OPERATIONAL DESIGN AND QUALITY-CONTROL IN THE CATCH MULTICENTER TRIAL

Background. The Child and Adolescent Trial for Cardiovascular Health ( CATCH) was the first multicenter school-based research study to employ the fundamentals of clinical trials including the standardized protoc ol and Manuals of Operation, a steering committee for study governance , a distributed data system, an extensive quality control system, and a Data and Safety Monitoring Board. Method. CATCH tested the effective ness of changes in school lunches, physical education, smoking policy, curricula, and family activities. Ninety-six elementary schools in fo ur states were randomized to intervention or control conditions. The b aseline cohort comprised 5,106 ethnically diverse third graders follow ed through fifth grade. Results. The percentages of calories from fat and saturated fat were reduced significantly more in the intervention school lunches than among the controls. Significant increases in moder ate to vigorous activity levels in existing physical education classes mere made as well. Changes in self-reported dietary, physical activit y, and psychosocial measures were significant. There were no significa nt differences in the physiological measures. Measurement error was ge nerally low for all physiologic measures except skinfolds, indicating a high level of reliability. Across all sites, the coefficients of var iation for lipids, height, and weight were less than 3%, whereas for s kinfolds, they were considerably higher, ranging from 6 to 8%. Intracl ass correlations for Lipid studies were also uniformly high at 0.99. I nterobserver agreement scores for SOFIT were greater than 90% for 9 of the 11 activities observed. Data entry error rates were low with less than five errors per 1,000 fields for all forms. Conclusions. The CAT CH results provided more scientific evidence on the importance of scho ols in the population approach to health promotion. Many of the strate gies used in this complex multicenter trial in the areas of design and analysis, measurement, training, data management, and quality control protocols might be appropriate for adoption in other studies. (C) 199 6 Academic Press, Inc.