We will draw on the management information system of a national behavioral managed care organization (MCC Behavioral Care) to assess the cost offsets of alcoholism treatment, the impacts of that treatment on a variety of functional outcomes, and the contributions of both patient characteristics and treatment program characteristics to outcomes. Data will be collected on the costs and functional outcomes of some 2100 cases at baseline, 6-month follow-up, and 12-month follow-up. We will analyze the contribution of patient characteristics in explaining treatment outcomes (specifically examining the effects of discontinued treatment as part of the analysis of treatment effectiveness). Using data taken from administrative records, we will analyze the contribution of provider characteristics (e.g., training, experience, alcohol/substance abuse history) and functional treatment outcomes, (i.e., outcomes other than costs). Using the full data set (with baseline measures and utilization data), we will calculate the medical and alcohol abuse costs of alcohol abuse treatment and the costs of all medical treatment by analyzing MCC Behavioral Care and its parent CIGNA claims data. We will estimate the relationship between various measures of patient characteristics, including baseline functional status, and each of the outcome measures, using various forms of regression. These regressions will use the logistical form when the outcome variable is expressed as a dichotomous variable and OLS (ordinary least squares) for continuous variables. We will use these analyses as a base from which to test the explanatory power of various treatment variables by adding them to the predictive equations in different combinations. Finally, we will test the potential for interactions between client characteristics and treatments. The proposed cost analysis will examine three sets of use and expenditure variables: (1) the use/expenditures for all alcohol/substance abuse services; (2) the use/expenditure of all alcohol/substance abuse related services (including those in (1)); and (3) the use/expenditure of all services. We will examine the last two to see if alcohol/substance abuse treatment of a particular type is partially or fully offset by changing medical treatment patterns. We distinguish between alcohol/substance abuse related costs and total health care costs because the causal linkages to all health care use are weaker and because of a concern over increased noisiness in the aggregate numbers. As a result, we expect less precision in category (3) (all health care expenses) than we do for either (1) (alcohol/substance abuse treatment) or (2) (use/expenditure of alcohol/substance abuse related services).
|Wall, Melanie M; Dai, Yu; Eberly, Lynn E (2005) GEE estimation of a misspecified time-varying covariate: an example with the effect of alcoholism treatment on medical utilization. Stat Med 24:925-39|
|Kane, Robert L; Wall, Melanie; Potthoff, Sandra et al. (2004) The effect of alcoholism treatment on medical care use. Med Care 42:395-402|
|Kane, Robert L; Wall, Melanie; Potthoff, Sandra et al. (2004) Isolating the effect of alcoholism treatment on medical care use. J Stud Alcohol 65:758-65|
|Wall, Melanie M; Stromberg, Kurt D; Pothoff, Sandra et al. (2004) Alcoholism treatment episodes validly defined using mental health care utilization records. J Clin Epidemiol 57:373-80|