Psychometric evaluation of the COPD assessment test: Data from the BREATHE study in the Middle East and North Africa region

The objective of this study was to assess the validity and performance of the Arabic and Turkish versions of the COPD Assessment Test (CAT) for evaluating the severity and impact of COPD symptoms. The data were obtained from the BREATHE study in the Middle East and North Africa region, a large general population survey of COPD conducted in ten countries of the region (Algeria, Egypt, Jordan, Lebanon, Morocco, Saudi Arabia, Syria, Tunisia, Turkey and United Arab Emirates), using a standardised methodology. A total of 62,086 subjects were screened, of whom a random sample of 5,681 subjects were administered the CAT by telephone. 5,639 evaluable questionnaires were recovered, representing a completion rate of 99%. In addition, the CAT was administered to an additional 833 subjects fulfilling the epidemiological diagnostic criteria for COPD. Mean scores in the general population were 6.99 +/- 6.91 for the Arabic version and 9.88 +/- 9.04 for the Turkish version. In patients with COPD, mean scores were 16.2 +/- 9.1 and 20.9 +/- 10.2 respectively. Scores were consistently higher in smokers than in non-smokers. In the general population, the proportion of respondents fulfilling criteria for COPD rose with higher CAT scores, and particularly above the 80th percentile, where 63% of COPD cases were to be found. This suggests that the CAT may be useful as a case-finding tool in the general population. In the CORD population, healthcare resource consumption rose linearly with CAT score above a threshold score of twenty, arguing in favour of the good criterion validity of the CAT. The internal consistency of the CAT was high (Cronbach's alpha 0.85 for the Arabic and 0.86 for the Turkish versions) and the factorial structure was unidimensional. In conclusion, this study performed in Arabic and Turkish speaking populations confirms the utility and validity of the CAT as a simple tool to collect data on the severity and impact of CORD symptoms, and suggests that it may potentially be useful as a case-finding tool to identify people at risk for CORD in the general population. (C) 2012 Elsevier Ltd. All rights reserved.