To the Editor.
—The study by Dr Hurt and colleagues1 evaluates the effect of nicotine patch therapy combined with physician advice on smoking cessation. Although it was a randomized study, there were important differences between the active and the placebo groups for several baseline characteristics, including the proportion of those employed (85% vs 73%), depressed or "down" (17% vs 23%), previously treated for depression (8% vs 13%), mean pack-years of cigarette smoking (35 vs 40), mean serum cotinine level (261 vs 295 ng/mL), and mean baseline withdrawal symptom score (0.95 vs 1.11). The direction of the differences for these variables could lower the a priori chance of cessation success for the placebo group.Randomization does not always result in groups that are alike, especially when the sample is small. If randomization fails with respect to a known confounder, such discrepancies can be controlled for in the analysis.2The
Lahad A. Nicotine Patch for Smoking Cessation. JAMA. 1994;272(1):32. doi:10.1001/jama.1994.03520010042028