Background The COPD Assessment Test (CAT) is a valid disease-specific questionnaire measuring health status. However, knowledge concerning its use regarding patient and disease characteristics remains limited. Our main objective was to assess the degree to which the CAT score varies and can discriminate between specific patient population groups. Methods The Canadian Cohort Obstructive Lung Disease (CanCOLD) is a random-sampled, population-based, multicenter, prospective cohort that includes subjects with COPD (Global Initiative for Chronic Obstructive Lung Disease GOLD classifications 1 to 3). The CAT questionnaire was administered at three visits (baseline, 1.5 years, and 3 years). The CAT total score was determined for sex, age groups, smoking status, GOLD classification, exacerbations, and comorbidities. Results A total of 716 subjects with COPD were included in the analysis. The majority of subjects (72.5%) were not previously diagnosed with COPD. The mean FEV1 /FVC ratio was 61.1 ± 8.1%, with a mean FEV1 % predicted of 82.3 ± 19.3%. The mean CAT scores were 5.8 ± 5.0, 9.6 ± 6.7, and 16.1 ± 10.0 for GOLD 1, 2, and 3+ classifications, respectively. Higher CAT scores were observed in women, current smokers, ever-smokers, and subjects with a previous diagnosis of COPD. The CAT was also able to distinguish between subjects who experience exacerbations vs those who had no exacerbation. Conclusions These results suggest that the CAT, originally designed for use in clinically symptomatic patients with COPD, can also be used in individuals with mild airflow obstruction and newly diagnosed COPD. In addition, the CAT was able to discriminate between sexes and subjects who experience frequent and infrequent exacerbations. Trial Registry ClinicalTrials.gov; No.: NCT00920348 ; Study ID No.: IRO-93326.