The paucity of effects in community trials: is secular trend the culprit?

Abstract

The most recent major U.S. trials that evaluated community-level programs to influence risk factors and health behaviors identified secular trends in the risk factors and health behaviors among the factors that might have limited community-level effects. The research reported in this paper uses data from one of the trials to examine the secular trend explanation directly. Data from the 22-community Community Intervention Trial for Smoking Cessation (COMMIT) were analyzed to test a hypothesis based on secular trend reasoning: program effects for smoking prevalence were larger for treatment communities matched to control communities with small declines in smoking than in treatment communities matched to control communities with larger declines in smoking. Consistent with the secular trend explanation, program effects were larger when control communities had relatively small declines in smoking prevalence. The findings suggest that secular trends masked community-level program effects in COMMIT.