Reliability of Height and Weight Measurements Taken by Physical Education Teachers for a School-based BMI Surveillance System

Background: School-based body mass index (BMI) surveillance is one approach to address childhood obesity. Cambridge Public Schools (CPS) in Massachusetts has maintained a BMI surveillance system since 2000, using trained physical education (PE) teachers to annually collect height/weight measurements of K-8th grade students.

Purpose: To estimate the reliability of height/weight measurements collected by PE teachers in a controlled and natural setting.

Methods: In the controlled setting, five randomly selected PE teachers measured 15 students' heights/weights in duplicate. In the natural setting, during class, eight PE teachers from four schools measured 105 students' heights/weights once, and a trained researcher measured the same students in duplicate. Intra- and inter-rater reliability were estimated by calculating intra- and inter-rater intraclass correlation coefficients (ICC), mean absolute differences, and mean range of measurement values per student.

Results: Controlled setting intra- and inter-rater height/weight ICCs were ³0.96. PE intra-rater mean absolute differences for height/weight were 0.52” and 0.8 lbs. Mean range of all measurement values were 2.18” and 4.6 lbs. In the natural setting, inter-rater ICCs were ³0.98. The overall mean height/weight absolute differences were 0.22” and 0.7 lbs, and of the eight PE vs. researcher comparisons, differences ranged from 0.10” to 0.35” and 0.1 lbs to 1.9 lbs.

Conclusions: Our ICC results demonstrate non-clinical school staff can collect student height/weight reliably for monitoring obesity and evaluating intervention programs. Equipment problems may have impacted weight measurement variability. Given the variability among the other measurements, attention to standardized rater training and precise equipment are recommended to improve reliability.