Reliability and validity of the Child and Adolescent Trial for Cardiovascular Health (CATCH) Food Checklist: A self-report instrument to measure fat and sodium intake by middle school students

Objective To develop a scoring algorithm and evaluate the reliability and v alidity of scores from the Child and Adolescent Trial for Cardiovascular He alth (CATCH) Food Checklist (CFC) as measures of total fat, saturated fat, and sodium intake in middle school students. Design Randomized, controlled trial in which participants were assigned to 1 of 3 study protocols that varied the order of CFC and 24-hour dietary rec all administration. Criterion outcomes were percent energy from total fat, percent energy from saturated fat, and sodium intake in milligrams. Subjects/setting A multiethnic sample (33% ethnic and racial minorities) of 365 seventh-grade students from 8 schools in 4 states. Statistical analyses Multivariable regression models were used to calibrate the effects of individual food checklist items; bootstrap estimates were u sed for cross-validation; and kappa statistics, Pearson correlations, t tes ts, and effect sizes were employed to assess reliability and validity. Results The median same-day test-retest reliability kappa for the 40 indivi dual CFC food items was 0.85. With respect to item validity, the median kap pa statistic comparing student choices to those identified by staff dietiti ons was 0.54. Test-retest reliability coefficients pranged from 0.84 to 0.8 9 for CFC total nutrient scores. Correlations between CFC scores and 24-hou r recall values were 0.36 for total fat, 0.36 for saturated fat, and 0.34 f or sodium; CFC scores were consistent with hypothesized gender differences in nutrient intake. Applications/conclusions The CFC is a reliable and valid tool for measuring fat, saturated fat, and sodium intake in middle school students. Its brevi ty and ease of administration make the CFC a cost-effective way to measure middle school students' previous day's intake of selected nutrients in scho ol surveys and intervention studies.