To examine the effectiveness and cost-effectiveness of nicotine replacement therapy (NRT) for 'cut down to quit' (CDTQ) smoking.
Major electronic databases were searched up to July 2006.
Data from studies meeting the criteria were reviewed and analysed. A decision analytical model was constructed to estimate the cost-effectiveness of CDTQ from the NHS perspective.
No systematic reviews of the effectiveness of CDTQ and no randomised controlled trials (RCTs) specifically addressing CDTQ were identified. Seven randomised placebo-controlled trials satisfied the inclusion criteria; six of these were industry sponsored. However, sustained smoking cessation was only reported as a secondary outcome in these trials and required commencement of cessation within the first 6 weeks of treatment. Meta-analyses of the study level results demonstrated statistically significant superiority of NRT compared with placebo. Individual patient data from unpublished reports of five RCTs were used to calculate sustained abstinence of at least 6 months starting at any time during the treatment period (generally 12 months). From this the meta-analysis indicated statistically significant superiority of NRT versus placebo [relative risk 2.06, 95% confidence interval (CI) 1.34 to 3.15]. The proportions achieving this outcome across all five RCTs were 6.75% of participants in receipt of NRT and 3.29% of those receiving placebo. The number-needed-to-treat was 29. This measure of sustained abstinence was used for economic modelling. No existing economic analyses of CDTQ were identified. A de novo decision analytic model was constructed to estimate the cost-effectiveness of making CDTQ with NRT available for smokers unwilling or unable to attempt an abrupt quit. The outcome measure was expected quality-adjusted life-years (QALYs). The model results suggest that CDTQ with NRT delivers incremental cost-effectiveness ratios (ICERs) ranging from around 1500 pounds/QALY to 7700 pounds/QALY depending on the age at which smoking cessation was achieved and the modes of CDTQ delivery. Assuming applicability to a single population, CDTQ was not cost-effective compared with abrupt quitting. If CDTQ with NRT were to be offered on the NHS as a matter of policy, the base-case results suggest that it would only be effective and cost-effective if a substantial majority of the people attempting CDTQ with NRT were those who would otherwise make no attempt to quit. This result is robust to considerable variation in the forms of CDTQ with NRT offered, and to the assumptions about QALY gained per quit success.
Meta-analysis of RCT evidence of quit rates in NRT-supported smoking reduction studies indicates that NRT is an effective intervention in achieving sustained smoking abstinence for smokers who declare unwillingness or inability to attempt an abrupt quit. The 12-month sustained abstinence success rate in this population (approximately 5.3% with NRT versus approximately 2.6% with placebo) is considerably less than that documented for an abrupt quit NRT regime in smokers willing to attempt an abrupt quit with NRT (which according to other systematic reviews is around 16% with NRT versus 10% with placebo). Most of the evidence of effectiveness of CDTQ came from trials that required considerable patient-investigator contact. Therefore, for CDTQ with NRT to generate similar abstinence rates for this recalcitrant population in a real-world setting would probably require a similar mode of delivery. The modelling undertaken, which was based on reasonable assumptions about costs, benefits and success rates, suggests that CDTQ is highly cost-effective compared with no quit attempt. CDTQ remains cost-effective if dilution from abrupt quitting forms a small proportion of CDTQ attempts. In an alternative analysis in which smokers who switch from an abrupt quit to CDTQ retain the success rate of abrupt quitters, all forms of CDTQ appear cost-effective. Randomised trials in recalcitrant smokers allowing head-to-head comparison of CDTQ delivered with various modalities would be informative.