Modifiable factors associated with cognitive performance in Chinese adolescents: a national environment-wide association study

Authors：Shanshan Wang, Ya Wang, Yuhui Wan, Puyu Su, Fangbiao Tao, Ying Sun

Source：Eur Child Adolesc Psychiatry. 2023 May 19

DOI: 10.1007/s00787-023-02229-1

Abstract：

Growing evidence exists about the candidate factors of childhood cognitive performance, but mainly limited to single-exposure studies. We sought to systematically and simultaneously identify and validate a wide range of potential modifiable factors for childhood cognitive performance. We used data from five waves of data from the China Family Panel Studies (CFPS-2010, 2012, 2014, 2016 and 2018). Our analytical sample was restricted to those children aged 2-5 at baseline with valid exposure information. A total of 80 modifiable factors were identified. Childhood cognitive performance was assessed using vocabulary and mathematics test at wave 5. We used an environment-wide association study (EnWAS) to screen all exposure-outcome associations independently and used the least absolute shrinkage and selection operator (LASSO) variable selection algorithm to identify factors associated with cognitive performance. Multivariable linear model was then used to evaluate causal relationships between identified factors and cognitive performance. Of the 1305 participants included in the study (mean ± SD, 3.5 ± 1.1 years age at baseline, 45.1% girls). Eight factors were retained in the LASSO regression analysis. Six factors across community characteristics (percentage of poverty in the community; percentage of children in the community), household characteristics (family size), child health and behaviors (mobile internet access), parenting behaviors and cognitive enrichment (parental involvement in child' s education), and parental wellbeing (paternal happiness) domains were significantly associated with childhood cognition. Using a three-stage approach, this study validates several actionable targets for improving childhood cognitive performance.

中文摘要：

越来越多的证据表明儿童认知表现的候选因素，但主要限于单次暴露研究。我们试图系统地同时识别和验证儿童认知表现的各种潜在可改变因素。我们使用了来自中国家庭样本研究（CFPS-2010、2012、2014、2016和2018）的五次随访数据。我们的分析样本仅限于基线时具有有效暴露信息的2-5岁儿童。总共确定了80个可改变的因素。在第5次随访时使用词汇和数学测试评估儿童认知表现。我们使用全环境关联研究（EnWAS）独立筛选所有暴露-结局关联，并使用最小绝对收缩和选择运算符（LASSO）变量选择算法来确定与认知表现相关的因素。然后使用多变量线性模型来评估已识别因素与认知表现之间的因果关系。在纳入研究的1305名参与者中（平均±SD，基线时3.5±1.1岁，45.1%为女孩）。LASSO 回归分析中保留了8个因素。社区特征的六个因素（社区贫困百分比；社区儿童百分比），家庭特征（家庭规模），儿童健康和行为（移动互联网接入），育儿行为和认知丰富（父母参与儿童教育）以及父母幸福（父亲幸福）领域与儿童认知显著相关。本研究采用三阶段方法，验证了改善儿童认知能力的几个可行目标。