Letters Section Editor: Robert M. Golub, MD, Senior Editor.
In Reply: Dr Berger is concerned about the conclusion of our cohort study because it implies a causal relationship between smoking reduction and lung cancer risk. However, I believe the article discusses strengths and limitations of the study given the observational nature, and I do not think we infer more than data allow.
While in general observational studies can only be suggestive of causal relationships, the vast majority of evidence regarding behavioral risk factors for disease, such as smoking, originates from cohort and case-control studies, as it may not be feasible or ethical to conduct randomized trials of risk behavior vs no-risk behavior. This is particularly the case when analyzing modification of risk (harm reduction) such as smoking reduction. This difficulty was demonstrated in 2 randomized trials.1,2 In the first study,1 400 healthy smokers who were unwilling to quit smoking enrolled in a 2-year smoking reduction program, but only 25 succeeded in reducing the number of cigarettes by at least 50%. The latter was a population-based smoking cessation intervention study,2 and smokers who were unwilling to stop smoking were offered smoking reduction intervention. Less than 2% of smokers participated in the smoking reduction group.
Godtfredsen NS. Smoking Reduction and Lung Cancer Risk—Reply. JAMA. 2006;295(9):1001-1002. doi:10.1001/jama.295.9.1001-b