This study provides firm experimental evidence that employment-based abstinence reinforcement can increase cocaine abstinence. All participants were offered paid employment in a model work setting. Participants who were randomly assigned to a condition in which daily access to the workplace was made contingent on providing urinalysis evidence of recent cocaine abstinence achieved significantly higher rates of cocaine abstinence than did those who were allowed to work independent of their urinalysis results. The study has important implications for four major public health domains: drug addiction treatment, employment programs for chronically unemployed and drug-addicted individuals, HIV risk prevention, and workplace practices.
As a drug addiction treatment, this study shows that employment-based abstinence reinforcement can be an effective intervention, even for extremely persistent drug users who fail to respond to conventional treatment approaches. The study was conducted in individuals who persisted in high rates of cocaine use despite participation in community methadone treatment. During the 4 weeks in the workplace prior to random assignment, about 6% of the urine samples provided by both groups were negative for cocaine (, and ), and 75% of participants in both groups failed to provide even one cocaine-negative urine sample (). Given that a high rate of cocaine use is a robust predictor of poor treatment outcome (Preston et al., 1998
; Silverman et al., 1996
), and given the limited number of treatments that have shown efficacy in treating cocaine abuse in methadone patients (Silverman et al., 1998
), the demonstrated effects on cocaine use in this population are particularly noteworthy.
The study has direct implications for employment programs for chronically unemployed adults with long histories of drug addiction. A number of employment programs for chronically unemployed adults have provided intensive education, job-skills training, and supported work (Dickinson & Maynard, 1981
; Drebing et al., 2005
; Kashner et al., 2002
; Magura et al., 2004
; Milby et al., 1996
); some of those programs have provided stipends for participation in training and supported work (Dickinson & Maynard; Drebing et al.; Kashner et al.; Milby et al.); one of the programs provided additional payments for abstinence and for meeting program objectives (Drebing et al.); and some programs have required participants to maintain abstinence from drugs to continue to work and earn wages (Kashner et al.; Milby et al.). One of the largest and most recognized of those programs is the U.S. Department of Veterans Affairs' Compensated Work Therapy (CWT) program (Kashner et al.). A recent randomized study evaluated the program in homeless veterans with substance abuse problems and showed that compared to individuals who did not participate in the program, CWT participants experienced significant reductions in drug-related problems, homelessness, and incarcerations (Kashner et al.). The CWT program arranged contingencies for abstinence; however, the nature of the contingencies and the consequences for drug use were guided by clinical judgment and were not well defined. Furthermore, the study did not show direct effects of the intervention on objective measures of drug use, or isolate which features of the CWT intervention were critical in improving the drug-related outcomes. For the large numbers of drug-addicted and chronically unemployed individuals who participate in employment programs, the current study suggests that paid training and supported employment programs could be used to simultaneously train participants and promote drug abstinence if daily access to the training program were made contingent on verified abstinence. Further, the current study provides precise guidelines as to how those contingencies can be effectively arranged and provides clear evidence that arranging abstinence-contingent access to paid training can substantially increase abstinence rates above rates observed under conditions in which paid training and supported employment are offered without such contingencies.
The study suggests that employment-based abstinence reinforcement may be effective in reducing the spread of HIV infection. All participants in this study were injection drug users, over 20% reported being HIV positive (), and over 20% reported sharing injection equipment, going to shooting galleries or crack houses, and trading sex for drugs or money (). In this population, employment-based abstinence reinforcement increased urinalysis-verified cocaine abstinence ( and and and ) as well as the proportion of participants who self-reported abstaining from injection drug and crack use (), behaviors that have been associated with increases in risk of HIV infection. The study did not show significant effects on other HIV risk behaviors (e.g., trading sex for money or drugs), possibly due to the relatively small sample size.
The study provides a model for drug-addiction treatment in the workplace. Through the growth of employee assistance programs, the workplace has increasingly become recognized as an important context in which to detect and treat drug addiction (Hartwell et al., 1996
; Office of Applied Studies, 2002
). It is important to note that urinalysis-testing programs, including random testing, are already used in many workplaces (French, Roebuck, & Kebreau Alexandre, 2004
). Although those programs are mainly used as employment screening tools to eliminate people who abuse drugs from workplaces, this study suggests that parameters of those testing programs could be altered to achieve therapeutic benefits for drug-addicted employees. The integration of employment-based abstinence reinforcement contingencies into community workplaces will not be simple, and systematic research will be needed to determine how such integration might be accomplished. The current study suggests that employees who have persistent substance-abuse problems could be offered to participate in an employment-based abstinence reinforcement program as an alternative to termination or another option that might be less desirable to the employee. Under such a program, the employee could be required to provide urine samples under observation on a routine basis. If the employee provides a positive urine sample or fails to provide a scheduled sample, the employee would not be allowed to work that day or any day thereafter until he or she provides a drug-negative urine sample. The employee would also experience a temporary decease in pay.
Imposing contingencies on abstinence in this study had the undesirable effect of reducing attendance in the workplace. Future research should investigate methods to minimize this undesirable effect of the abstinence contingency. Nevertheless, almost all participants exposed to the abstinence requirement continued attending the workplace throughout the intervention period, at least intermittently (); participants in the two groups achieved comparable skill levels on the training programs; and participants exposed to the abstinence contingencies still achieved approximately a three-fold increase in cocaine abstinence rates. Thus, the training objectives of the program were achieved for both groups while substantially increasing abstinence in the group exposed to the abstinence reinforcement contingencies.
One limitation of the study is that participants in the abstinence-and-work group provided fewer mandatory Monday, Wednesday, and Friday urine samples than did work-only participants ( and ), which makes comparison of the two groups more difficult. Different methods of replacing the missing data were used, and all methods generally showed similar outcomes.
Although the employment-based abstinence reinforcement contingency significantly increased rates of cocaine abstinence in this study, many participants did not initiate cocaine abstinence. This is not completely surprising, because they had extremely high baselines rates of cocaine use. As described above, individuals with the highest baseline rates of cocaine use have historically been the least likely to respond to abstinence reinforcement interventions (Preston et al., 1998
; Silverman et al., 1996
). Prior research has shown that abstinence can be promoted in treatment-resistant patients by manipulating reinforcement parameters such as magnitude (e.g., Silverman et al., 1999
), and similar investigations will be required using employment-based abstinence reinforcement. The Web-based therapeutic workplace intervention should provide an efficient and highly controlled context to explore the utility of potential methods to improve treatment outcomes (Silverman, 2004
Heroin use was not targeted in the abstinence reinforcement contingencies in this study, and many participants continued to use heroin. Prior research has shown abstinence from both heroin and cocaine can be achieved through abstinence reinforcement contingencies (Silverman et al., 2004
). Future studies will need to develop procedures to produce abstinence from multiple drugs using employment-based abstinence reinforcement contingencies.
There were no differences in cocaine abstinence outcomes at the single-point follow-up assessment conducted 6 months after the end of treatment (). Cocaine abstinence in the abstinence-and-work group at follow-up appeared to be slightly lower than the cocaine abstinence levels observed during treatment; however, cocaine abstinence in the work-only group appeared to increase. Although we do not fully understand why we obtained this pattern of results, some observations and speculation might be useful. Careful review of the data shows that some of the data seem orderly and somewhat predictable, and some of the data are less easy to understand. As might be expected based on posttreatment outcomes from prior related research (Silverman et al., 1996
), a few participants in the abstinence-and-work group who achieved substantial amounts of cocaine abstinence during treatment (S50, S55, and S56; ) appeared to be abstinent after treatment and provided a cocaine-negative sample at the 6-month follow-up (data not shown). Also consistent with prior research, several other abstinence-and-work participants who achieved abstinence during treatment (S47, S51, S52, S53, and S54; ) appeared to return to cocaine use and provided a cocaine-positive sample at the 6-month follow-up (data not shown). As might also be expected, many participants in the work-only control group who provided relatively frequent cocaine-negative samples during treatment (S23, S24, S25, S26, S27, and S28; ), presumably representing their usual patterns of cocaine use, also provided a cocaine-negative urine sample at the 6-month follow-up (data not shown). Somewhat less easy to understand is the fact that 3 work-only participants who never provided a single cocaine-negative sample throughout the 7 months of participation (S5, S9, and S11; ) provided a cocaine-negative urine sample at the 6-month follow-up (data not shown). This study was designed primarily to study the effects of the interventions on cocaine use during treatment, and therefore assessed patterns of cocaine use thoroughly during treatment through frequent and repeated urine collection and testing. It is possible that collection of a single urine sample 6 months after treatment did not fully or accurately capture the patterns of cocaine use and abstinence in the two groups after treatment ended. Alternatively, the 6-month follow-up results might accurately reflect the patterns of cocaine use in the two groups after treatment. If so, those results might reflect changes in life circumstances for some of the individuals (e.g., S5, S9, and S11) that we do not fully understand but that have little to do with the treatment assignments. Whatever the case, it will be important for future research to investigate posttreatment abstinence outcomes more fully, possibly by obtaining more frequent measures of cocaine abstinence during follow-up.
Overall, this study shows that employment-based abstinence reinforcement contingencies can be effective in promoting abstinence from cocaine in a population of treatment-resistant methadone patients who are at considerable risk for spreading or contracting HIV due to their continued high rates of injection and crack cocaine use and associated HIV risk behaviors. Use of employment-based abstinence reinforcement contingencies for this and other drug-addicted populations could be useful in addressing the critical and often intractable public health problem of drug addiction.