Academic detailing and data feedback are two methods that have been used to change provider behavior. Academic profiling is proposed as an intervention that combines provider educational outreach and peer-comparison feedback of data generated from chart reviews and health plans. This project assessed the feasibility of academic profiling, using baseline measures to assess provider performance in identifying and treating patients who smoke. The pilot study was undertaken with four primary care practices in Maine. Two health plans shared administrative claims data on adult patients of participating providers. Two educational sessions were conducted: one including feedback of tobacco-related chart documentation and claims for nicotine replacement and bupropion (Zyban), and the other, coding for tobacco use (ICD-9 305.1) in adults enrolled in two health plans during 1998. A mailed survey assessed provider attitudes following the intervention. Among 24 providers, 80% attended the first session and 70% attended the second session. Provider documentation of tobacco status in the medical records varied from 68% to 100%. The frequency of tobacco pharmacotherapy claims for adult health plan enrollees having a provider visit in 1998 varied from 0% to 4.6% (mean 1.5%) by provider. The frequency of tobacco use diagnosis claims (ICD-9 305.1) varied from 0% to 19.8% by provider. More than 90% of the providers who reviewed the profiling graphs found the data were understandable, and 66% reported that the sessions helped them improve the ways they interact with patients who smoke. Practices vary in tobacco-related documentation, the prescribing of tobacco pharmacotherapy, and the coding for tobacco use. Providers are willing to participate in educational outreach using peer-comparison feedback, presenting opportunities to improve performance in the treatment of tobacco dependence.