OBJECTIVE: This study models the cost-effectiveness of tiotropium versus ipratropium in a veterans population while incorporating information on medication compliance, disease severity, and history of health care utilization. METHODS: Electronic medical records from the Veterans Affairs (VA) Maryland Health Care System for 2004 were analyzed. Inclusion criteria: 1) filled prescription for ipratropium in 2004; 2) pulmonary function test (PFT) results; 3) PFT-based evidence of chronic obstructive pulmonary disease (COPD). Hospitalizations and emergency room (ER) visits for COPD exacerbations, COPD severity, and medication adherence were identified via chart review.  The relative effectiveness of tiotropium was based on published clinical trial results. The incremental cost-effectiveness ratio (ICER) was calculated for 702 actual ipratropium patients and 702 modeled tiotropium patients for two effects: avoided exacerbations (ICERex) and avoided hospitalizations (ICERhos).  Sensitivity analysis was also conducted.  RESULTS: The ipratropium sample characteristics were: mean age of 69 years; 98 percent male; 21 percent Black; and 40 percent smokers.  The distribution by severity was: mild (7%), moderate (42%), severe (40%) and very severe (11%).  The total (exacerbation-related) ER visits and hospitalizations were 879 (171) and 462 (75), reflecting the exclusion of 40 ER and 9 hospital encounters following missed medications.  The overall ICERex and ICERhos were $1318.38 and $4284.75.  Tiotropium was dominant in very severe patients: ICERex of -$2099.00 and ICERhos of -$6297.00.  Tiotropium was dominant (-$4215.46) for patients with one hospitalization and ranged from $433.29 to $117.27 for patients with one to three ER visits.  The results were most sensitive to variation in tiotropium compliance, the cost of ipratropium, and the relative efficacy of tiotropium. CONCLUSIONS:  Assuming VA efficacy similar to published estimates, tiotropium is more cost-effective in patients with more severe disease or a history of ER visits and is dominant when considering patients with very severe COPD or with a previous hospitalization.