Clinical outcomes studies frequently provide payment incentives to research participants for completing follow-up assessments. The payments are intended to compensate participants for their time and effort and to encourage those who might otherwise be difficult to relocate to stay in contact with the research team.
Maintaining adequate follow-up rates is of particular concern in drug abuse research studies, which are often hindered by substantial attrition of research subjects prior to follow-up assessments. Dozens of drug abuse treatment outcome studies report average attrition rates ranging from 20% at six months postadmission to 35% at 12 months postadmission to 65% at 36 months postadmission. (1) A meta-analysis of 85 drug abuse prevention studies similarly reported average attrition rates ranging from 27% at 12 months postadmission to 33% at 36 months postadmission. (2) Importantly, these reviews and meta-analyses only included published studies, and it is reasonable to assume that many studies with lower follow-up rates were never published for the very reason that their follow-up rates were deemed to be unacceptable. Indeed, follow-up rates in large-scale community evaluations of drug abuse treatment programs often fall well below an acceptable threshold of 70% and have been as low as 47%. (3)
Attrition from research cannot be assumed to be random or ignorable. The likelihood that a research subject will complete a follow-up assessment is influenced by such confounding factors as the participant's demographic and drug-use characteristics, as well as the nature of the research interventions. (5) This can contribute to systematic differences in attrition rates between research conditions, which would constitute a serious threat to the internal validity of a study. For example, a larger proportion of participants in an experimental treatment arm of a study might be lost to follow-up as compared to control participants attending treatment as usual. This could have the effect of undermining random assignment by making the experimental and control conditions systematically different from each other on outcome-biasing factors. Because such confounding effects may go unmeasured or undetected, they can never be confidently ruled out, thus leaving the findings in irresolvable doubt. Moreover, it cannot be confidently determined whether the follow-up sample sufficiently represented the original baseline cohort, which presents a serious threat to the external validity or generalizability of the study. (6)
Differential attrition also raises ethical concerns about the potential for systematic exclusion of certain subgroups of clients from treatment-outcome studies, such as clients who are geographically transient, emotionally withdrawn, or socially disenfranchised. If such individuals self-select out of research studies, then new treatment-related findings and therapeutic advances might be less applicable to those very clients who may need the services the most. This process could potentially violate the ethical principle of justice, which necessitates publicly funded research programs to benefit all members of society in a fairly equivalent manner. (7)
One strategy that has proven successful in improving follow-up rates is the use of payment incentives. The literature on "contingency management" or "operant conditioning" is replete with evidence that payments can improve a wide range of behaviors among substance abusers, including increasing treatment attendance and drug abstinence, (8) as well as enhancing recontact rates in research. (9)
In a prior study, consenting research subjects recruited from drug-free, outpatient substance abuse treatment programs were randomly assigned to receive various magnitudes of payment incentives ($10, $40, or $70) in either cash …