Multi-dimensional predictors of first drinking initiation and regular drinking onset in adolescence: A prospective longitudinal study

Tam T. Nguyen‐Louie, Wesley K. Thompson, Edith V. Sullivan, Adolf Pfefferbaum, Camila Gonzalez, Sonja Eberson, Natasha E. Wade, Duncan B. Clark, Bonnie J. Nagel, Fiona C. Baker, Beatríz Luna, Kate B. Nooner, Massimiliano de Zambotti, David B. Goldston, Brian Knutson, Kilian M. Pohl, Susan F. TapertDevelopmental 2024 · DOI

TL;DR

This study identifies key behavioral and environmental factors that predict when adolescents first start drinking alcohol and when they transition to regular drinking. It finds that parental involvement, alcohol expectancies, and home access are strong predictors, and importantly, that delaying any alcohol initiation, not just regular use, significantly reduces the likelihood of experiencing withdrawal symptoms.

What Problem Does This Paper Address?

The transition into adolescent alcohol use is a critical developmental period, as early initiation is strongly linked to a host of negative consequences, including alcohol use disorder, academic problems, and other substance use. While many studies have identified risk factors for problematic drinking, there's a need to understand the specific predictors of when adolescents first start drinking (initiation) and when they move from occasional use to regular use. Dissociating these two onset points is crucial because not all who try alcohol become regular drinkers. This paper aims to identify robust, pre-drinking predictors of both first and regular drinking onset, and then to examine how these onset ages forecast later negative outcomes like binge drinking and withdrawal symptoms.

Methods

The study employed a two-part analytical approach using data from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) study.

Part 1: Predictor Identification * Best Subsets Selection & Cox Proportional Hazards (PH) Regression: This powerful combination was used to sift through 100 candidate baseline predictors (covering youth psychodevelopmental characteristics, cognition, brain structure, family, peers, and neighborhood factors) to identify the most robust predictors of both "First Drinking Onset" and "Regular Drinking Onset." Cox PH models are ideal for time-to-event data, accounting for participants who haven't yet experienced the event (censoring). * Optimism Bootstrapping: Used to assess model bias and overfitting, ensuring the identified predictors are likely to generalize. * False Discovery Rate (FDR) Correction: Applied to control for multiple comparisons, increasing confidence in the statistical significance of the findings.

Part 2: Forecasting Outcomes * Robust Linear Regressions: Used to examine the longitudinal relationship between drinking onset ages and subsequent binge drinking onset. * Zero-Inflated Poisson (ZIP) Regressions: Employed to model the number of lifetime withdrawal symptoms, which often have an excess of zero occurrences (i.e., many participants never experience withdrawal).

Datasets

The study utilized data from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) study, specifically baseline to Year 8 data releases. The sample consisted of 538 participants who were alcohol-naïve at baseline.

Key Insights and Ideas

  • Dissociating First vs. Regular Drinking: The study successfully differentiated predictors for first drinking initiation versus regular drinking onset, highlighting that these are distinct pathways.
  • Parental Solicitation as a Double-Edged Sword: While parental control and knowledge were protective, parental solicitation (actively inquiring about activities) was a strong predictor of earlier drinking onset. This suggests that while monitoring is good, the way it's done might influence initiation.
  • Home Access is Key: Easy access to alcohol at home emerged as a significant predictor for both first and regular drinking onset, underscoring the importance of environmental factors.
  • Alcohol Expectancies Matter Early: Positive alcohol expectancies, particularly those related to social changes and relaxation, were strong predictors of earlier drinking. This suggests interventions targeting these beliefs could be effective.
  • Brain Structure's Limited Predictive Power: While some brain structure measures (insula volume, parahippocampal surface area) showed associations, their predictive power was weaker than behavioral and environmental factors, suggesting a complex interplay.
  • Delaying Any Initiation is Crucial for Withdrawal: The most striking finding for Part 2 is that delaying first drinking onset had a significantly larger impact on reducing the likelihood of experiencing withdrawal symptoms (62% increase in odds of never experiencing withdrawal per year delay) compared to delaying regular drinking onset (29% increase). This emphasizes the critical window of early adolescence.

Key Results

  • Strongest Predictors of Earlier Onset: Positive alcohol expectancies (HRs=1.67-1.87), easy home alcohol access (HRs=1.62-1.67), more parental solicitation (HRs=1.72-1.76), and less parental control/knowledge (HRs=.72-.73).
  • Predicting Binge Drinking: Earlier first and regular drinking onset predicted earlier transitions into binge and regular binge drinking (βs=0.57-0.95).
  • Predicting Withdrawal: A one-year delay in first drinking onset increased the odds of never experiencing withdrawal symptoms by 62% (IRR=1.62), and a one-year delay in regular drinking onset increased these odds by 29% (IRR=1.29).

Limitations and Future Work

  • Sample Selection: Participants were excluded if they had already initiated drinking before study entry, which might limit generalizability to populations with earlier onset. The included sample was also younger at entry than the excluded group.
  • Predictor Exhaustiveness: While comprehensive, the study acknowledges that not all potential predictors were included (e.g., childhood trauma, sexual orientation, gender identity, family attitudes towards minority sexual identities). Future research should explore these.
  • Brain Structure Replication: The study notes that some brain structure findings were not replicated from prior literature and that the two significant predictors emerged in opposing directions, suggesting they might have been chance occurrences.
  • Independent Replication: The study used bootstrapping for bias assessment rather than replicating results in an independent sample, which is a standard limitation when dealing with complex models.
  • Future Directions: The authors suggest future studies should investigate brain-wide regional metrics and explore how predictors of drinking onset might also contribute to heavy episodic drinking and withdrawal. They also highlight the potential for using this foundational model with other large-scale national studies like ABCD to examine additional predictors and complex interactions.

Summary updated

Abstract

Early adolescent drinking onset is linked to myriad negative consequences. Using the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) baseline to year 8 data, this study (1) leveraged best subsets selection and Cox Proportional Hazards regressions to identify the most robust predictors of adolescent first and regular drinking onset, and (2) examined the clinical utility of drinking onset in forecasting later binge drinking and withdrawal effects. Baseline predictors included youth psychodevelopmental characteristics, cognition, brain structure, family, peer, and neighborhood domains. Participants (N=538) were alcohol-naïve at baseline. The strongest predictors of first and regular drinking onset were positive alcohol expectancies (Hazard Ratios [HRs]=1.67-1.87), easy home alcohol access (HRs=1.62-1.67), more parental solicitation (e.g., inquiring about activities; HRs=1.72-1.76), and less parental control and knowledge (HRs=.72-.73). Robust linear regressions showed earlier first and regular drinking onset predicted earlier transition into binge and regular binge drinking (βs=0.57-0.95). Zero-inflated Poisson regressions revealed that delayed first and regular drinking increased the likelihood (Incidence Rate Ratios [IRR]=1.62 and IRR=1.29, respectively) of never experiencing withdrawal. Findings identified behavioral and environmental factors predicting temporal paths to youthful drinking, dissociated first from regular drinking initiation, and revealed adverse sequelae of younger drinking initiation, supporting efforts to delay drinking onset.
BibTeX
@inproceedings{nguyenlouie2024multi,
  author = {Tam T. Nguyen‐Louie and Wesley K. Thompson and Edith V. Sullivan and Adolf Pfefferbaum and Camila Gonzalez and Sonja Eberson and Natasha E. Wade and Duncan B. Clark and Bonnie J. Nagel and Fiona C. Baker and Beatríz Luna and Kate B. Nooner and Massimiliano de Zambotti and David B. Goldston and Brian Knutson and Kilian M. Pohl and Susan F. Tapert},
  title = {{Multi-dimensional predictors of first drinking initiation and regular drinking onset in adolescence: A prospective longitudinal study}},
  booktitle = {Developmental},
  year = {2024},
  doi = {10.1016/j.dcn.2024.101424},
}
Figure