Factorial, Construct, and Predictive Validity of the Motivation for Treatment Scale in Alcohol-Use Disorder Withdrawal Treatment.

Introduction: The aim of the present study was to examine for the first time the factorial, construct, and predictive validity of the motivation for treatment (MfT) scale in a cohort of patients undergoing inpatient-qualified alcohol withdrawal treatment with the goal of referring patients to further treatment. The MfT scale has previously been evaluated in different settings of substance abuse treatment, revealing factorial ambiguity. To the best of our knowledge, the present study is the first study that conducted comprehensive factor analyses versus separate analyses of the factors conducted in prior studies in order to clarify the aforementioned factorial ambiguity.

Methods: A total of 249 patients (mean age 45.2 years (SD = 10.3); 34.4% females) with alcohol dependence were assessed. Data were obtained from four inpatient clinics specialized in qualified alcohol withdrawal treatment in Germany. First, confirmatory factor analyses were carried out to examine the fit of the four models discussed in the literature. Second, an exploratory factor analysis was conducted. Correlations of the new factors with other motivational constructs and referral to a subsequent treatment were investigated as measures of construct and predictive validity.

Results: None of the four models showed an acceptable fit to the data in confirmatory analyses. The exploratory analysis suggested to eliminate seven items because of inappropriate factor loadings and resulted in a shortened MfT scale, which consists of three factors based on 17 items. For the latent variables "problem recognition," "desire for help," and "treatment readiness," satisfactory composite reliability was found with 0.82, 0.80, and 0.78, respectively. Evidence for predictive validity was found in the correlation between "treatment readiness" and referral to a subsequent treatment.

Discussion/conclusion: The new shortened MfT scale exhibited remarkable parsimony, which is desirable in settings such as withdrawal treatment, where patients frequently are cognitively or physically impaired. Despite its briefness, construct and predictive validity were better than in the original version of the MfT scale. The factorial validity of the suggested scale needs to be corroborated in further research.