Patient-Centered Placement Matching of Alcohol-Dependent Patients Based on a Standardized Intake Assessment : Primary Outcomes of an Exploratory Randomized Controlled Trial

Background: Placement matching guidelines are promising means to optimize patient-centered care and to match patients' treatment needs. Despite considerable research regarding placement matching approaches to optimize alcohol abuse treatment, findings are inconclusive. Objectives: To investigate whether the use of patient-centered placement matching (PCPM) guidelines is more effective in reducing heavy drinking and costs 6 months after discharge from an inpatient alcohol withdrawal treatment compared to usual referral to aftercare. Secondary aims were to investigate whether age, gender, trial site or level of care (LOC) are moderators of efficacy and whether patients who were actually referred to the recommended LOC had better treatment outcomes compared to patients who were treated under- or overmatched. Methods: Design. Exploratory randomized controlled trial with measurements during withdrawal treatment and 6 months after initial assessment. Setting. Four German psychiatric clinics offering a 7-21 day inpatient qualified withdrawal program for patients suffering from alcohol dependence. Participants. From 1,927 patients who had a primary diagnosis of alcohol dependence and did not have organized aftercare when entering withdrawal treatment, 299 were invited to participate. Of those, 250 were randomized to the intervention group (IG, n = 123) or the control group (CG, n = 127). Intervention. The PCPM were applied to patients of the IG by feeding back a recommendation to a LOC for aftercare that was calculated from the Measurements in the Addictions for Triage and Evaluation (MATE) and discussed with the staff of the treatment unit. Patients of the CG received a general feedback regarding their MATE interview on request. Measurements. The MATE, the Client Socio-Demographic and Service Receipt Inventory--European Version and the MATE-Outcomes were administered. Data were analyzed using generalized linear models. Results: In the intention-to-treat analysis, there wer