Incentivizing Recovery: Payment, Policy, and Implementation of Contingency Management

Alison Buttenheim, PhD, MBA, keynoted the convening.

It doesn’t have to be an either/or. Implementation science frameworks can provide the right structure to identify research or practice questions. Design thinking can complement that by offering processes/methods for understanding provider or patient perspectives; refining and focusing the challenge; ideating solutions like implementation strategy designs; and prototyping or de-risking early-stage designs before pilot testing. Use both!

Design thinking is just a structured approach to problem solving.  Design thinking can be applied to the development and testing of intervention and implementation strategies. It could be equally helpful for formulating an advocacy or advertising campaign. We have used design thinking at Penn in the early phases of funding proposal development to inform the design of the project.

I believe that there is strong support among federal health (HHS) and policy (ONDCP) leaders for the use of contingency management as an evidence-based practice for the treatment of stimulant use disorder. There appears to be a reluctance to eliminate the $75 cap (the current major limiting factor) because the practice of providing higher amounts of federal money to individuals with stimulant use disorder may be viewed as a government “give away” program, and this may not have broad taxpayer support.  So, in my opinion, this issue blocking the effective use of contingency management is less a policy issue than a political issue.

We are aware of two states that have committed to using opioid settlement dollars for contingency management: Rhode Island and Vermont. Both of these states are in the preparation stage, and neither is operational yet (to our knowledge). There may be other states, but this is what we know.

Higgins has published data that documents that the use of community reinforcement approach, together with contingency management, can extend the effects of contingency management and strengthen some of the other treatment effects. I do not know of research that has looked at education per se as an adjunct to contingency management.  In the research I did at UCLA, in two large clinical trials, I found that adding cognitive behavioral therapy to contingency management did not increase or extend the effects of contingency management.

As of now, a Medicaid 1115 waiver is one way that states could make federal money at higher than $75 per individual available for use with contingency management.

There is most research support for abstinence-based contingency management, but also lots of studies showing that incentivizing other behaviors can be effective (e.g., medication adherence, treatment attendance). It’s important to keep in mind that in contingency management you “get what you pay for”, so the outcome that you are incentivizing (e.g., treatment attendance) tends to improve while other related outcomes (e.g., abstinence) improve much less if they are not what you are incentivizing. As long as programs implement contingency management with this knowledge and the right goals, contingency management has the potential to improve treatment engagement for individuals not ready for/desiring abstinence. One could argue, however, that for the sake of advocacy and given the research base, the focus should stay on abstinence-based contingency management at the present time.

Contingency management is most straightforward for stimulants because it is relatively easy and cheap to test for stimulants on point-of-care tests and because we lack other evidence-based treatments for stimulant use disorder. However, there is great research (by Mike McDonell, Kait Hirchak, Sara Parent, and their team) showing that contingency management is effective for other substances, like alcohol, as well. These protocols/tests are much newer and have not thus far been the focus of widespread implementation efforts.

There is not good empirical guidance on this issue.  There is one study by Petry and colleagues that shows that $80 per patient does not produce a contingency management effect.  So we know $75 is inadequate.  Many of the studies using the prize-based approach have used amounts in the $400 range.  The voucher studies have used amounts in the $1000 range (per patient per 12 weeks).  However, we don’t have parametric studies to identify how much is necessary for optimal response.  And in some current projects, like California, the maximum amount used is $599 per patient.  $599 is chosen to stay under $600 per patient, per year, to avoid giving participants a 1099 tax form.  So, $599 is not based on research data, but on the most that can be used and avoid giving tax documents as part of treatment.

To answer the cash part of this question– given the language in the OIG’s Final Rule, there would appear to be a much higher concern about “diversion” (i.e., fraud) when cash incentives are used. While cash is highly motivating because it is the most easy to use on a wide range of purchases that people desire or need (I know some people want to pay their rent or car loan), there are other ways to distribute motivating reinforcers (such as restricted gift cards), that still afford the recipient a fair amount of choice. I think we shouldn’t let the perfect be the enemy of the good, and restricted gift cards also offer effective reinforcement. There is already enough fraud concern among regulatory bodies that may be part of slowing the effort to get the $75 cap lifted. If introducing a request to also allow the use of cash slows that effort further, it may have the unintended consequence of hamstringing the effort to get an evidence-based dose of contingency management to people who need it.

There is no evidence of increased use after incentives are stopped. A recent meta-analysis suggests that contingency management has positive effects on use even after incentives are stopped. 

To add to this, the research on post-treatment effects of contingency management is limited and mixed, with some showing contingency management as no different than control at 6 months post completion of a contingency management program. However, the rate of return to use is similar (and often a bit lower– see meta-analysis linked above) when contingency management is compared to other substance use disorder treatments. All the substance use disorder treatments available include a risk of return to use after the treatment is completed. However, it is also true that periods of abstinence have a positive effect on health and wellbeing and people tend to make many quit attempts in their lifetime. That contingency management is a great opportunity for a quit attempt and/or a period of abstinence, with the potential for effects to last at least a year post-treatment cessation, is a very positive outcome.

California is the first non-VA large scale trial under a Medicaid 1115 waiver. The priority in this protocol was to use an empirically sound protocol and determine if it could be delivered with fidelity and with fraud prevention guardrails in standard community treatment programs. So far, this looks promising.

There have been surveys done on provider attitudes about contingency management and, yes, some providers have disagreements with use of contingency management.  However, over time, these reservations appear to be less of a problem.

Montana has the distinction of starting their statewide contingency management program BEFORE California and Washington did! This is a great example that contingency management is appropriate and well-accepted in rural communities (anecdotally, we’ve seen a greater rate of contingency management adoption in rural areas of Montana and Washington!). As a trainer, I have noticed regional differences between provider orientation toward a harm-reduction (non-abstinence) approach versus a total-abstinence approach, which may create some contingency management adoption barriers in both cases, and inform the need for contingency management education and myth-busting to encourage buy-in.

This is a great question. It’s important to keep in mind that contingency management is an individual-focused intervention that uses fairly minimal monetary incentives (not enough to cover basic needs) to incentivize a very specific outcome. Individual-focused interventions for substance use like contingency management need to be implemented together with systemic interventions that address factors like housing insecurity, access to health care, etc. – especially those that promote prevention as we know that prevention of SUDs is preferred to having to treat SUDs once they have developed.

The California initiative did this – a description is provided here.

The Penn LDI website has some information about this! 

There is at least one document that addresses this issue of tax exemption of contingency management dollars through the Welfare Exclusion principle.  This was done by folks in Allegheny County (PA).

Contingency management is a potential tool in any medical setting! One thing to keep in mind is that contingency management is most effective when the incentives are provided frequently to incentivize behaviors that can be objectively verified. For that reason, MOUD programs where patients dose daily, intensive outpatient programs and programs where patients present to care more frequently are natural fits. Other settings require more creativity (for instance there are emerging digital tools) and/or patient willingness to present to treatment regularly (and staff availability to monitor the target behavior).

Contingency management is an evidence-based adjunct to MOUD and as noted by a few panelists, the VA is an incredibly successful example of contingency management implementation (and many patients were on MOUD). Other examples outside the VA are described in Substance Abuse and Addiction Science & Clinical Practice. 

The VA currently has programs incentivizing abstinence and also incentivizing adherence to injectable MOUD, see description here.

This is another example of form vs. function – contingency management is effective for all, but if cash is triggering then the form of incentives for that patient could be gift cards to a grocery store, prizes like toiletries, etc.

As far as a “screening” process, I would suggest it is more of a conversation that is had during the consenting/patient agreement. This is where the exact features of the program are shared so that the person receiving contingency management gets the opportunity to understand expectations, get excited about the program, feel inspired that it can help them, and share concerns. This is a perfect opportunity to hear what might be challenging about the program for the patient/client so those can be addressed within the structure of the program.

In our experience the major issue that has prompted action is recent data documenting that stimulants are playing a role in over 50% of the overdose deaths in the US.  And in some states, more than 50%.  And if policymakers want to reduce overdose deaths they have to provide effective treatment for individuals with stimulant use disorder.  And contingency management is the only approach with robust evidence of efficacy.

The VA has sent out “Contingency Management Success Stories” every month to a VA listserv for substance use disorder treatment providers for most of the time the VA has been doing contingency management.  These have not been published, however, outside of VA.  I don’t know of any other published testimonials. 

Check out the DynamiCare app. Several studies have been published about it. There are other apps that have contingency management as a core feature: AFFECT, Q2i and CHESS (and there may be others).

It’s important to understand the distinction between apps that deliver the full package of contingency management and more (such as DynamiCare) and the apps that act as “incentive managers” so that contingency management is still conducted by local staff at a clinic/agency, but the incentive (gift card) distribution is handled by a third party. Each version has their own strengths.

Contingency management is a behavioral approach to treating stimulant use disorder that systematically reinforces (aka rewards) reduced use or abstinence from cocaine and meth and/or other recovery-related behaviors.

Generally, the more specific the behavioral target the better results will be. So, if you want to address co-occurring stimulant and opioid use, reinforce abstinence from just those two substances. The more substances you reinforce simultaneously, the smaller the effects will be. You can certainly monitor many at once, but it is worth being aware of the evidence!

Our training emphasizes picking one substance class at a time to provide the best opportunity for successfully accomplishing the recovery goal. (As mentioned above, the research indicates that focusing contingency management on a narrow behavior goal, like abstinence from a single drug class,  has the biggest effect size.) In the case of opioid and stimulant co-use, we emphasize the importance of MOUD to address the opioid use and suggest that the contingency management focus only on the stimulant use.

Our training team has fielded interest calls from some providers or healthcare organizations located in the south and at least 1 southern state, however, we have not yet had the opportunity to provide training to these sites.

Dynamicare Health is an app that allows for remote self-administered testing that might be worth looking into.