Impact of Substance Abuse Risk, Age, Gender, and Comorbidity Status on Behavioral Health Treatment Utilization and Cost

Abstract

In the U.S., criteria for a substance abuse or dependence diagnosis were met by approximately 22.2 million in 2008, while managed healthcare covered 176.3 million Americans. This study examined how risk for substance abuse disorders interacted with age, gender, and mental health and medical comorbidity status to affect behavioral health treatment utilization and cost. Individual impacts include risks to physical and psychological well-being. The disruptive behaviors of substance use disorders influence families through both social impairment and an inability to maintain role functions, leading to discord within couples and an increased risk of cognitive, emotional, and behavioral problems in offspring. Substance use disorders affect businesses through absenteeism, reduced productivity, and termination and new-hire costs. Local, state, and federal governments all attempt to minimize the impact of substance use disorders on communities, which significantly affects the workload and fiscal solvency of those governments through public welfare and law enforcement programs. Treatment of substance use disorders has been shown to effectively mitigate the immediate impacts on individuals, families, and businesses; yet, the impact on governments has not been explored, though they may be inferred to receive some benefit through the same mechanisms by which they are negatively influenced. The study sample consisted of 13,417 adult members of a managed behavioral healthcare company who went to at least one outpatient mental health session and reported a significant level of psychological distress on a self-report measure administered at that session. Substance use risk, age, gender, comorbidity status, treatment utilization, and treatment cost for a four-month period were examined, based on data from that assessment and claims data. Correlations initially ruled out any significant effect of psychological distress on the other predictor variables, indicated that Facility and Other treatment types were highly interrelated, and set the foundation for later multiple regressions. Factor analysis then determined that Facility and Other treatment types were part of a single factor for both treatment utilization and cost; so, these criterion variables were then combined. Multiple regressions then examined the relationships between multiple predictors on each criterion. Results indicated significant relationships for most analyses which was expected given the large sample size, so Cohen's d was used to discuss the relative strength of the effects. No relatively weak results were found between all of the predictor and criterion variables. Suggested causes for this pattern of outcomes included methodological flaws in operationalization or minimal relationship between the constructs of interest and the assessment questions aimed at evaluating those constructs. Future research is discussed in terms of improving methodological problems of the current study.