Premium
Breastfeeding and the risk of childhood asthma: A two‐stage instrumental variable analysis to address endogeneity
Author(s) -
Sharma Nivita D.
Publication year - 2017
Publication title -
pediatric allergy and immunology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.269
H-Index - 89
eISSN - 1399-3038
pISSN - 0905-6157
DOI - 10.1111/pai.12750
Subject(s) - endogeneity , breastfeeding , medicine , instrumental variable , asthma , confounding , confidence interval , relative risk , pediatrics , demography , econometrics , economics , sociology
Abstract Background Several explanations for the inconsistent results on the effects of breastfeeding on childhood asthma have been suggested. The purpose of this study was to investigate one unexplored explanation, which is the presence of a potential endogenous relationship between breastfeeding and childhood asthma. Endogeneity exists when an explanatory variable is correlated with the error term for reasons such as selection bias, reverse causality, and unmeasured confounders. Unadjusted endogeneity will bias the effect of breastfeeding on childhood asthma. Methods To investigate potential endogeneity, a cross‐sectional study of breastfeeding practices and incidence of childhood asthma in 87 pediatric patients in Georgia, the USA, was conducted using generalized linear modeling and a two‐stage instrumental variable analysis. First, the relationship between breastfeeding and childhood asthma was analyzed without considering endogeneity. Second, tests for presence of endogeneity were performed and having detected endogeneity between breastfeeding and childhood asthma, a two‐stage instrumental variable analysis was performed. The first stage of this analysis estimated the duration of breastfeeding and the second‐stage estimated the risk of childhood asthma. Results When endogeneity was not taken into account, duration of breastfeeding was found to significantly increase the risk of childhood asthma (relative risk ratio [RR]=2.020, 95% confidence interval [CI]: [1.143‐3.570]). After adjusting for endogeneity, duration of breastfeeding significantly reduced the risk of childhood asthma (RR=0.003, 95% CI: [0.000‐0.240]). Conclusion The findings suggest that researchers should consider evaluating how the presence of endogeneity could affect the relationship between duration of breastfeeding and the risk of childhood asthma.