Association of outpatient alcohol and drug treatment with health care utilization and cost: Revisiting the offset hypothesis

Objective: This study examines the hypothesis that treat ment reduces medic al utilization and costs of patients with substance use problems. Method: A dult patients (N = 1,011; 67% men) entering the outpatient chemical depende ncy recovery program at Sacramento Kaiser Permanente over a 2-year period w ere recruited into the study. Medical utilization and costs were examined f or 18 months prior and 18 months after intake. To account for overall chang es in utilization and cost, an age, gender and length-of-enrollment matched nonpatient control group (N = 4,925) was selected from health-plan members living in the same service area. Multivariate analyses controlling for age acid gender were conducted using generalized estimating equation methods, allowing for correlation between repeated measures and nonnormal distributi ons of the outcome variable. Results: The treatment cohort was less likely to be hospitalized (odds ratio [OR] = 0.59; p < .01) and there was a trend for having spent fewer days (rate ratio [RR] = 0.77; p < .10) in the hospit al in the posttreatment period compared to pretreatment period. These patie nts were also less likely to visit the emergency room (ER) (OR = 0.64; p < .01)and had fewer ER visits (RR = 0.81; p < .01) following treatment. Inpat ient, ER and total medical costs declined by 35%, 39% and 26%, respectively (p < .01). Reductions in cost were greater for the treatment cohort when c ompared with the matched sample (p < .05). Among women, there were signific ant reductions (p < .05) in inpatient, ER and total costs for the study coh ort when compared with the matched sample; among men, the reductions in inp atient and ER cost (but not total cost) were significantly larger (p < .05) for the study cohort when compared with the matched sample. For the treatm ent cohort, the change in medical cost was not significantly different by g ender. Changes in cost were significantly different across the various age groups (p < .05) for the study cohort and the matched sample. Among those i n the group aged 40-49 years, the decline in cost fur study cohort was sign ificantly larger (p < .05) than for the matched sample. Conclusions: For pa tients with substance use disorders entering treatment, there was a substan tial decline in inappropriate utilization and cost (hospital and ER) in the posttreatment period. The disaggregated pattern of posttreatment decline i n utilization and cost is suggestive of longterm reductions that warrant a longer follow-up.