Using Big Data Techniques to Study Social Factors that Affect Child Health

Researchers at Weill Cornell Medicine recently developed an AI-based method to identify patterns in the conditions in which people are born, grow, live, and age commonly referred to as social determinants of health (SDoH), linking the patterns to the health outcomes of children. This novel approach provides a more holistic and objective picture of potential and current social factors which can affect the health of children.


The research team analyzed data obtained from over 10 thousand children across 17 states in the U.S. Each child had over 80 neighbourhood SDoH factors, with the analysis revealing four broad patterns in the sample. They include high-stigma environment, affluence, high socioeconomic deprivation, and high drug and crime sale rates. The team discovered statistical connections between these patterns relating to the developmental, mental, cognitive, and physical health of children.


‘A complex set of social factors can influence children’s health, and I think our results underscore the importance of using methods that can handle such complexity,’ said Dr. Yunyu Xiao, the lead author of the study and assistant professor of population health sciences at Weill Cornell Medicine.


In this study, lead authors Drs Xiao and Chang Su-experts in machine learning and AI- applied these ‘big data’ methods on social epidemiology problems. One of such problems was the examination of factors influencing the mental health of children during the COVID-19 pandemic.


‘Our approach is data-driven, allowing us to see what patterns there are in large datasets, without prior hypotheses and other biases getting in the way,’ said Dr Su.


The dataset in the novel study was generated by the Adolescent Brain Cognitive Development (ABCD) Study with a cohort of 10,504 American children.


The research’s key finding was categorized into four broad patterns- high-stigma, affluence, low level of education and urban high crime, and high socioeconomic deprivation. White children were the majority in the high-stigma and affluent patterns, and Hispanic and Black children were over-represented in the other two.


On the average, the high socioeconomic deprivation pattern was associated with the worst health outcomes including worse physical health and cognitive performance, and more signs of mental illness. The two non-affluent patterns were generally associated with more adverse outcomes.


However, the study was not without its limitations. The ABCD study data is not considered as reliable as objectively measured data, and epidemiological analyses are only capable of revealing relationships between health outcomes and social factors. They cannot prove that one can influence the other. Nonetheless, the researchers explained that the results show that a relatively impartial machine learning approach can be used to reveal connections and can help further studies discover causative mechanisms linking social factors to the health of children.


‘This multi-dimensional, unbiased approach in principle can lead to more targeted and effective policy interventions that we are investigating in a current NIH-funded project,’ said Dr Xiao.


By Marvellous Iwendi.

Source: Weill Cornell Medicine