Psychology ETDs

Publication Date

Spring 8-3-2018


Given the widespread costs associated with alcohol use disorder (AUD; World Health Organization, 2011), it is unsurprising that many treatments exist for AUD. Moreover, many treatments have been rigorously studied via experimental research designs. In such research, treatment success has been defined predominantly as abstinence from alcohol or, more recently, no heavy drinking days. Consumption-based definitions of treatment success, rather than alternative non-consumption based definitions, have dominated in the field for at least two reasons. First, there are multiple measures of similar non-consumption constructs (e.g., quality of life, psychosocial functioning), and very little research has been conducted to direct researchers toward the best non-consumption measures to use among AUD populations. Second, it is assumed that non-consumption measures are insensitive and, therefore, consumption must be used as a surrogate measure for more clinically meaningful non-consumption measures. The present research study empirically addressed these two barriers that have thwarted attempts to shift toward including non-consumption variables in our definitions of treatment success. Using secondary data analysis of data collected from the COMBINE Study (Anton et al., 2006) and Project MATCH (Project MATCH Research Group, 1997), the present study conducted several tests of measurement stability, reliability, validity, sensitivity, and specificity. To test measurement stability the current study examined effect sizes and measurement invariance across time to test if non-consumption measures may be viable options for comparing pre- and post-treatment scores on these measures. The present study also conducted analyses on psychometric properties of extant measures: internal consistency reliability, construct validity, convergent validity. Finally, receiver operating characteristic curve analyses were conducted of total scale scores, subscales, and individual items when available and appropriate to test the sensitivity and specificity of non-consumption measures in detecting post-treatment and 12-month outcomes. The Brief Symptom Inventory (BSI), Beck Depression Inventory (BDI), and the brief World Health Organization Quality of Life measure (WHOQOL-BREF) were invariant across time and performed the best overall across all psychometric and sensitivity/specificity analyses conducted in the present manuscript. All other measures examined in the current study had at least some promising results, with the sole exception of the Addiction Severity Index (ASI), which had weak findings across all analyses. Moreover, some non-consumption measures (e.g., Drinker Inventory of Consequences, Obsessive-Compulsive Drinking Scale) had baseline to post-treatment effect sizes as large as some consumption-based outcome effect sizes. The results of the present study have identified gold standard measures for assessing mental health and quality of life. Future research should use the BSI, BDI, and WHOQOL-BREF to examine clinically-relevant changes beyond consumption outcomes. The present findings also indicate that consumption measures may not be needed to serve as surrogates for these clinically relevant constructs. These findings represent the possibility of a paradigm shift in the field of AUD treatment research evaluation to incorporate non-consumption outcomes.

Degree Name


Level of Degree


Department Name


First Committee Member (Chair)

Katie Witkiewitz

Second Committee Member

Theresa B. Moyers

Third Committee Member

Matthew R. Pearson

Fourth Committee Member

Dennis M. Donovan


Alcohol Use Disorder, Treatment Outcomes, Harm Reduction, Quality of Life, Measurement Invariance, Psychometric Properties

Document Type



Dissertation was defended May, 2017. Because the PhD for Clinical Psychology is awarded upon completion of a 1-year internship, the graduation date is 8/3/18, when internship ends.