COMBINE Manuscript Analysis Request Form:
For Administrative Use Only (CC)MS #: CM13 / Date Received: 3/4/04 / Date Reviewed: 3/29/04
Priority Set: / Scheduled for Review: 3/29/04 / Action: Approved
Title Matching CBI to client profile to increase treatment effectiveness.
Identified Authors: Longabaugh, Donovan, LoCastro, Zweben
Rationale: The underlying assumption is that some clients will respond more favorably to CBI treatment than to MM, whereas others will respond at least equally well and perhaps better to MM. It is hypothesized that differential responders to CBI will be predicted from dimensions of the client’s pre-treatment profile. Two overlying predispositional factors are readiness and capacity.
Ho.1. All other things being equal, clients less ready to change their drinking (as measured by URICA) will be more likely to do so during CBI treatment than MM; whereas this difference will be less for those who are more ready to change.
Ho 2. All other things being equal, clients who have fewer psychosocial resources prior to treatment will be more likely to have enhanced quality of life and reduce their drinking during CBI treatment than during MM; whereas this differences will be less for those with greater psychosocial resources.
Inclusion / Exclusion Rules:
All patients randomized to Combine treatment will be included, as MM only patients will be included as the comparison condition.
Proposed Analysis:
Dependent Variables:
Primary: 1) PDA during weeks 1-16; 2) % of heavy drinking days during treatment (males 5 or more, females 4 or more), 3) # of days to first heavy drinking day during treatment.
Secondary: Quality of Life at the end of treatment.
Covariates:
1) Baseline level for each dependent variable.
2) Assigned Treatment Condition:
2.1. Active medication group (NTX, Acamp, Combination) vs. placebo
2.2. MM only vs. CBI &MM vs. CBI only
3) Baseline level of psychosocial resources (see below)
4) Baseline level of readiness to change (see below).
5) Baseline level of “alcohol dependence” (see below)
Independent Variable: Product Term of Treatment Assignment (MM alone vs. CBI&MM vs. CBI) X baseline matching variable (i.e., either Readiness to change or psychosocial resources; each dichotomized into high and low groups, based on a median split)
Analysis: ANCOVA’s followed by planned comparisons.
Evidence required to support the matching hypotheses:
For both hypotheses it is necessary that the interaction effect be significant, after main effects of treatment condition and baseline client variable are accounted for.
Planned comparisons must show that the magnitude of difference between CBI/MM vs. MM only be significantly greater for the low resource(motivation) group than for the high resource (motivation) group.
Planned comparisons must show that the magnitude of difference between the CBI only vs. MM only be significantly greater for the low resource (motivation) group than for the high resource (motivation) group.
For the three alcohol variables, the alpha level for the planned comparisons would be reduced to p.0125; for the Quality of life variable the alpha would be p.05.
Operational Definitions of patient alcohol dependence, psychosocial resources, and readiness to change variables
Alcohol Dependence: It is proposed that we develop a composite alcohol dependence variable derived from baseline measures that are intended to measure this construct: # of dependence symptoms, compulsive drinking, etc. Such a measure might be arrived at by conducting a PCA or Confirmatory factor analysis of all measures intended to directly measure dimensions of this construct. (Alternatively, another group working on baseline measures may provide such a measure).
Readiness to Change: Baseline descriptive data are already available on the URICA. It is proposed that we review these data and decide on a cut point based on distribution and theoretical considerations (e.g. action or more vs. preparation or less).
Psychosocial Resour ces: Psychosocial resources can be conceptually divided into psychological resources and social resources. It is proposed that significant preliminary analyses of selected baseline data be conducted to develop a composite measure for this construct.
Psychological resources will be measured by one or more instruments indexing psychological health and/or deficits (TBD).
Social resources would be measured by baseline indices of: a functional relationship with a most important significant other; non-problematic work/school role; supportive family & social networks, stable residence & adequate income (TBD). These measures will be derived from baseline demographic and IP data.
Once each psychosocial domain is indexed, they will be aggregated to yield a single overall psychosocial resource variable for testing the a priori hypothesis.
Analysis proposals should be a concise but detailed summary of the requested data analysis for an approved writing group