We made the ad hoc observation that the benefits of the intervention were even more pronounced (ORa, 0.16; p = 0.01) when the comparison was restricted to the subset (approximately 59%) of the children in the intervention group and the control group whose families cooperated with all the active data collection requirements of the study during the follow-up. Within the intervention group, those who participated in the active follow-up also proved more likely to attend intervention sessions, supporting the conclusion that the observed reduction in healthcare utilization was in some manner directly related to exposure to the intervention. For the subgroup with follow-up cotinine data, no baseline difference in cotinine levels or the number of smokers in the home was observed between children in the intervention and control groups. Failure to participate in active follow-up was associated with greater ETS exposure at baseline in both experimental groups but was unrelated to participants race, gender, or baseline health-care utilization. In focusing on the group that complied with the follow-up requirements, one is presumably comparing a subset of the families in the intervention and control groups that were somewhat more positively predisposed toward the goals and requirements of the study. In the framework of the transtheoretical model of stages of change, a higher proportion of these individuals may have been at the “contemplation” stage with regard to changing smoking practices, rather than the “precontemplation” stage. Nevertheless, within this subset, children of the control group families showed little or no reduction in health-care utilization simply as a result of their participation in the study and the data collection, whereas those exposed to the intervention showed a marked reduction. buy claritin online
The point estimate of the effect of the intervention on the proportion of families allowing smoking in the home (OR, 0.24) is even larger than the estimate of the effect on utilization but is not statistically significant (p = 0.11). The reduced statistical precision due to the loss to follow-up may account for the failure to detect an effect on this outcome as well as in the CCR. The change in the CCR favored the intervention group, and the effect size (ie, the adjusted difference divided by the control group baseline SD) was moderately large ( — 0.38/1.11 = —0.34 SD) but was not statistically significant (p = 0.26). It is possible that a reduction in ETS exposure, on the order of the effect size observed in the present study, may be clinically significant, if not statistically significant in a sample of only 51 cases.