Comparing participants and nonparticipants recruited for an effectiveness study of nicotine replacement therapy

Interventions for smoking cessation have been typically evaluated on reactively recruited samples in clinical trials (efficacy trials). However, to have an impact on smoking rates in a general population, the intervention should also be evaluated with proactively recruited representative samples (effectiveness trials).

The characteristics of participants and two groups of nonparticipants recruited for a population-based nicotine replacement therapy study were compared.

All members of a large New England Veterans' Administration Medical Center were contacted, and interviews were completed with 3,239 identified smokers (at least 10 cigarettes per day). At the end of the interview, all smokers were offered participation in a multiple intervention study. Of the interviewed smokers, 2,915 verbally agreed to participate in the study (90%). Of those who gave initial verbal consent, 2,054 returned the written informed consent form and became participants (70%).

The participants (full consent group) differed significantly from both nonparticipant groups-that is, the smokers who were interviewed but declined participation by active refusal (survey only group) and those who gave verbal consent but passively refused participation by failing to return the written consent form (verbal consent only group). Participants were more likely to be married, younger, and female; to live with others; and to have previously used or considered using nicotine replacement therapy. The survey only group was also more likely to be in the precontemplation stage (54%), whereas the participants were more likely to be in the contemplation (46%) or preparation stage (35%). The verbal consent only group was intermediate of the other two groups in stage-of-change characteristics.

An important finding was that it is possible to recruit a large proportion of a sample of identified smokers to an nicotine replacement therapy study. However, the participants are likely to differ in significant ways from those who either actively or passively decline participation.