OBJECTIVES: To develop and validate a patient-reported outcome (PRO) instrument for measuring frequency, severity, and duration of COPD exacerbations in international trials.  METHODS: Multi-sponsor initiative involving experts and the FDA.  Phase I:  Instrument development informed by literature and qualitative data from focus groups and interviews with U.S. COPD patients.  International experts addressed concept relevance, transferability, translation ease. Phase II:  Prospective, 2-group, observational study of patients with 1) confirmed exacerbation (PDA diary Days 1–28 and 60–67) or 2) stable COPD (PDA for 7 days).  Item analyses and Rasch IRT were used to reduce the items from 23 to 14. Validity testing included known-group differences and relationship to clinical history, SGRQ-C, and clinician and patient assessment of exacerbation severity.  RESULTS: Phase I: N= 83; 45% male; 16% African American; 13% Hispanic; mean age = 65 (±10); stable FEV-1 = 44% pred (±16).   Phase II: N=410 (222 acute; 188 stable); 48% male; mean age = 65 (±10); stable FEV-1 = 51% pred (±20).  Factor analyses supported a unidimensional structure suitable for Rasch analysis.  Overall chi-square=149.3 df=84; person-separation index=0.92. Three respiratory domains (breathlessness (5 items), cough-sputum (2 items), chest symptoms (3 items)) account for 68% of the variance; 4 items address systemic manifestations. Internal consistency (Day 1, N=410) total = 0.91; domains = 0.87; 0.69; 0.87. Reproducibility (ICC) in stable patients (n=171) Day 1 to 7 = 0.77; 0.71; 0.65; 0.64. Correlation with SGRQ-C Day 1 (N=395): r = 0.64; 0.66; 0.35; 0.46 (p<0.0001). Significant differences in scores were found across exacerbation severity Day 1 (p<0.001) and between responders and non-responders Days 1 to 10 (p<0.0001).  CONCLUSIONS: The EXAcerbations of Chronic Pulmonary Disease Tool (EXACT) is a 14-item diary for evaluating exacerbation outcomes of COPD with measurement properties demonstrated to be adequate for use in clinical trials with a similar population.