Identifying risk profiles for childhood obesity using recursive partitioning based on individual, familial, and neighborhood environment factors. - RIIP - Réseau International des Instituts Pasteur Accéder directement au contenu
Article Dans Une Revue International Journal of Behavioral Nutrition and Physical Activity Année : 2015

Identifying risk profiles for childhood obesity using recursive partitioning based on individual, familial, and neighborhood environment factors.

Résumé

Few studies consider how risk factors within multiple levels of influence operate synergistically to determine childhood obesity. We used recursive partitioning analysis to identify unique combinations of individual, familial, and neighborhood factors that best predict obesity in children, and tested whether these predict 2-year changes in body mass index (BMI). Data were collected in 2005-2008 and in 2008-2011 for 512 Quebec youth (8-10 years at baseline) with a history of parental obesity (QUALITY study). CDC age- and sex-specific BMI percentiles were computed and children were considered obese if their BMI was ≥95th percentile. Individual (physical activity and sugar-sweetened beverage intake), familial (household socioeconomic status and measures of parental obesity including both BMI and waist circumference), and neighborhood (disadvantage, prestige, and presence of parks, convenience stores, and fast food restaurants) factors were examined. Recursive partitioning, a method that generates a classification tree predicting obesity based on combined exposure to a series of variables, was used. Associations between resulting varying risk group membership and BMI percentile at baseline and 2-year follow up were examined using linear regression. Recursive partitioning yielded 7 subgroups with a prevalence of obesity equal to 8%, 11%, 26%, 28%, 41%, 60%, and 63%, respectively. The 2 highest risk subgroups comprised i) children not meeting physical activity guidelines, with at least one BMI-defined obese parent and 2 abdominally obese parents, living in disadvantaged neighborhoods without parks and, ii) children with these characteristics, except with access to ≥1 park and with access to ≥1 convenience store. Group membership was strongly associated with BMI at baseline, but did not systematically predict change in BMI. Findings support the notion that obesity is predicted by multiple factors in different settings and provide some indications of potentially obesogenic environments. Alternate group definitions as well as longer duration of follow up should be investigated to predict change in obesity.
Fichier principal
Vignette du fichier
art%3A10.1186%2Fs12966-015-0175-7.pdf (460.68 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...

Dates et versions

pasteur-01352669 , version 1 (08-08-2016)

Identifiants

Citer

Andraea van Hulst, Marie-Hélène Roy-Gagnon, Lise Gauvin, Yan Kestens, Mélanie Henderson, et al.. Identifying risk profiles for childhood obesity using recursive partitioning based on individual, familial, and neighborhood environment factors.. International Journal of Behavioral Nutrition and Physical Activity, 2015, 12, pp.17. ⟨10.1186/s12966-015-0175-7⟩. ⟨pasteur-01352669⟩

Collections

RIIP INRS-IAF
84 Consultations
147 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More