Do Treatment Services for Drug Users in Outpatient Drug-Free Treatment Programs Affect Employment and Crime?

Abstract

Using the 1992–1995 National Treatment Improvement Evaluation Study data, we employ principal components analysis to create treatment service factors based on both patient self-reports and treatment record extracts. We included these factors in multivariate models for 1,136 outpatient drug-free treatment patients to estimate the relationship between services and post-treatment employment and crime. Although our models indicated some significant predictors of employment and crime, the overall effects of services were generally insignificant. We conclude that either services are unrelated to outcomes or if they are related then we are not measuring the key aspects of service provision that may be important. The study's limitations are noted.

Keywords:

• crime
• employment
• substance user treatment
• treatment services

Notes

Notes

*The journal's style utilizes the category substance abuse as a diagnostic category. Substances are used or misused; living organisms are and can beabused. Editor's note.

^†Treatment can be briefly and usefully defined as a planned, goal-directed change process, which is bounded (culture, place, time, etc.) and can be categorized into professional-based, tradition-based, mutual help–based (AA, NA, etc.) and self-help (“natural recovery”) models. There are no unique models or techniques used with substance users—of whatever types—that aren't also used with non-substance users. Editor's note.

^‡The reader is reminded that programs treating the broad range of heterogeneous substance users consistently do not assess the patient's use of tobacco products, their patterns of eating (for eating disorders), and rarely if ever, include these behaviors in treatment planning, implementation, and assessment (process and outcomes). Editor's note.

¹ Statistical tests of mean differences in baseline variables between those outpatient drug-free patients who completed all three interviews and those who did not complete all interviews revealed no statistical differences on most variables. Those patients who did not complete all interviews were slightly more likely to be male (74% versus 70%) and slightly less likely to have had the intake interview in a prison setting (7% versus 5%) compared to those patients who completed all three interviews.

² Although we rotated to the simplest structure, we still had multiple loading for several variables. We decided to allow the variables to define more than one factor and accept the naturally occurring intercorrelation among the factors.

³ Due to the skewness of the crime and hours data, we estimated two specifications of each model—with and without logged crime (and hours) variables as dependent variables. The results differed little between the two specifications, and we present the unlogged results for ease of interpretation. Logged results are available from the corresponding author upon request.

⁴ We also ran an interactive model in which the indicator variable for treatment completion was interacted with the treatment service variables, and these models did not produce any statistically significant findings. This suggests that the effect of services on posttreatment crime and employment are not different between these 2 groups. The output for these analyses is available from the corresponding author upon request.