Bridging the Gap Between Hospital and Community-Based Services for People with Substance Use Disorders

By Vanessa Finisse, Center for Health Care Strategies


Personalized patient navigation services can be valuable for people with co-occurring behavioral and physical health conditions a population that is often at a higher risk for acute care needs, hospital readmissions, and poor outcomes. As patient navigation programs have different structures and staffing models, health systems looking to implement a navigation program can assess the relevant evidence for programs based on their population of focus, setting, and other characteristics. A randomized controlled trial in 2021 evaluated Navigation Services to Avoid Rehospitalization (NavSTAR), an intervention designed to reduce hospital readmissions by combining substance use disorder (SUD) consultation services with intensive case management and linkages to community resources. Participants in the intervention group received consultation services from a multidisciplinary team comprised of an addiction counselor, nurses, a social worker, psychiatrists, and psychiatry residents with addiction medicine specialty, in addition to coordinated case management by navigators who were social workers. As a result, these patients had significantly fewer inpatient readmissions and were more likely to enter community SUD treatment within three months of discharge.

The Better Care Playbook recently spoke with lead researchers of the NavSTAR evaluation Jan Gryczynski, PhD, senior research scientist at Friends Research Institute, and Christopher Welsh, MD, medical director at the University of Maryland Medical Center (UMMC) Substance Abuse Consultation Service to learn about the trial and the value of providing personalized navigation services to patients with SUD.

Q: What led to the development of the NavSTAR model?

It's one of the oldest addiction consultation services in the country that's integrated within a hospital, and it’s well-staffed and resourced. But despite the availability of those services, we were still seeing the same patients come back again and again.

A: Dr. Gryczynski: When I was first introduced to what Dr. Welsh was doing at UMMC, it was impressive to see the array of services that were provided in addiction medicine. It's one of the oldest addiction consultation services in the country that's integrated within a hospital, and it’s well-staffed and resourced. But despite the availability of those services, we were still seeing the same patients come back again and again. There was a capacity to do referrals, but not to follow up with patients in the community. That's the gap that we sought to fill with the NavSTAR model – to create a bridge from the hospital to community services.

Q:  Who were eligible participants for the study of the NavSTAR model?

A: Dr. Gryczynski: Participants were UMMC patients seen for medical or surgical problems who had also received addiction consultation services. Some people were hospitalized directly as a result of their SUD, while others were hospitalized for unrelated conditions, but their SUD would affect their prognosis and long-term health outcomes.

Q: What are the most important elements of the model and the services that are provided?

A: Dr. Welsh: One of the key elements of NavSTAR was extending the navigation services from 30 days to 90 days to allow more time for engagement and support. Another important element was engaging clinical social workers to serve as navigators. We also secured visitations for our navigators to follow patients discharged to the skilled-nursing facilities, which was challenging due to the laws and regulations related to treating opioid use disorder in these facilities. 

A: Dr. Gryczynski: Additionally, our navigators were trained to use motivational interviewing techniques to address the internal barriers to change that people often experience, in tandem with providing care coordination services to resolve practical barriers to treatment.

Many of our study participants also experienced housing insecurity and homelessness, and we saw a lot more people getting discharged from the hospital to subacute nursing facilities and homeless shelters than we had expected. We know that the intervention was particularly helpful in linking people to care that included some form of a recovery housing component. Nonetheless, it was challenging to figure out the housing part and deal with some of the persistent issues in the skilled nursing facilities. Oftentimes, these facilities may not have the ideal environment or care teams best suited for the patient’s needs, but are at least a warm place to stay while we determine longer-term placement and stabilization. While our experienced navigators were able to overcome those challenges, it really hammered home for me the difficulties that some of these patients who don't have support face.

Q: What are the key takeaways from the study? Is there anything that surprised you about the findings?

The main takeaway for me is that finding and using a proactive approach to help people get the care they need and link them to community resources can improve people's outcomes and have a real impact on hospital readmission.

A: Dr. Gryczynski: The main takeaway for me is that finding and using a proactive approach to help people get the care they need and link them to community resources can improve people's outcomes and have a real impact on hospital readmission. I was surprised with how high the service utilization rates were for emergency department visits and inpatient hospitalizations, at baseline and during the study, as well as the role of social determinants of health on outcomes, and the levels of complexity and multiple comorbidities that this patient population faces. I was also surprised by the degree of care fragmentation — we had patients whom we could see went to multiple emergency departments and hospitals in the same day. We are all aware of the social barriers and systemic issues, but the extent to which they played a role was quite striking.

A: Dr. Welsh: No matter how many navigation services we offer, some things must change at the system level to support improved patient outcomes. The navigator can certainly help, and statistically it helps save money, but even with the navigation group, there were a lot of problems that probably no one could have helped with because we don't have nearly enough safe recovery housing that will accept someone if they’re receiving methadone or buprenorphine.  

The other takeaway is that we were one of the first major studies in the state that used CRISP, our statewide electronic health information system. At the time, most of the outcomes reported in behavioral intervention studies were based on a patient self-report. By using CRISP, we were able to look at utilization data across other systems for a more comprehensive assessment of hospital use and found that there were study participants who did not report visits in other emergency departments.

A: Dr. Gryczynski:  That was an important advantage for us because we gained access to objective data on hospital utilization for the whole sample, even when we couldn’t find the patient for follow-up, which is a common challenge with this population.

Q: What are considerations for organizations that are thinking about refining their patient navigation approaches or adopting new programs for individuals with multiple complex conditions, including SUD?

A: Dr. Welsh: If you do not have a grant, one of the hardest things to figure out is how to bill for those services. For example, we were debating what a peer recovery specialist could do. In Maryland, peer specialists can’t bill Medicaid for recovery support services, although other states do allow them to bill for services.

You need to think about the community capacity for treatment. Where do you refer people? Is there enough medication treatment for opioid use disorder? Is there enough housing?

A: Dr. Gryczynski: Different hospitals are going to start out with varying degrees of addiction medicine expertise and capacity. Therefore, hospitals interested in establishing this type of program would need to decide how best to make an impact. If there isn’t an addiction consult service in the hospital, the first step would be to establish one to identify patients with severe SUD, then ensure that there are appropriate resources and services in place to respond to patients’ needs while they’re admitted. Within the University of Maryland system, we had an outpatient buprenorphine treatment and outpatient counseling service across the street to refer people to, and a methadone clinic down the road. We also developed strong relationships with other treatment providers. Lastly, you need to think about the community capacity for treatment. Where do you refer people? Is there enough medication treatment for opioid use disorder? Is there enough housing?

Q: How can providers and care team members promote the use of patient navigation?

A: Dr: Welsh: You must be proactive and problem solve. You may have to make phone calls or set up meetings, things that are not necessarily part of your job at the hospital, but that could support patients in getting the help they need. It also involves understanding your local resources. For example, for larger hospitals and systems, it makes sense to employ navigators who are based within the hospital but for other settings, including rural, it might be more effective to partner with the local health department.

Q: What are the next steps for the program and for your research?

A: Dr. Gryczynski: The next step I see is replication in a larger and more diverse sample across multiple hospitals. We showed compelling outcomes that are worthy of a larger, broader study to test whether this model can work in different communities and with different patient populations. There are a lot of sub questions within that: What is the optimal length of the intervention? What’s the best kind of staffing model? Do you really need master's level social workers, or peers, or a team-based approach? How do you get hospitals up to speed to be ready to implement some of these services? How do you make them work within each individual system? We hope to be fortunate enough to do that kind of study in the future.