A dropout analysis of the second phase of the Swedish part of the European Youth Heart Study

Abstract

The European Youth Heart Study (EYHS) addresses cardiovascular disease risk factors and their determinants in European children and adolescents. The Swedish part of the study began with cross-sectional data collection in 9- and 15-year-old schoolchildren in 1998–1999 (EYHS-I). Repeated observations of the key indicators were performed in 2004–2005 (EYHS-II). The purpose of this study was to assess potential dropout effects in EYHS-II. Participants in both EYHS-I and EYHS-II (n=459) were compared with dropouts who participated only in the EYHS-I (n=678) in relation to baseline physical activity, cardiorespiratory fitness and socioeconomic and anthropometric characteristics. Bivariate comparisons were performed using chi-square tests and gamma tests for nominal and ordinal data, respectively. Continuous data were compared by t tests and Mann−Whitney tests depending on the distribution. The Bonferroni correction was used to control for multiple hypothesis testing. Multiple logistic regression with backward elimination of variables was applied to study independent effects of variables on the probability of becoming a dropout. Analyses were performed separately for the younger and older age groups. The dropout proportion in EYHS-II was 60%. Subjects from the older age group were less likely to participate in the follow-up study (32% vs. 50%, p<0.001). In bivariate analyses, only maternal education was associated with dropout rates in the younger age group after Bonferroni correction. Males were more likely to drop out in both younger [odds ratio (OR)=1.72; 95% confidence interval (CI): 1.10, 2.96] and older (OR=1.96; 95% CI: 1.09, 3.54) age groups while basic maternal education was associated with outcome only in the younger group (OR=4.31; 95% CI: 1.78, 2.95) in regression analysis. The Swedish EYHS-II had high dropout rate after EYHS-I, but the dropouts did not differ from the participants in relation to physical activity, physical fitness, and anthropometric indices. Males were more likely to drop out than were females in both age groups. Differential dropout in relation to maternal education was observed in the younger age group.