Gregg Masters 00:10
You’re listening to PopHealth Week on HealthcareNow Radio. I’m Gregg Masters Managing Director of health innovation media publisher of ACOwatch.com and your PopHealth week co host with my partner co founder Fred Goldstein, the president of Accountable Health, LLC, a Jacksonville, Florida based consulting firm. Our very special guest today is Jeffrey Brenner, MD, Senior Vice President of Integrated Health and Human Services at United Healthcare. Jeff is also the founder and former executive director of the Camden Coalition of healthcare providers, Dr. Brenner weighs in on his work in the social determinants of health and addresses the just reported results published in the New England Journal of Medicine titled Healthcare Hotspotting, a Randomized Control Trial, he closes with a preview of his keynote at the 20th anniversary of the Population Health Colloquium hosted by the Jefferson College of Population Health in Philadelphia, PA from March 30, through April 1 2020. So, Fred, over to you help us get to know Dr. Brenner, and what he’s up to at United Healthcare.
Fred Goldstein 01:24
Thank you so much, Gregg. And Jeff, welcome to PopHealth Week.
Jeff Brenner 01:27
Thank you. Great to be here.
Fred Goldstein 01:28
Yeah, it’s a pleasure to have you really is fantastic. And before I get started, I just want to thank you for actually all the work you’ve done to open people’s eyes and get us all to account for some of these other issues, where to healthcare, through your early work that we’ll get into. So really appreciate your efforts around that.
Jeff Brenner 01:44
Thank you.
Fred Goldstein 01:44
Yeah. So why don’t we start off, give us a little bit of background of how you really got into this whole area and some of the work you did in Camden. And perhaps we start there. And then we’ll move a little bit deeper into some of the more recent studies.
Jeff Brenner 01:54
So I’m a family doc and graduated residency out in Seattle, and wanted to move close to where I’d grown up and wanted to work in a community that really needed access to care. So got a job and decided to move back east and work in Camden, New Jersey. So for folks who might have heard of the place, it’s usually in the headlines with East St. Louis and Flint, Michigan. So it’s very poor, small urban city, it’s about nine square miles. And it’s directly across the Delaware River from Philadelphia, you come across the bridge, and you go through Camden, it’s had all the challenges of cities in America de industrialized and hit on hard times, it lost a lot of its population. So I moved there and worked in a small three exam room office, saw kids, adults and delivered babies, and it really dug in and committed myself to learning about the community and working there as well.
Fred Goldstein 02:49
Fantastic. And in terms of that work, can you talk a little bit about the the early work and your your hotspotting? And what you found and how you then went about trying to solve that?
Jeff Brenner 02:59
Yeah, so like lots of primary care providers, you, you know, particularly in an underserved community, you get overwhelmed with the things that you see. So I was often in my office until seven or eight o’clock at night doing paperwork I had, I was running from room to room to room trying to do the best I could about half the practice, was Spanish speaking. So you’d walk into a room and there might be a woman in a wheelchair, who is older, speaks Spanish is a little bit confused, was recently in the hospital and you play 20 questions, trying to figure out what’s going on. And you pull out all the paperwork from the hospital, you pull out of all the medication bottles, and it’s just totally confused. They changed up all the medications, she’s got Brand name and generic bottles, the family is overwhelmed. You know, meanwhile, you’ve got a waiting room full of patients and you’ve somehow you’re going to magically figure that out in 15 minutes. Then at night, you get a piece of paper from the insurance company that says, you know, we’ve been unable to contact your patient, can you fill this form out and tell us everything that’s been going on with your patient, then you’d call the case managers and say, like, I really need help with this case. And you know, they’re sitting in a cubicle far away and are really unable to, to help. So you experienced that day by day by day as a provider, and you’re not being paid more, you’re struggling to hire staff, you’re struggling to pay the bills, and meanwhile, hospitals building new wing, they’re expanding the size of the emergency room, and just the system seems out of whack. You know, that’s what life is like for lots of primary care providers. And it’s even more magnified in an underserved community. And I love taking care of my sickest patients. I really felt like when I spent time with them when I leaned in, it could make a difference for them. At what point the city was the most dangerous city in the country. It was a frightening time to live there. And every day I was coming to the office and I was taking out staples. I was taking out sutures and I’d have my younger patients tell me that some awful thing had happened to them. They’d been shot they’d been beaten They’d been stabbed and I’d say, let’s, that’s horrible, didn’t you? Did you call the police? And they would laugh at me. They’d say, like, not gonna call the police there, they’re gonna beat me up, they’re gonna hurt me. Like, are you crazy, and you just don’t want to think that social order is broken down to that extent. And it turned out everything they told me was true. Police would plant evidence on people, they would beat people up with a tremendous amount of corruption in the local police department. So here we’ve got a dysfunctional police department, and we’re the most dangerous city in the country. And I thought that the report the official police statistics, were an undercount. Because I had a lot of patients that had bad things happen to them, and they weren’t reporting it to the police department. So we started a project where we got all of the accident and injury claims data, and it was a summer student project with public health student and mapped it graphed it and charted it And sure enough, like the rate of accidents of of assaults in the census track right next to the hospital, just from one hospital’s data was one in 40 kids in a one year period, were injured badly enough to come to the hospital normally you count that stuff in like one in 10,000, like it was a staggering rate of violence. And the hospital data was a better reflection of the true rate of crime in the city. So that got me really turned on to this idea of administrative datasets that we have data sitting around, that could tell a story that could improve the populations health and change health outcomes. So we went to the other two hospitals in Camden, got partners in those hospitals, we went through the IRB, and now had a data set that had claims data, that was all payer, all citizen data for the entire city. And we learned how to link it match it. It was sitting in two 50 dollar hard drives on a desktop computer. And this, the drives were locked in a safe under my desk. And we had a bunch of student interns that were helping us figure all this out. We mapped it charted it it graphed it one of the things we found was the in the nursing homes, you could see this like hotspot of falls. And if you looked over time, you could see the fall rate went way up, and then went way back down. And during that time period the leadership and management had turned over in that nursing home. Everyone knew that that nursing home was unsafe. But here the data showed it. And we weren’t doing anything with that data. And we had even more data showing the patterns of falls injuries, accidents and assaults in the city. We went back and amended the data set and got all of the claims data from the city over a three year period, mapped it, graphed it charted it and we began to find buildings and neighborhoods and locations that had staggering rates, of readmissions and cost as well. And no one up to that point had ever really looked at hospital claims data where it was an all payer data set at a community level mapped out at like city block and building level, we found two buildings that were the most expensive buildings in the city. One was a nursing home. And the other was a building full of dual eligibles and spent a bunch of time in the buildings. And they were both had great management. They had people running the buildings that really cared about the patients. They were beautiful buildings. But the stories that we got from the patients were terrible. And from the staff, they just a lack of access and sort of the kind of the meanness of the healthcare system, and how inaccessible it was. So the other thing is, I had all the cost data, you could see like 1% of the city was 30% of the spend, and I could see my own patients in the data. And you know, you’re like holy bejesus, look how much the hospital made for Mrs. Rodriguez and Mr. Jones, and I’m getting like 19 to 35 a visit and they’re getting like 100,000 a pop, it just like made no sense like the the mismatch of need resources and spending. And the way we’re building new buildings and expanding the ER and primary care offices were getting boarded up and closed in the middle of that I had broken off from the hospital. And my office was a private practice. And I like couldn’t pay the bills, my payment rates were getting cut by the Medicaid HMOs. And I had to close the office. So you know, there’s just like this incredible dissonance between what I was seeing and all the patients and the kind of the just failures of the delivery system to deliver kind, compassionate, accessible and well structured care and the way we were spending money and resources and the lack of data to measure all that. That was the clarion call for the organization. And that’s what we spent the next you know, 10, 15 years tackling, and I had no idea what we were signing up for and how hard this was gonna be. You know, I thought this was just going to be like breaking down the silos and coordinating care and navigating care. I had no idea That this was gonna really require a wholesale transformation of how we deliver care to complex people.
Fred Goldstein 10:06
So before we get into delivering care, you mentioned very early on, you’re pulling together all this data and you talked about an IRB. So would you explain to the audience for those who may not know what an IRB is and their importance and looking at the review of the process? Yeah, sure.
Jeff Brenner 10:21
It’s called an Institutional Review Board. And every hospital and every university usually have them. And it’s meant to protect people who were part of research trials. So if you’re working with data, or you’re testing a medication or device or doing any kind of work with humans, you have to go through an Institutional Review Board, and make sure that you are protecting the data, protecting people from harm, and that you’ve got a real good purpose for the work that you’re not just experimenting for experiments sake, but you’re doing something that will improve the health outcomes for people.
Fred Goldstein 10:58
Talk a little bit about you about how difficult this is and how deep you had to get into it. What are some of the things that were done based upon the data sets that you put together in the analysis and determine there were issues, whether were the building or some of the other social determinants,
Jeff Brenner 11:13
so we decided to create an organization to tackle the problems. So it was imagined to be a city wide coalition of stakeholders. We formed the Camden Coalition of Healthcare Providers almost 16 years ago. It’s a membership nonprofit, and our members are all the unreleased stakeholders locally. So hospitals, FQHCS, primary care providers, nonprofit organizations, we even let church groups join and patients join as well, we had a patient advisory board, we made decisions pretty much by consensus, which is pretty hard to do. We increased the data holdings of the organization. So we started a health information exchange. Eventually, we had five health systems feeding in their data, both their ADT feeds, which are the notifications when a person gets admitted, discharged and transferred, as well as the electronic health record data. It was all in a data warehouse. We also ended up with all of the call and arrest data from the police department, the jail data, the housing and homelessness data set the perinatal co operative data for high risk pregnancies, and the school district data of household and truancy. And all the data is linked together and allowed us to really think about different populations in different segments, and what their unmet needs might be, we built a city wide care coordination model, so that we could, we had badges to all the local hospitals, we could walk in and out of the hospital, go right into a patient room, and enroll a patient in our care models. So we had business associate agreements and collaborative agreements with all the hospitals, we ended up forming a Medicaid Accountable Care Organization. So we passed state legislation to create a city wide geographic shared savings model, we ended up in contracts with the two largest Medicaid HMOs, United Healthcare and Horizon to do city wide shared savings. We built an infrastructure to feed data to the local primary care offices every day of who’s been admitted in the last 24 hours. And then we paid the primary care offices $150 for every patient that got in with within seven days. And we knew every day for every patient that was admitted to any hospital when they got back into primary care, because in order for the primary care providers to get the payment they had to encode in the system when the patient got seen. So is a remarkable sort of virtuous feedback loop of data. We ended up forming a national center for complex social needs, as we began to realize that there was no way we’re going to get this out in Camden, they had to link up with partners around the country that were doing similar work. And I think we started to realize this is way harder than we’d ever like set out to, to do and that it was going to take a community of learning to figure this out. And that was going to need to be a national movement. And that national field building effort to really build a field of complex health and social needs, we pivoted into doing a randomized control trial. So we felt very strongly that our work needed to be rigorous, that we needed to test it. And we needed to let the chips fall wherever the chips were going to follow. And we were going to publish the results no matter what the outcome was. And, you know, we had enough notoriety that we could have like just coasted on our laurels and just declared the intervention working. But we refuse to do that we raised a lot of money to to fund a randomized controlled trial to test our care coordination model.
Fred Goldstein 14:46
So you put together what’s really an incredible community based approach linking it all up all the data care coordination, really a grand vision of how you do population health is community based upon what you learned. In all the data and analytics and turn that over to ultimately, the early studies, as I understand it showed some success. And then you said, let’s do an RCT on this or the group did and, and lo and behold the RCT, which was recently released comes out, and the data doesn’t show the changes that you had originally expected or, or had had seen in earlier studies. So where is that now? What do you think might be some of the root causes for that? And where do you go next?
Jeff Brenner 15:28
Great, great questions. So let’s turn the clock back five years ago, so even longer than that, actually, we first started designing the RCT around 2013, a pretty long time ago. And we partnered with a group called the Poverty Research Lab and MIT that does RCTs around the world, and their lead researcher is someone in Amy Finkelstein, who led the Oregon Health Study where they randomized access to Medicaid. At that point, there had been very few RCTs testing care coordination and care management. And and we were determined to test ours we received a CMMI federal grant which just allowing us to go from like five staff up to about 15, 20 staff. If you look back at that point, the dominant view at that time was that healthcare is disorganized, patients are confusing, that if we just do care coordination, care navigation, that we could significantly reduce costs and improve quality. And we had the work of Mary Naylor and Eric Coleman, which had been focused on Medicare patients that showed within person care coordination, care, navigation, and engagement in that critical hospital care transitions period that they were able to statistically lower costs for Medicare patients. We also had the Medicare chronic care demo with 15 sites being rigorously tested. And within a year, they pulled the plug on 12 or 15. Because they drove cost up, only three were allowed to go forward. And they were all boots on the ground face to face messy, hard models. So you know, the dominant feeling at that point was like Healthcare’s disorganized, which we all agree it is and that if we just stepped into the middle of that with a rigorous in person care coordination model, that we’d be able to reduce costs and improve quality. So we set out to enroll 800 people 400 intervention, 400 controls, we went right to the bedside of the two local hospitals, enrolled people, we would then consent them, walk out, randomize them and walk back in and let them know if they were the intervention of control group. We found people with two or more admissions in six months. And we were finding them through the HIE data. So every morning would come in at six o’clock, and screen all the data for had been admitted. We did 90 days of in person wraparound services, we would go with people to their appointments, we tried to get them back into primary care, within seven days, we went to them went with them to the Board of social services, help them get IDs, birth certificates, Social Security cards, whatever they needed, the primary endpoint was 180 day readmission rates. And what’s amazing is that the 180 day readmission rate, the baseline was 62%. So this is the most complicated sickest group that’s ever been studied in any care management care coordination study. So we set out to find the hardest patients and, and we found them and enrolled them in it. So what’s amazing is the natural regression to the mean, was 38%. In the two groups, that means if you just do before and after and the claims data and the cost these patients if you do nothing with them will drop 38%. So we were seeing the treat, we were seeing what we thought was a treatment effect in the claims data because we were seeing 30, 40, 50% drops in people’s utilization. And we had a reasonable hypothesis that this model was working. But think about how hard that it is to test an intervention where there’s a 38% just natural regression to the mean, whether you do anything or not, that’s a profoundly difficult statistical and clinical research problem. So about halfway through the study, we were looking at the data people were working with, and I asked my team, I want you to pull out the patients who are not doing well and tell me why they’re not doing well. And what we saw was that if you were homeless, severe mental illness, severe addiction, very significant medical complexity, jail involvement, arrest, which was not uncommon in this population, where you had significant earlier, later life trauma, the model wasn’t working, and those folks were going up and I would sit at night and kind of do the rounds and talk with all my staff. And they would tell me sort of heartbreaking stories of 15 minute primary care visits, 15 minutes to gastric med checks, counseling and behavioral health organizations of new grads that just gotten out of school. There was no treatment in the jail and people were worsening when they got out and people could not access addiction treatment. You have to line up at eight o’clock in the morning at the methadone treatment center. At the window. Patients were being sent to the local shelter and it could take three years to move to permit housing. And if you’ve had a lot of early life trauma, physical abuse or sexual abuse, lying in a big open shelter is terrifying. We also found a lot of incorrect diagnoses and incorrect medications. For instance, bipolar disorder is very commonly incorrectly diagnosed and resulting in incorrect medications. So, you know, the irony of it all is that we’re a coalition and my board members are the people who are delivering all the care. And what we discovered is the problem was the care that we were navigating to nowhere, that we were correctly finding and engaging the sickest and most complex people in the city, but we couldn’t find the right care for them. So we did a huge pivot at that point and began to assemble a different kind of services and different set of services. There’s an emerging evidence base back then, of new thinking about underserved populations. One of those areas was a model called housing first came out of New York was invented by a guy named Samson Barris. And his innovation was the cure to homelessness, for people who are lying out on the streets, is to move them directly into an apartment with low barrier to entry. And in, in the old model, it takes two to three years of perfect behavior until you get into a unit. Whereas in this new model, you moved rightly right into an apartment. So it turns out that in the old model, only 20 to 30% of people end up housed in the new model. 90% of patients can be retained in housing first. So we brought this to Camden, I had no idea what we were. I was terrified to get us into housing. I didn’t understand it. I didn’t. It’s expensive. It’s complicated. And when we started housing patients, it was like magic fairy dust. It was the most profound thing I’d seen in my career. It almost made me want to sort of hang up my stethoscope and forget about what we do in medicine. We, you know, we often harm people in medicine, but housing is it’s just unbelievable what it does for people. So I had watched up in Trenton, they’d done a tiny pilot with a patient who did who’d gone to the ER 450 times in a year. They got her into an apartment, and she went back 18 times the following year. So we got 50 vouchers form the state, we raised a bunch of funds for the wraparound and and vaulted ourselves into housing. We also got very focused on improving treatment of addiction and mental health. So I got offered a part time job to run the large academic outpatient resident teaching office at the local academic hospital, Cooper hospital, and it was doing about 30,000 visits a year, it had about 100 different doctors at residents and fellows, it did all the specialty care and primary care. It had all the residents teaching clinics, and agreed to to run it. And probably the craziest job I’ve ever had in my career. If anyone’s ever done that sort of thing. And it turned out to be a turnaround. We ended up letting a lot of stuff go turning folks over. And one of the things we did there was to start an addiction treatment program. We hired the city’s first fellowship trained addiction specialists, we expanded access to treatment of Suboxone, we eventually got all of the emergency room x wavered. The ob department x wavered. And when patients came in with an overdose, they were started on Suboxone in the emergency room. We also started something called an ambulatory ICU, which is a primary care office with behavioral health and addiction treatment that could see people with hour long appointments, and we took a bunch of these folks and put them in that clinic. So between the Housing First, the addiction treatment, and this more sophisticated, integrated ambulatory ICU model, we began to really feel like we’re making some headway. I think the key evidence base that we’re you know, not delivering with fidelity in care of underserved populations is access to Suboxone access to intermuscular anti-psychotics, access to housing, and access to trauma informed and sophisticated integrated primary care services. And I think that’s going to become the key to beginning to make progress with these more difficult populations. So the conclusion from the study is that navigation and coordination alone are they’re necessary, but insufficient to reduce costs and improve quality for the most complex people. And I think we’re at a fork in the road, which is we can throw our hands up and say, forget outliers, no one’s going to work with them, let them keep going to the hospital 450 times or we can dig in and figure this out. And there’s a field of knowledge to build here. A field of complex home social needs that I think we’ve got glimmers of and we’ve got some working models out there. There’s been a lot more work done in Medicare in geriatrics and palliative care. Then there has been in sophisticated addiction models, behavioral health models and primary care models that include social determinants for underserved populations. So I think there’s a real opportunity to move forward in that. And I hope that this trial starts to point in that direction.
Fred Goldstein 25:07
Absolutely. And I think you’ve done a great job of talking about how you transitioned that program and started to focus on a few key areas. We’ve got just about 30 seconds left, could you provide a little bit of a taste of what you’ll be speaking about at the population health colloquium in Philadelphia coming up?
Jeff Brenner 25:22
Yeah, so I got hired by United Healthcare about three years ago, couldn’t have imagined that my career, and it’s been an incredible opportunity to figure out how to scale the nuggets that are working. And we’ve been very focused on scaling Housing First, for medically complex patients, we’ve housed about 350 patients around the country, we’re in about 15 communities. And we’ll be in about 30 communities by the end of the year. And it’s just an incredible learning opportunity to figure out how to put this work in a context of an insurer. And what are the elements of scale, healthcare interventions don’t tend to scale very well. So it’s been a great learning opportunity. So I’ll talk about that. And I’ll talk about the where this hopefully, where this field will go in the future.
Fred Goldstein 26:03
Well, that’s fantastic. Thank you so much, Jeff, for joining us on PopHealth Week. And back to you, Greg.
Gregg Masters 26:08
And thank you, Fred. That is the last word on today’s broadcast. I want to thank Dr. Jeff Brenner, MD, as Senior Vice President of Integrated Health and Human Services at United Healthcare for his generous time and many insights today. For more information or to follow Dr. Brenner’s work, go to www.unitedhealthcare.com and search for social determinants of health. As mentioned, Dr. Brenner will be keynoting at the 20th anniversary of the Population Health Colloquium in the city of Philadelphia from March 30 through April 1. And for more information or to register for this annual goto event, go to www.populationhealth olloquium.com for PopHealth Week, my colleague Fred Goldstein and HealthcareNOW Radio This is Gregg Masters saying bye now.