|
Article Excerpt
SUBSTANCE USE-DISORDERS OUTCOMES affect a wide range of domains, including productivity, wages, fulfillment of societal responsibilities (e.g., as parent, partner, or spouse), social, criminal justice, mental and physical health, and health care utilization (Farahati et al., 2003; French and Zarkin, 1992; Mullahy and Sindelar, 1989; Rice et al., 1991). There is also substantial evidence suggesting that substance use-disorder treatments can effectively decrease substance use, individual costs, and social costs and improve health status (Holder, 1998; Kraemer et al., 2002; Volk et al., 1997). From a health policy perspective, next steps include understanding the relative efficiency (or "bang for the buck") of investing in substance use-disorder treatments compared with other mental and physical health interventions.
Previous studies have evaluated substance use-disorder treatments according to two broad outcomes: effectiveness and cost. Effectiveness refers to the effect of substance use-disorder treatment on symptoms, health status, and quality of life. Several studies have shown that substance use-disorders treatment improves substance use-disorders symptoms, health status, and quality of life using the Addiction Severity Index (Cisler and Zweben, 1999; McLellan, 1994), Medical Outcomes Study Short Form (SF)-36 (Kraemer et al., 2002; Romeis et al., 1999), other quality-of-life measures (Cisler and Zweben, 1999; Foster et al., 2000), and other study-specific measures (Kairouz and Dube, 2000).
Costs of substance use-disorder interventions are usually defined as direct costs (health care resources, non-health care resources, informal caregiver time, patient time) and indirect costs (changes in work performance, earnings, drug traffic control, crime prevention, and cost to social welfare system) (Cartwright, 1999). Several studies have shown that substance use-disorder treatment benefits may actually exceed treatment costs (cost/benefit ratio <1), even without including many indirect costs (Cartwright, 2000; Fletcher and Battjes, 1999; French et al., 2000; Holder, 1987, 1998). A gap in the current research literature is the lack of outcomes studies combining cost with quantity and quality of life from the patient perspective (Godfrey and Parrott, 2000).
The primary objective of this study is to empirically investigate the relationship between outcomes specific to substance use-disorder treatment and preference-weighted health status based on the visual analog scale (VAS) and standard gamble (SG) conversion formulas for the SF-36 score (Brazier et al., 1998, 2002). The SF-36 is a commonly used measure of general health status (Ware and Sherbourne, 1992). The VAS and SG conversion formulas combine quantity of life and health status using SF-36 data to produce an effectiveness term called a quality-adjusted life year (QALY). Because these QALYs are based on a general health status measure, they can be compared across disorders. Cost-per-QALY ratios for health care interventions provide a common yardstick by which the cost-effectiveness of interventions for different illnesses can be directly compared.
To our knowledge, there have been no studies that have estimated the cost-effectiveness of a substance use-disorders intervention using general QALYs as the measure of effectiveness. To date, the substance use-disorder interventions that reported cost-per-QALY ratios have used QALYs based on HIV disease for an opiate-addiction intervention (Barnett and Hui, 2000; Barnett et al., 2001) or based on an unvalidated general QALY conversion formula for smoking interventions (Cromwell et al., 1997). One advantage of using a general QALY measure in a cost-per-QALY analysis is that the cost-per-general QALY ratios can be directly compared across a wide variety of mental health and physical health interventions, and these ratios can be used to identify the most efficient use of limited health care resources (Gold et al., 1996).
The focus of this study is methodological in nature. For cost-per-QALY analyses to be useful in substance use-disorders treatment research, the measurement of QALYs must be correlated with outcomes that are meaningful to substance use-disorders treatment. Therefore, the primary aim of this study is to examine the construct validity of two methods for converting SF-36 outcomes into QALYs in the context of substance use-disorders treatment. The secondary aim is to present examples of simulated cost-per-QALY ratios using the QALY estimates from our data and common cost-effectiveness thresholds to determine a range of hypothetical costs that could be associated with substance use-disorders treatment.
Method
Design
The design of the parent study was observational and included subjects that initiated outpatient treatment for substance-use disorders. Baseline data were collected via inperson interview, and 6-month follow-up data were collected via mail-out survey. There were a wide variety of times for returning the 6-month survey, so we defined an eligible 6-month follow-up as 2 weeks before and 4 weeks after 6 months from the baseline survey (a 6-week window) to be more consistent with intervention trials. Substance use-disorder treatment services were provided by a regional managed behavioral-health care organization. In 1997, the 24 highest-volume substance use-disorders treatment providers in this region, which delivered care for 80% of the patients, were invited to participate in this study. The parent study aimed to generate data that could be used to guide clinicians in tailoring treatments and identifying effective and efficient treatment approaches (Smith et al., 1997). Nine treatment centers were eliminated from the original study due to practical barriers such as bankruptcies, closures, reluctance to participate, or administrative changes. The remaining 15 centers participated in the evaluation.
Subjects
For the main analysis, we used data from 165 outpatients who met baseline criteria for substance dependence within the past year, initiated substance use-disorder treatment, returned the 6-month survey within the 6-week window, and had sufficient SF-36 data to calculate preference-weighted health status scores. Additional subject-exclusion criteria included substance-induced cognitive impairment (e.g., intoxication or psychosis) at the time of the baseline interview or inability to speak English. There were 883 subjects who were eligible, completed the baseline interview, and met baseline criteria for substance dependence. Only 185 subjects remained after applying the 6-week window for completing the 6-month follow-up survey. Of these, 20 subjects did not have sufficient SF-36 data to calculate VAS and SG scores, leaving a main analysis sample size of 165. There were no statistical differences between those who completed the 6-month survey within the 6-week window and those who did not on the basis of age, gender, ethnicity, marital status, education, income, number of drinks per day, drinking-quantity category (heavy, moderate, abstinent), or baseline VAS or SG scores. The only baseline difference between those eligible and not eligible for the analysis was that those who were eligible were less likely to...
|
