Back to search

Cornell University

DEVELOPING BETTER TREATMENT FOR ALCOHOL USE DISORDERS WITH THREE TOOLS: DECISION THEORY, BIG DATA, AND MACHINE LEARNING

Abstract

dc:description.abstract

The core of addiction science is the conceptualization, development, testing, and dissemination of increasingly effective treatments for alcohol and substance use disorders (SUD). This dissertation presents contributions at multiple stages of this developmental process, using three high-impact tools to address novel, timely research questions. Paper 1 addresses the emerging field of support services for people in long-term recovery. In the early stages of treatment development, high-quality descriptive research designs using large datasets provide foundational information about exactly what needs treating. This project examined impairment in several domains of everyday functioning, including self-care, work performance, and social relationships. A key contribution was testing multiple potential drivers of impairment – including concurrent mental illness and social determinants of health – to avoid the stigmatizing assumption that all problems experienced during remission are caused by residual effects of SUD itself. Indeed, results point toward a mental-health focused agenda for treatment after one year of recovery. Rather than defining treatment targets, Paper 2 takes a conceptual, theory-driven approach to examine why harm reduction strategies may be ineffective at achieving well-defined harm reduction goals: reducing total alcohol use and high-risk drinking behaviors. By applying fuzzy-trace theory, we find evidence that low-risk drinkers likely think about alcohol-related risk using gist-based, categorical thinking, rather than taking a “trading off” approach to alcohol risk, wherein large amounts of alcohol are deliberately substituted with small amounts. This thinking is not about avoiding alcohol altogether, although such thinking would produce less risk taking. Results provide an empirical foundation for the development of intervention strategies aimed at training gist-based thinking about alcohol use. Project 3 aims to facilitate greater dissemination and uptake of an emerging treatment strategy known as peer recovery support services. Services with greater evidence of cost-effectiveness are more likely to be supported by insurance funding. This project combines the use of two existing datasets and machine learning to overcome a practical obstacle to estimating PRSS cost-effectiveness. Using the algorithm that resulted from this work, researchers can conduct cost-effectiveness analyses of prior clinical trials on PRSS even when a utility-compatible measure was not administered during data collection.

Degree

thesis:*
Name thesis:degree_name
Ph. D., Psychological Sciences and Human Development
Level thesis:degree_level
Doctor of Philosophy
Discipline thesis:degree_discipline
Psychological Sciences and Human Development
Grantor
Cornell University
Year dc:date.issued
2025

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Hayes, Bridget
Chair dc:contributor.chair
  • Reyna, Valerie
Committee members dc:contributor.committeemember
  • Sonnenstuhl, William
  • Mendle, Jane

Subjects

dc:subject × 6

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Bibid
16938316
Dc Identifier Other
ProQuest Submission ID: 14951
ProQuest Publication ID: 31999926

Chain of custody

source
Harvested from
Cornell University
Base URL
ecommons.cornell.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Hayes, Bridget. DEVELOPING BETTER TREATMENT FOR ALCOHOL USE DISORDERS WITH THREE TOOLS: DECISION THEORY, BIG DATA, AND MACHINE LEARNING. Doctor of Philosophy thesis, Cornell University, 2025. https://hdl.handle.net/1813/117568