BACKGROUND AND OBJECTIVE:The aim of this study was to develop and validate two models to estimate the probabilities of frequent exacerbations (more than 1 per year) and admissions for chronic obstructive pulmonary disease (COPD) that can be used in a primary care setting. METHODS:Information was obtained in a cross-sectional observational study on ambulatory COPD patients performed in 201 general practices located throughout Spain. The model for admissions included 713 cases, 499 for the developmental sample and 214 in the validation sample; the model for frequent exacerbations included 896 patients, 627 in the developmental sample and 269 in the validation model. Candidate variables to be included in both models were: age, sex, body mass index (BMI), FEV(1) as percent predicted [FEV(1 )(% pred.)], active smoking, chronic mucus hypersecretion (CMH) and significant comorbidity. RESULTS:The admission model contained 2 readily obtainable variables: comorbidity (OR = 1.97; CI 95% = 1. 24-3.14) and FEV(1)(% pred.) (OR = 0.72; 0.58-0.88, for every 10 units), and well calibrated in developmental and validation samples (goodness-of-fit tests: p = 0.989 and p = 0.720, respectively). The model for frequent exacerbations included 3 variables: age (OR = 1. 21; 1.01-1.44; for every 10 years of increasing age), FEV(1 )(% pred. ) (OR = 0.82; 0.70-0.96, for every 10 units) and CMH (OR = 1.54; 1. 11-2.14) and also well calibrated (p = 0.411 and p = 0.340 in the developmental and validation samples, respectively). CONCLUSIONS:Our results suggest that FEV(1) impairment explains part of the risk of frequent exacerbations and hospital admissions. Furthermore, CMH and increasing age are significantly associated with the risk of frequent exacerbations, but severity of exacerbations provoking hospital admissions is associated with the presence of significant comorbidity. These important and easily measurable variables contain valuable information for optimal management of ambulatory patients with COPD.