Building research partnerships to address the opioid crisis in Minnesota
In 2019, Minnesota Management and Budget and the Minnesota Board of Pharmacy submitted a proposal to develop a randomized evaluation of the state’s prescription monitoring program (PMP) to J-PAL North America’s State and Local Innovation Competition. The PMP is an electronic database that tracks controlled substance prescriptions in the state.
Their proposal was awarded funding and technical support, including matchmaking with affiliate researchers with expertise in the field. The researchers then applied in 2020 to the State and Local Innovation Initiative and received additional funding and STReaM support to conduct the evaluation. The ongoing study evaluates the impact of sending informational letters on the dangers of co-prescribing opioids with benzodiazepines and gabapentinoids on the PMP usage of providers.
We chatted with Senior Manager Weston Merrick of Minnesota’s Management and Budget Impact Evaluation Unit and lead researcher Adam Sacarny of Columbia University, who shared their takeaways from the process of launching a randomized evaluation through the State and Local Innovation Competition.
Why was MMB and the Board of Pharmacy interested in rigorously evaluating the prescription monitoring program?
Weston Merrick: Like many other states, Minnesota is very focused on the opioid epidemic. In Minnesota, the Board of Pharmacy, and the Prescription Monitoring Program (PMP) that it runs, plays an important role in mitigating and preventing the harm of opioid use disorder. They were interested in using rigorous evaluation to improve their operations and maximize the positive impact of the PMP on prescribing.
Over the last few years, Minnesota’s Board and the Federal Food and Drug Administration (FDA), for that matter, had been putting a lot of thought into the potential harm of co-prescribing opioids with other drugs, like benzodiazepines and gabapentinoids. So when we started to discuss places the Board could use evidence to inform operations, co-prescribing came up.
How did the research partnership between Minnesota and the research team come about?
Weston: After our proposal to the Innovation Competition was accepted by J-PAL North America, the technical assistance team started to identify J-PAL researchers who might have experience in these domains. Fortunately, there was an affiliate, Adam Sacarny, who had a related study. He and J-PAL brought researchers Mireille Jacobson and David Powell, who had similar expertise. We had an initial meeting where we laid out Minnesota’s interests and goals, which turned out to align well with their research interests. They've been wonderful to work with, and we are grateful to get to work with academics that are so knowledgeable and so pragmatic.
Adam Sacarny: The fact that this research was proposed by the implementing partner in the first place was a really good sign. It signaled that they were going to have the capacity and resources to make this project a success. The questions they wanted to study were also really interesting and aligned with our own research interests.
How did the team arrive at a research question that was of mutual interest and feasible to assess through a randomized evaluation?
Adam: Minnesota was initially very interested in doing something to try to increase the use of the PMP. That was interesting to me because there's research by Thomas Buchmueller, Colleen Carey, and others that show that when prescribers are required to check the PMP, there is a decrease in inappropriate prescribing. To arrive at our specific research question, we brainstormed the potential intervention mechanisms we could use, and we chose sending letters, an approach which I've studied in the past.
What conditions were necessary for this project to be possible?
Weston: We're fortunate in Minnesota that the legislature, a few years ago, invested in my unit, which is a group of data and social scientists that use state administrative data to evaluate the efficacy of state investments in human services. So we already had researchers on staff who were able to be a match-maker or a broker between the agency and the academics.
After you decided to move forward with the partnership, how did the collaboration continue to be successful?
Adam: It has been a huge help to have the support of J-PAL, dealing with a lot of the logistics and project management. There are so many different, often minor, things that need to get done to make the RCT happen, and J-PAL helped ensure that everything stayed on track.
Our collaborators on the state-side, who have been terrific to work with. They were pushing this research forward so quickly, and we never could have made that happen without their amazing support.
And what was exciting as well as challenging, about the experience?
Weston: It's always difficult to make sure that we are doing a study that is timely for the state but also current to the academic literature. I'm incredibly impressed with the academics’ ability to take our initial question and alter it in ways to make it relevant to theory, the field, and our project. It's been very exciting to work with and learn alongside Adam, Mireille, and David.
Sharing and cleaning data can be complicated, but we were able to do both efficiently while ensuring privacy and data security. Ultimately, there's a positive benefit to having this research tell us whether something is working or not. The Board of Pharmacy has been a wonderful partner and understood the potential to improve operations of the PMP.
Adam: We’re excited to work on a project that has come together quickly to address such an important health care issue. It did take some time to arrive at the intervention and to get all the relevant approvals, but once that was in place, the process of randomizing, intervening, and collecting post-intervention data should only amount to a matter of months. It feels good to be able to turn around results quickly to the implementing partner because they need to know if the interventions we’re testing are effective and if they should invest more resources in them.
Do you have any advice for researchers who are partnering with the state and local government to conduct a randomized evaluation?
Adam: I would recommend having an internal advocate at the implementing partner who can flag potential red tape. An internal advocate can also help you connect with staff on-site who will support the project and answer questions about internal processes and administrative data.
Looking at the long term is important. In our case, our timeline changed a bit, which raised the question of whether other policy changes that could affect the intervention. It’s useful to map what’s happening in the next year policy-wise to consider how time-sensitive your intervention is.
Another thing of note—there were some data gaps that we needed to fill to make the study happen. We could have produced that data ourselves painstakingly, but we chose to rely on a third party vendor that had produced the relevant data already. It was expensive, but we could be sure that the data was high quality, since the major players in the healthcare industry used it. I’d recommend utilizing third party vendors that you and your implementing partners trust when these sort of data needs arise.
Do you have any advice for other states and policymakers who are interested in doing something like this?
Weston: First, I’d re-emphasize the importance of a broker. It’s critical to identify people in your organization who understand the needs and interests of each of the stakeholders.
A second thing is to understand that high-quality research takes time. This has been a relatively quick study, but it still is going to be two to three years from funding to having results that we can use. It's slow, it's challenging, but it's worth it in the end.
Finally, it's meaningful to manage expectations with state partners. It’s important to communicate from the outset that we may get a result that isn't positive, but it still tells us something valuable. It’s important to discuss at the outset that if the research tells us that a service isn't working, we should change it by doing ‘X’ or ‘Y’. Or if service is working, we can take these steps to scale it up. This planning helps ensure that findings are ultimately used.