Page images
PDF
EPUB

treatment compared with the treatment group.

The 1991-92 cohort of treatment completers produced cost savings of $83,147,187 for the two and a half years following treatment. The cost for treating all adults in 1991-92 was $14,879,128. Thus, every tax dollar spent on treatment produced $5.60 in avoided costs to the taxpayer. This is most conservative for the following reasons:

No unemployment cost savings are included.

We can assume some benefit accrued to those clients treated for weeks and/or months but who did not complete treatment. These savings are not included in this study.

There are other potential cost avoidances not included in this study, e.g., federal and local prison costs saved, institutional costs avoided,

intoxicated driver costs avoided, business losses avoided, healthy rather than drug-affected babies born, etc.

The accrual of positive societal outcomes resulting from alcohol and drug treatment were found to be significant for a period of at least three years.

In the summer of 1994, the Office of Alcohol and Drug Abuse Programs and the Governor's Council on Alcohol and Drug Abuse Programs requested an independent study of treatment outcomes for drug and/or alcohol treatment clients in publicly funded residential, outpatient, and methadone settings. Included in that request was a desire to augment the research with an assessment of the savings that might accrue (or cost that would be avoided) to Oregon taxpaying citizens from any positive outcomes of treatment. Because this assessment would be limited to the avoided costs that are measurable using existing Oregon state agency databases, it would be only a conservative estimation.

Over the past 30 years there have been a number of studies involving economic analyses of the benefits and costs of drug and/or alcohol treatment.1 The usefulness of their results has often been weakened by limitations in their methodologies. These limitations include the following:

No comparison or control group

Failure to use a representative sampling design in selecting subjects

Exclusive use of self-reported data

Brief observation periods (usually focused on the time just before or just after treatment—not necessarily representative periods)

Use of limited populations (e.g., enrollees in an HMO)

Costs and benefits assessed only in a limited number of areas

Several studies of the costs and benefits of alcohol and drug abuse treatment have recently been conducted and have received national attention. The most notable include a national study by the Center for Substance Abuse Prevention-Costs of Alcohol-Connected Crime (1995) and the CALDATA study (1994).2 The CSAP study is useful in that it provides some cost estimates of alcohol-related crime specific for Oregon; however, it is not a study of the outcomes of treatment. The CALDATA study is one of the largest studies ever attempted that combines a study of treatment outcomes with an estimate of the cost savings of treatment. The researchers selected a

1For a thorough examination of all but the most recent research, the reader is referred to Socioeconomic Evaluations of Addiction Treatment prepared by the Center of Alcohol Studies at Rutgers University.

2CSAP Prevention Monograph Costs of Alcohol-Connected Violent Crime, 1995; Dean Gerstein, et. al. "Evaluating Recovery Services: The California Drug and Alcohol Treatment Assessment," Report to State of California Department of Alcohol and Drug Programs, (April 1994), 63.

representative sample of discharged clients from substance abuse treatment programs (outpatient, residential, and methadone) statewide. They located 1826 clients who agreed to complete a retrospective interview that included questions about criminal and medical issues, substance use, income, and other concems for the period 12 months before and after treatment. The study received national attention with its conclusion that for every dollar spent in treatment, seven dollars were saved in avoided costs to society. The study has had its critics who have accurately noted its methodological limitations. These include the following:

:

The study relied primarily on retrospective self-reported data from clients who were required to remember the occurrence of events up to 36 months after they occurred.

The observation time period was very brief. The period of 12 months before and after treatment is not necessarily representative of the before treatment and after treatment periods.

The study used a pre-post design and had no comparison group.

The subjects of the study were clients who had been discharged, not
necessarily clients who had completed treatment.

This Oregon study has been designed to avoid (where possible) the methodological problems that have plagued previous studies. To this end the research design has been created with the following characteristics:

:

Representative sampling of treatment completers

Development of a matched comparison group using clients who enrolled in treatment but left before receiving an appreciable amount of treatment

Use of existing Oregon state agency databases rather than self-reported data A time period of two years prior and three years subsequent to treatment completion

METHOD

Using a quasi-experimental design, groups of clients who completed treatment were compared with groups of clients who had enrolled in treatment programs, but who terminated after receiving only minimal services. The sample was drawn from the 1991-1992 fiscal year in order to have up to three years of post-treatment outcome data. Using the Client Process Monitoring System (CPMS) database3, a representative random sample of clients for each service element (outpatient,

3 This database is the Oregon management information system for alcohol and/or drug treatment programs receiving public funds. Programs must report on clients at intake and at termination.

residential, and methadone4) was selected. A comparison group of those who began treatment but did not follow through in keeping appointments was randomly selected and matched to the treatment completers so that no differences existed between the groups on age, gender, race, drug type, and severity of drug abuse. Based on a power analysis of needed sample size, a target of 250 treatment and 250 comparison clients was set for each module, outpatient and residential. A total of 1267 clients was originally selected for the study sample.5

Existing state databases were used to collect outcome data for these clients from the periods prior and subsequent to their treatment episodes. The databases included the following: CPMS (Client Process Monitoring System), LEDS (Law Enforcement Data System), OPS (Offender Profile System),7 AFS (Adult and Family Services),8 OMAP (Office of Medical Assistance Programs [Medicaid]),9 and CSD (Children's Services Division). Permission to access these databases was gained and confidentiality of clients was protected at all times.10

4 Because of its small size, the entire population of methadone treatment completers was selected.

5There was, however, some inevitable attrition. For some individuals, insufficient identifiers were available to locate them in the database (or the identifiers were incorrect). Some individuals in the sample were deceased; others had moved. This reduced the useable sample to 1125. Finally, with the residential module, some of the non-completers were found to have had considerable treatment experience. For some analyses, they and their matches among the treatment completers were excluded. This attrition did not significantly diminish the power of the subsequent data analysis.

6 The statewide arrest database.

7 Oregon Department of Corrections database.

8 The statewide database containing public assistance payments which include welfare, food stamps, emergency assistance, etc.

9 The statewide database containing medical assistance (medicaid) payments.

10Access to the full Employment Division database was denied; however, some employment data were gathered in an earlier data collection from AFS case files. The employment data in this report are therefore limited to those individuals with AFS case files. These data were not used in the avoided cost analysis.

RESULTS

ARRESTS AND CONVICTIONS

PERCENTAGE OF CLIENTS WHO HAVE ARRESTS AND CONVICTIONS (Total-all service modules)

Treatment completers had fewer arrests and convictions in the three year period following treatment than did the non-completers. This is illustrated in the table below.

TABLE 1

Percentage of clients with subsequent arrests and/or convictions11

[blocks in formation]

There are no statistically significant differences in the arrest and conviction histories between treatment completers and non-completers prior to treatment.

Nearly half of those who completed treatment who had prior arrest records were arrest free in the three years subsequent to treatment. In comparison, only a third of treatment non-completers with prior arrest records were arrest free in the subsequent period.

11 Chi-square tests, p<.05.

« PreviousContinue »