortality.\ud Conclusions\ud DECAF is a robust predictor of mortality,\ud using indices routinely available on admission. Its\ud generalisability is supported by consistent strong\ud performance; it can identify low-risk patients (DECAF\ud 0\ud –\ud 1) potentially suitable for Hospital at Home or early\ud supported discharge services, and high-risk patients\ud (DECAF 3\ud –\ud 6) for escalation planning or appropriate early\ud palliation.\ud Trial registration number\ud UKCRN ID 14214.