Integrating Community Based Participatory Research and Machine Learning Methods to Predict Youth Substance Use Disorders for Urban Cities in New Jersey

点击翻译按钮获取中文摘要

Project Summary My lab uses a community-based participatory research approach to reduce health disparities in substance use among Black and Hispanic youth in urban communities. We primarily work in New Jersey (NJ) due to our close ties with Paterson and East Orange, NJ which both have the highest number of substance use disorders in the State and the largest group of racial-ethnic minorities (e.g. Black and Hispanic) in the state. My lab intentionally works on the community level as we have found that one-size-fits all approaches to ending the youth substance use epidemic will not work, particularly in communities that have been historically marginalized. Our recent work has discovered that targeting individual level behaviors to promote behavior change may not be enough to end the youth substance use epidemic and in fact, understanding the role of neighborhood characteristics may be a more plausible strategy. In our work, we have shown that predominantly urban communities such as Paterson, NJ and East Orange, NJ have some of the lowest neighborhood resources associated with healthy youth development and therefore can contribute to likelihood of using substances and becoming addicted. In addition, the use of complex statistical methods and study designs, may contribute to lack of mistrust of researchers, participation in studies and of the data by community members. We hypothesize that within predominantly urban communities, there is variability in structural risk and asset-based neighborhood characteristics associated with youth substance use. In line with using a social determinants of health approach, environmental and place-based factors have long been equated with health outcomes such as respiratory conditions (e.g. asthma) among youth. However, determining the exact resources within the community that contributes to substance use disorders have not been discovered. The field of addiction does not know the exact characteristics within a neighborhood that can serve as either protective or risk factors to substance use disorders within an urban community. In this Pioneer proposal which is responding to the RFA-DA-23-026, “NIDA Racial Equity Visionary Award DP1 mechanism”, we will combine innovative approaches and multiple forms of data to investigate neighborhood level factors by using participatory methods to co-create machine learning systems to predict and prevent substance use disorders with community members. We intend for this project to promote co- learning between community members and researchers that can lead to sustainable solutions for the community. The proposed work will shed light on the importance of place in addiction and also work towards eliminating racial bias in data sets and predictive algorithms by incorporating community members in all stages of the model development process. Findings from this study have the potential to change the way we as researchers conduct substance use and misuse prevention research and the way in which we formally engage with community members. This work can contribute significantly to achieving health equity for Black and Hispanic youth in urban communities.