BACKGROUND: Chronic obstructive pulmonary disease (COPD) exacerbations are associated with adverse weather conditions. We examined if a forecasting alert service available to general practices in Bradford and Airedale for the winter of 2007-08 reduced COPD admissions. METHODS: We compared admissions in 2007-08 with 2006-07 when the service was not available. We examined admissions in December to March, and in the 7- and 14-day post-alert periods. There were five alerts in 2006-07 and four in 2007-08. We compared practices using the service to varying extents with practices not using it. RESULTS: Admission rate ratios for practices using the service were 0.98 [95% confidence interval (CI): 0.78-1.22] for December to March, and 0.82 (CI: 0.57-1.18) and 0.95 (CI: 0.72-1.26) for the 7- and 14-day post-alert periods, respectively. When we took into account the proportion of patients entered on the alerts system and the duration for which practices participated in the service, admission rate ratios for practices fully using the service were 1.11 (CI: 0.80-1.52), 1.22 (CI: 0.73-2.04) and 1.21 (CI: 0.82-1.78) for the three corresponding periods. CONCLUSION: We failed to show that any change in admissions associated with the forecasting service was significant. More research on the effectiveness of the COPD forecasting service is needed.