Analytical approach:
The analysis synthesised data from recent clinical trials and epidemiological evidence. A Markov model, named the BENESCO (Benefits of Smoking Cessation on Outcomes) model, was developed. It was based on the established Health and Economic consequences of Smoking (HECOS) model and included enhancements. The model covered different time durations (two, five, 10 and 20 years) and a lifetime horizon. The authors stated that the perspective was that of the US health care system.
Effectiveness data:
The clinical estimates included continuous abstinence rates at one year, numbers of smoking-related diseases (coronary heart disease, lung cancer, chronic obstructive pulmonary disease and asthma) and deaths. Evidence of abstinence rates were abstracted from two head-to-head clinical trials of varenicline, bupropion and placebo, a Cochrane review of NRT that included a meta-analysis and epidemiological evidence of disease incidence.
Monetary benefit and utility valuations:
Utility valuations were based on a cost-effectiveness study that provided values by sex- and age-bands for individuals without smoking-related morbidity (Fiscella et al. 1996, see Other Publications of Related Interest for bibliographic details) and from eight published studies for individuals with smoking diseases.
Measure of benefit:
The measure of benefit used was quality-adjusted life years (QALYs) discounted at 3% per annum.
Cost data:
Direct medical costs were included in the analysis. The cost categories comprised drugs, dispensing fees, physician fees and treatment costs of subsequent smoking-related diseases. Drug costs were obtained from the US Red Book. A weighted average cost was used for NRT based on established US market shares of these options. Resource quantities and unit costs for the smoking-cessation treatments and annual costs for smoking diseases were presented and referenced. Costs were provided in US dollars ($) for the year 2005, inflated when necessary (index not stated) and discounted at 3% per annum.
Analysis of uncertainty:
Parameter uncertainty was addressed using one-way sensitivity analyses of the key variables (costs, efficacy rates, relapse rates, disease incidence, prevalence and mortality rates). Results were illustrated in a tornado plot. Probabilistic sensitivity analyses were performed and the results were provided in the text.
