Do implementation issues influence the effectiveness of medications? The case of nicotine replacement therapy and bupropion in UK Stop Smoking Services

BACKGROUND Effective pharmacotherapies are available for smoking cessation but their efficacy is established through randomised controlled trials where the medication is supplied direct to subjects. In health care settings patient access to medicines is often less direct. The process for obtaining supplies of nicotine replacement therapy (NRT) is relatively easy for smokers attending National Health Service (NHS) Stop Smoking Services in the UK, whilst this is not necessarily the case for those wishing to using prescription only medicines (e.g. bupropion and varenicline). This study was a direct comparison of the short-term validated abstinence rates of NRT and bupropion in a clinical setting. METHODS Data were routinely collected from 2626 clients setting a quit date (82% of those registering) with two London NHS Stop Smoking Services that offered behavioural support combined with pharmacotherapy (NRT and bupropion). RESULTS Contrary to what would be expected from multiple randomised controlled trials, the CO-validated 3-4 week abstinence rate in clients using NRT was higher than for bupropion (42% versus 34%, p = .003). This difference persisted even when controlling for smoking characteristics, demographic variables and treatment variables 1.40 (95% CI = 1.08 - 1.83). CONCLUSION Given that the level of behavioural support received by clients on each medication was identical, the most plausible explanation for the difference in effectiveness between NRT and bupropion perhaps lies with how clients of the Stop Smoking Services obtained their medications. Obtaining NRT was relatively easy for clients throughout the study period whilst this was not the case for bupropion. This study suggests that implementation issues and/or self-selection may influence the effectiveness of medications in health care, as opposed to research, settings.