Financial Incentives for Therapy Completion: A Randomized Controlled Trial with U.S. Veterans

Recovery Resource Council (RRC) is one of the largest and most comprehensive non-profit mental and behavioral healthcare providers in North Texas. Accredited by the Joint Commission in Behavioral Health and licensed by the State of Texas as an Outpatient Treatment Center, RRC strives to promote wellness and recovery through a variety of services and programming. An important component of RRC programming is providing free counseling services to hundreds of U.S. veterans annually. While RRC observes great success for veterans who complete counseling, attendance can be a major obstacle.
We will evaluate the impact of RRC’s therapy incentives using a randomized-controlled trial research design. Veterans who are referred to RRC are subject to an eligibility screening via phone, during which time they must ensure that they are 18 years of age or older, hold veteran status, are deemed not suicidal by RRC staff, and will participate in individual counseling. Eligible veterans are then assigned an intake session time, during which they show any requested documentation, complete an intake form, complete an initial mental health assessment, and go through the informed consent process with RRC staff. Any participant who does not want to participate in the study will still receive all services as usual. After a participant gives consent, they are enrolled in the study. Randomization will be done through the Qualtrics platform, so that results can be communicated to veterans immediately during their intake session.
The treatment group will be offered a $500 gift card for completing 6, 12, and 18 counseling sessions; up to $1,500 in total. Because RRC’s typical practice involves prescribing and scheduling counseling in 6-session blocks, the financial awards are revealed to veterans in the treatment group in stages. After intake, treatment group individuals are told that they will receive a $500 gift card for completing the first 6 sessions; later sessions and awards are not mentioned. If the counselor decides that another 6-session block should be prescribed, then at the completion of the 5th session (i) the next block of 6 sessions is scheduled (i.e., sessions 7 through 12) and (ii) the counselor reveals that the participant can earn another $500 for completing the 12th session. The same process takes place during session 11. As part of RRC’s typical practice, no veterans are offered more than 18 sessions. The counseling received by the control group is identical to that of the treatment group, as is the scheduling of additional sessions. Other than informing individuals in the treatment group of the gift cards, counseling services will not change as a result of this study. Intake will be different than usual for all participating veterans (treatment and control) due to the informed consent process and collecting some additional personal information via tablet. We will also ask all participants two questions via tablet at the conclusion of each counseling session about the expected costs and benefits of therapy.

2024-01-02

2027-01-02

Total number of counseling sessions attended
Completion of therapy, as defined by the counselor

An indicator variable for completion of therapy will be constructed using discharge codes that RRC assigns to each patient. In particular, there is a discharge code for program completion (regardless of the number of sessions).

Completion of every possible number of counseling sessions, e.g., at least 1 session, at least 2 sessions, etc. In other words, we will plot survival curves by treatment status.

Start-of-Session Mental Health Index and sub-components

Expected cost and benefit of therapy

Structural parameters (if applicable)

Using the responses to start-of-session questions, we will create a summary standardized index of mental health at the start of each session.

Using the responses to survey questions asked at the end of each session regarding how a respondent thinks their symptoms will change after the next two sessions as well as how upsetting a respondent expects the next two sessions will be, we will create indicator variables corresponding to the categories of these expectations.

Should we find that the financial incentives increase therapy attendance, we intend to estimate the structural parameters of a dynamic model of optimal therapy exit timing. Experimental variation creates exogenous differences in the marginal benefit of a counseling session, allowing us to ‘price’ the marginal benefit of therapy to the patient, identify the deep structural parameters, and to conduct counterfactual exercises.

Study enrollment will occur starting in January 2024 and continuing through 2027. Veterans who are referred to RRC are subject to an eligibility screening via phone, during which time they must ensure that they are 18 years of age or older, hold veteran status, are deemed not suicidal by RRC staff, and will participate in individual counseling. Eligible veterans are then assigned an intake session time, during which they show any requested documentation, complete the standard RRC intake form,, complete an initial mental health assessment, and go through the informed consent process with RRC staff. Any participant who does not want to participate in the study will still receive all services as usual. After a participant gives consent, they are enrolled in the study and complete additional intake questions via the Qualtrics platform. Randomization will be done through the Qualtrics platform, so that results can be communicated to veterans immediately during their intake session. Randomization odds will be 50-50. We plan to enroll 600 individuals, of which approximately 300 individuals will be offered financial incentives for therapy.
The experimental portion of these procedures is the random assignment of individuals to be offered or to not be offered financial incentives for therapy. The treatment group will be informed that if they attend their first six counseling sessions, they will be given a $500 gift card. If it is determined that they need more sessions, they will be given the same incentive to attend 6 more sessions. The same process continues again after the completion of 12 sessions. The content of counseling sessions will not be impacted by the designation of treatment or control group. Individuals who are not offered financial incentives for therapy will have access to the normal counseling services provided by RRC staff. That is, those in the control group will receive “business as usual” services offered by the RRC, which includes the same counseling services and evaluation of how many sessions an individual is recommended to receive.
During the consent process, participants consent to be followed in administrative records. We will track all outcomes of interest as they become available from RRC’s start-of-session and end-of-session questionnaires and mental health screenings.
At the point that 300 individuals are enrolled in the study sample, we plan to estimate the treatment effect to determine the feasibility of adding a third study arm. If the treatment effect on either average number of sessions completed or the probability of completing six sessions is sufficiently large and there is support from RRC to do so, we plan to begin randomizing participating veterans into a second treatment arm which will receive a gift card of value less than $500 for each completed 6-session block (up to 18 sessions total). If this third study arm becomes a reality, power calculations will be re-evaluated to shift to an updated optimal sample size.

Not available

Qualtrics

Individual

No

600 individuals

600 individuals

300 in treatment group, 300 in control group

Using historical data from RRC, the average number of attended sessions is 7 (standard deviation of 6.5) and 46 percent of patients complete at least 6 sessions. Assuming these figures apply to control-group outcomes, with 50/50 randomization, 600 total veterans, and the probability of a type II error at 0.90, the minimum detectable effect (MDE) for total sessions attended is 1.7 sessions and the MDE for the probability of attending 6 sessions is 13 percentage points.