COMBINE Manuscript Proposal Form

For Administrative Use Only
MS # B15 / Date Received: 3/1/04 / Date Reviewed: 3/16/04
Priority Set: / Scheduled for Review: 3/16/04 / Action: Approved

Title: The Obsessive Compulsive Drinking Scale among a Large Alcohol Dependent Sample

Proposed Writing Group Members: Roger Weiss, Raymond Anton, Andrew Saxon, Alan Shields

Proposal: This paper will examine the psychometric properties of scores produced by the Obsessive Compulsive Drinking Scale among a diverse and large sample of alcohol dependent patients at baseline in the NIAAA COMBINE study. More specifically, reliability will be evaluated via estimates of internal consistency; concurrent, incremental, and predictive validity will be evaluated by relating OCDS total and factor scores to variables representing other important aspects of the alcohol dependence process. Both reliability and validity data will be subjected to sub-group analyses to determine if sociodemographic (e.g., sex) or clinical (e.g., increased levels of psychiatric distress) patient characteristics predict OCDS score variability. These results can be used to a) inform future outcome analyses and b) provide insight for clinicians and researchers as to the proper use of the instrument. Additionally, understanding sample characteristics that predict pre-treatment craving can guide treatment selection and delivery.

Introduction: The experience of craving has long been recognized as important among individuals with alcohol use disorders (Jellinek et al., 1955). Nevertheless, due to its multi-factorial nature and subjective nature, it has been a difficult concept to define and only recently have behavioral scientists taken more seriously the complex task of developing and validating self-report instruments explicitly designed to assess the experience of craving (cf., Drobes & Thomas, 1999). Introduced almost 10 years ago, the Obsessive Compulsive Drinking Scale (OCDS; Anton, Moak, & Latham, 1995) has emerged as a common craving assessment tool in both research and treatment settings.

The OCDS is a 14-item self-report questionnaire that can be completed in about 5 minutes. Items are scored on a 0 – 4 Likert scale, lower scores representing no craving and higher scores representing increased craving for alcohol, and summing all items will produce a Total Score. Like the scale from which it was adapted (the Yale-Brown Obsessive Compulsive Scale; Goodman et al., 1989), the OCDS conceptualizes craving as consisting of two latent factors. First, the obsessive factor assesses the rate and intensity at which thoughts about drinking alcohol intrude in the respondent’s daily life. Second, the compulsive factor assesses the extent to which loss of control over drinking interferes with the respondent’s daily life.

Evidence suggests that the OCDS is capable of producing reliable and valid scores among a variety of alcohol treatment-seeking samples, can track change in the phenomenon of alcohol craving, can distinguish important treatment groups, and predict treatment outcomes. The NIAAA COMBINE study, because of its large and diverse sample, provides a first-rate environment in which this instrument can be examined more thoroughly.

Hypotheses: This is not a hypothesis-driven paper. Rather, it will be a description and evaluation of the performance of OCDS scores produced among a large alcohol dependent sample.

Approach / Data Analysis: This paper will address the sample dependent nature of OCDS score reliability and validity and provide a descriptive analysis of these properties among a large alcohol dependent sample. Results will be presented in the context guiding future use of the instrument both as research instrument in clinical trials and as a clinical tool for making important treatment decisions. Score reliability will be evaluated via estimates of internal consistency (Cronbach’s α) and score validity, concurrent, incremental, and predictive, can be examined via correlational analyses (e.g., Pearson’s r and multiple regression and correlation procedures). We will, for example, correlate OCDS scores with length of time since the last drink. Subgroup analyses will involve examination of individuals with very high and very low craving scores. Different dimensions of the OCDS, e.g., the degree of resistance to thoughts about drinking, will be correlated with number of abstinent days in the past 90 days, degree of AA attendance, treatment history, and motivation. Finally, OCDS scores will be correlated with negative consequences of drinking, i.e., DRINC scores.