Reliability of height and weight measurements collected by physical education teachers for a school-based body mass index surveillance and screening system.

BACKGROUND To address the obesity epidemic among children and youth, school-based body mass index (BMI) screening and surveillance is proposed or mandated in 30 states. In Cambridge, MA, physical education (PE) teachers are responsible for these measurements. This research reports the reliability of height and weight measures collected by these PE teachers. METHODS Using Bland-Altman plots, mean absolute differences, and intraclass correlation coefficients (ICC), we estimated intra- and inter-rater reliability among PE teachers in a controlled setting and PE teacher-vs-expert inter-rater reliability in a natural classroom setting. We also qualitatively assessed barriers to reliability. RESULTS For the controlled setting, of 150 measurements, 3 height (2.0%) and 2 weight (1.33%) measurement outliers were detected; intra-rater mean absolute differences for height/weight were 0.52 inches (SD 1.61) and 0.8 lbs (SD 3.2); intra- and inter-rater height/weight ICCs were ≥0.96. For the natural setting, of 105 measurements, 1 weight measurement outlier (0.9%) was detected; PE teacher-vs-expert-rater mean absolute differences for height/weight were 0.22 inches (SD 0.21) and 0.7 lbs (SD 0.8), and ICCs were both 0.99. Equipment deficiencies, data recording issues, and lack of students' preparation were identified as challenges to collecting reliable measurements. CONCLUSION According to ICC criteria, reliability of PE teachers' measurements was "excellent." However, the criteria for mean absolute differences were not consistently met. Results highlight the importance of staff training and data cleaning.