How rapidly does the excess risk of lung cancer decline following quitting smoking?
A quantitative review using the negative exponential model
Demonstrates that the predictions of the decline in excess risk following quitting are quite similar to those of the multistage model, often used for detailed analysis of individual subject data on smoking and lung cancer, and that estimates of H depend little on the assumed magnitude of the excess risk in current smokers.
Gives brief details of each of the 85 studies providing the data used for the modelling, including the six character study reference code (REF); a brief study description including the location, main characteristics of the study population, study design and study duration; the total number of lung cancers studied; and data sources.
Gives complete details of the analyses carried out using the negative exponential model, including the main analysis and the two sensitivity analyses. For each analysis, the output includes the 95% CI of the RR estimates shown in Table 3, the estimates and standard errors of all the fitted parameters, and the observed and fitted numbers of cases and controls at each level for each block.