Methodologic issues in evaluating stop smoking programs.

Three methodologic issues (definitions of smoking, period of follow-up, composition of study group) were assessed in 426 persons five years after participation in a stop smoking program of a prepaid medical plan. When smoking was defined by measurement interval or by type or amount of tobacco smoked, smoking rates varied only slightly. Little information was gained by extending the follow-up period beyond the first year. However, study group composition (as defined by attendance at program sessions) had a pronounced effect on smoking rates. Those who attended fewer sessions were more likely to smoke during the follow-up period.