Key points are not available for this paper at this time.
To better understand the generalizability of experimental smoking studies, the authors compare volunteer and population-based recruitment in two randomized clinical trials of smoking cessation methods, one that used a volunteer sample and one that used a random, population-based sample. Participants are compared with nonparticipants within each study, to the participants in the other study, and to the general population. Smokers aged 18-65 years without contraindications to using nicotine polacrilex who were able to quit unassisted were eligible for the two studies. In the first study (1984-1989), advertising recruited 3,758 eligible smokers (75% of all inquiries) who were willing to give a baseline telephone interview. Of these 3,758 eligible smokers, 2,592 attempted to quit and 1,226 succeeded for 48 hours and were randomized. In the second study (1988-1993), 7,135 smokers were identified through a random household telephone survey, and of these 7,135 smokers, 5,124 (72%) agreed to the telephone interview, 2,209 made a quit attempt, and 1,170 succeeded for 24 hours and entered the study. Compared with nonparticipants who provided an interview, subjects in both studies were more likely to be male, married, and employed, and were older and better educated. Compared with population-based recruitment, volunteer recruitment produced subjects who were less likely to be male and married, and were older, more educated, only slightly less representative of all ethnic subgroups, and heavier, more addicted smokers. Women were less likely than men to quit in the first 24-48 hours. The authors conclude that volunteer recruitment to smoking cessation studies appears to produce a reasonable sample of smokers, especially of heavy smokers, and that, compared with men, women may have more difficulty with the initial stages of smoking cessation.
Fortmann et al. (Fri,) studied this question.