Estimating the Prevalence of Childhood Obesity in Alaska Using Partial, Nonrandom Measurement Data

Area of Science:

• Public Health
• Biostatistics
• Epidemiology

Background:

• Childhood obesity prevalence monitoring is crucial for public health initiatives.
• Few states possess comprehensive body mass index (BMI) measurement systems.
• School districts often collect student height and weight data during health screenings.

Purpose of the Study:

• To estimate childhood obesity prevalence in Alaska.
• To develop a logistic regression model utilizing existing school measurement data.
• To validate the model's estimates against established methodologies.

Main Methods:

• A logistic regression model was developed using student height and weight measurements.
• Public demographic and socioeconomic status data were incorporated.
• A mixed-effects model was employed to account for school-level and student-level variations.
• Model-generated estimates were validated against weighted estimates.

Main Results:

• The mixed-effects model provided validated prevalence estimates for childhood obesity.
• Estimates showed good agreement with weighted estimates, with overlapping 95% confidence intervals in 7 of 8 districts.
• The methodology effectively accounted for demographic and socioeconomic variations.

Conclusions:

• The described methodology offers a viable approach for estimating childhood obesity prevalence.
• This method can be applied by other states with existing nonrandom student measurement data.
• It supports public health programs in assessing obesity trends and intervention effectiveness.