Targeting childhood obesity in schools: an examination of the stability and utility of the Value Added Index

• Published in: Pediatric obesity Vol. 9; no. 3; pp. 197 - 208
• Main Authors: Moreno-Black, G., Stockard, J.
• Format: Journal Article
• Language: English
• Published: England Blackwell Publishing Ltd 01-06-2014; Wiley Subscription Services, Inc
• Subjects: Adolescent; BMI screening; Body Mass Index; Child; childhood obesity; Childrens health; Feeding Behavior; Female; Health behavior; Health promotion; Health Surveys; Humans; Male; Obesity; Pediatric Obesity - epidemiology; Pediatric Obesity - ethnology; Pediatric Obesity - prevention & control; Pediatrics; Prevalence; School Health Services - organization & administration; school policy; Social Class; Socioeconomic Factors; United States - epidemiology; Value added; value added modelling; school policy; value added modelling; BMI screening; childhood obesity
• ISSN: 2047-6302; 2047-6310
• DOI: 10.1111/j.2047-6310.2013.00161.x

Summary
What is already known about this subject
High rates of childhood obesity and overweight have promoted interest in school‐based interventions.
Procter's Value Added Index (VAI) identifies schools with higher‐than‐predicted rates of obesity.
What this study adds
The VAI can be calculated with much simpler techniques than those proposed by Procter and associates.
VAI data from a span of 4 years indicated that rank orderings of schools vary widely over time.
This casts doubt on the utility of the VAI measure for targeting policy interventions, and we suggest that using a simpler method would be more effective, especially when obesity rates are high.
Background
High rates of childhood obesity and overweight have promoted interest in school‐based interventions. As a way to identify schools with high unexpected prevalence of obesity and the greatest need, Procter and associates developed a ‘Value Added Index’ (VAI). It compares rates of obesity in entry level and advanced students in elementary schools, quantifying the extent to which rates for advanced students are higher than what would be expected given entry level rates and socio‐demographic characteristics.
Methods
This paper replicates their analysis using data over a 4 year time span from 17 schools in the western United States. Our analysis compared results obtained with the relatively complicated mixed‐model approach, which was used by Procter and associates, and a more simple linear regression, which could be easily used by local school officials. Results were also compared across the 4 years for which data were available.
Results
Identical results were found when the two methods were compared. There was little stability in the rank ordering of schools, based on the VAI, from 1 year to another.
Conclusions
Our results cast doubts on the utility of the VAI for policy makers and suggest policy makers consider a universalistic, rather than targeted, approach to interventions.