Official websites use .gov

A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS

A lock ( 🔒 ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Need Help

Application of Mediation Mixture Modeling and Machine Learning to Understand Heterogeneity of Risk Pathways on Youth's Onset of Substance Use

Award Number: R03DA065201
ORGANIZATION: NATIONAL INSTITUTE ON DRUG ABUSE
OPDIV: NIH
AWARD CLASS: DISCRETIONARY
AWARD ACTIVITY TYPE: SCIENTIFIC/HEALTH RESEARCH (INCLUDES SURVEYS)
PERIOD OF PERFORMANCE START DATE: 09/01/2026
PERIOD OF PERFORMANCE END DATE: 08/31/2028

Group Awards By:

View Award Description

Application of Mediation Mixture Modeling and Machine Learning to Understand Heterogeneity of Risk Pathways on Youth's Onset of Substance Use - Abstract The literature highlights the complex interplay of factors influencing the development of substance use. To increase the accuracy of identifying at-risk groups for substance use, this secondary data analysis project aims to apply advanced quantitative methods in studying the heterogeneity of risk pathways leading to youth's early initiation of substance use, using the Adolescent Brain Cognitive Development (ABCD) Study®, a national longitudinal dataset in tracking youths' behavioral health. Informed by two models in explaining the etiology of early substance use, the Deviance Proneness Model and the Negative Affect Regulation Model, the proposed study will examine the following mediation pathways: parent mental health  Youth externalizing spectrum  Substance use; in the Negative Affect Regulation Model: parent mental health  Youth internalizing spectrum  Substance use. With the recognition of the multifaceted nature of the etiology of substance use, prior studies assume homogeneity in testing risk mechanisms, revealing only dominant pathways shared by most participants, neglecting subgroup-specific mechanisms. Prior studies using machine learning data-driven approach primarily focused on the prediction of substance use, with less emphasis on connecting with theory- driven mediation approach. Built on recent advances in methodology, this is the first study to ensemble approaches, combing theory-driven mediation approach and data-driven machine learning approach, to explore the complex interactive features of factors in characterizing group-specific risk mechanisms. The study aims to assess heterogeneity in mediation pathways in influencing early substance use among youth. Specifically, this study will explore two different approaches in ensembling multiple methods: decision tree- structural equation modeling (DT-SEM), and combined mediation mixture modeling (MMM) and classification tree method to address the mediation pathways and associations involving distinct genetic, family and environmental backgrounds. In the formal approach, the mediation model will be split into different subclasses based on the predictive factors via decision tree modeling. In the latter approach, the MMM combines the mixture model and mediation model to estimate the heterogeneity of theory-driven pathways into subclasses, then, the classification tree estimates the prediction of the enumerated subclasses with a large number of observed genetic, family and environmental variables to understand the features of the subclasses. The employment of these two approaches aims to provide complementary perspectives for identifying subgroups to ensure a deeper understanding of heterogeneity of mediation pathways to youth substance use. Results will identify subgroups that are vulnerable to specific risk mechanisms, in an effort to unravel the multifaceted nature of early substance use. Ultimately, results will inform personalized interventions for identifying at-risk youth with early initiation of substance use.


Issue Date FY	Funding FY	Legal Entity Name	Legal Entity Address	Legal Entity City	Legal Entity State	Legal Entity Zip Code	Legal Entity COUNTY	Legal Entity COUNTRY	Assistance Listing	Award Code	Budget Year	Action Date	Action Type	Action Amount

Issue Date FY: 2026 ( Subtotal = $314,000 )
2026	2026	ARIZONA STATE UNIVERSITY	1475 N SCOTTSDALE RD STE 200	SCOTTSDALE	AZ	85257	MARICOPA	USA	Drug Use and Addiction Research Programs	000	1	8/31/2026	NEW	$314,000
														Subtotal = $314,000

Grand Total All Awards = $314,000

Top

All Categories

About

Search

Reports

Data Submission

Award Information

Application of Mediation Mixture Modeling and Machine Learning to Understand Heterogeneity of Risk Pathways on Youth's Onset of Substance Use

Award Number: R03DA065201

ORGANIZATION: NATIONAL INSTITUTE ON DRUG ABUSE

OPDIV: NIH

AWARD CLASS: DISCRETIONARY

AWARD ACTIVITY TYPE: SCIENTIFIC/HEALTH RESEARCH (INCLUDES SURVEYS)

PERIOD OF PERFORMANCE START DATE: 09/01/2026

PERIOD OF PERFORMANCE END DATE: 08/31/2028

Federal Websites

Department of Health & Human Services

HHS Operating Divisions

HHS Staff Divisions

Download A Document Viewer