The aim of this study is to investigate the health care utilization and costs, and the medical and psychiatric conditions, for spouses and children of persons admitted to a chemical dependency (CD) program. If family members of CD treatment patients have higher costs and utilization than persons from a general membership sample of families, and if these are affected by CD treatment, this would have implications for the cost-effectiveness of treatment. Few studies have looked at costs and utilization of family members of persons with alcohol or drug problems, and fewer still have been able to assess the effect on them of treatment. We propose to study 3011 family members of 1241 patients who were admitted to treatment at a Northern California Kaiser Permanente (KP) CD recovery program. Extensive interview data, including 5 year treatment outcomes, have been collected on the CD sample. Using KP electronic databases we will collect 6 years of cost, utilization and diagnostic-related data for the family members, as well as for members of a large comparison group of families randomly selected from the general health plan membership. We will investigate whether the spouses and children of the CD patients have higher costs and utilization (and different types of utilization such as inpatient and ER) than the comparison group, after adjusting for age, gender, family size and other individual and family characteristics, and whether treatment outcome affects the cost differences. We will also investigate whether family members of CD patients have heightened medical and psychiatric conditions compared to the comparison group. In investigating the role of treatment outcome on family utilization, we use a model that includes characteristics of: the treatment sample (e.g., demographic and problem characteristics), the family environment (e.g., family stressors and family composition), and treatment (e.g., treatment outcome and readmissions) in their relationship to family members' utilization. We will use linear and non-linear mixed models to investigate cost and utilization differences. These models appropriately handle repeated measures and intra-family correlation, while adjusting for important confounders. Our large sample and wealth of detailed self-report and automated data over a long period of time will allow us to answer important epidemiologic and health policy questions regarding the effect on utilization of having family members with alcohol and drug problems, the medical needs of family members, and the effect on family members of the treatment of those problems.
