From 3M Health Information Systems
Podcast Episode Transcript: Social determinants: Changing our view of the patient
Dr. Gordon Moore: Welcome to the Inside Angle podcast from 3M Health Information Systems. This is Dr. Gordon Moore and today, I’m talking with David Buck and Kallol Mahata from the Patient Care Intervention Center in Houston Texas. A little bit about where this conversation came from: David and I were in residency in family medicine together back in Rochester, New York and ran into each other online after many years. I found out that he and Kallol had been doing fascinating work in health care delivery to homeless people in Houston, bringing together information and care delivery across multiple spheres of influence in a way that I think you’ll find fascinating and useful. So with that, hello, David and Kallol.
Dr. David Buck: Hello.
Kallol Mahata: Hello.
David: Thanks for inviting us on the show.
Gordon: Yeah. You bet. And thanks for agreeing. And I’d love for maybe David, if you would start with describing how this thing came about.
David: Well, Kallol and I met in 1984 when I worked with his father with Mother Theresa, and that’s really where my interest in working with underserved populations began. He was three years old at the time. But in developing over the last 30 plus years, how we could be more effective with underserved populations, at the heart of it is how to integrate data more effectively so that we can help serve the needs of the patient. The more complex a given patient or client is, the more systems they interface with and the less capacity there is to provide an integrated care plan for them. And I learned that through caring for the homeless and watching the sickest people consistently die because of our inability to coordinate care.
Kallol has brought in the technical expertise. He’s been the technical architect integrating datasets, but also developing a health record that allows multiple systems to access a common record so that we can push a common plan that’s a values-based care plan. It’s the patients’ values, so that we can make a difference by using social determinants of health, not just medical data.
Gordon: So Kallol, you have a background in computers and information and you bring information from disparate sources together. Can you describe a little bit about some of those disparate sources and the work that you’re doing there?
Kallol: Sure. Our data really comes from three different types of datasets. One is from the health system or hospital systems data. The second is from the payer’s side which would be insurance providers or managed care. And the third part is really social systems data. So that would be data from jail systems, police departments, community service agencies like shelters, food banks, and so on. Looking at the need of patients, we realize that 80 percent of a lot of those needs were driven by their social factors than just 20 percent from the medical side. Yet a lot of the analysis that had been done was very focused on the medical and clinical side. Our approach was to say, “Can we bring these different types of data together to get a much more holistic view of utilization and needs for our clients?”
Gordon: I think that the influence of nonmedical information on clinical outcomes is significant. A lot of people are interested in this and wonder how they get their hands on this kind of information. So you mentioned jails, and that’s fascinating. First of all, why are you interested? And then tell me, how do you come about getting that information and what use do you make of it?
David: About ten years ago, we did a study that looked at the jail system and we found that merely by linking services for psychiatric meds, the day of discharge, we were able to reduce re-arrest rates, like readmission rates, by 62 percent. Another effort was done at the same time by our county, and they did a much more sophisticated project that included actual housing (ours included shelter beds; it was really on the cheap). The only thing our study did that theirs did not do was same day connection to primary care services. That’s an example we learned early on that linking social need with a discharge from jail was able to reduce re-arrest rates or readmission rates significantly.
So often, we find that the people released from jail are, by definition, people that have not been able to be productive members of society, yet they’re expected to do really what is almost impossible. If they’re discharged with a serious mental illness, they’re trying to arrange follow-up plans and they often don’t even understand the fabric of the safety net system.
Gordon: And when you crossover into this persistent mental illness which may be combined sometimes with substance abuse, is that the main type of person that you’re identifying and reaching out to?
David: That’s a great question. No, it actually is just an example of how one tiny need can make it impossible for someone to get care. And because we don’t link care in that example, we end up having a very costly medical impact. But if we had social resources, we could reduce that medical impact and the psychological impact significantly. But that person could be a diabetic, it could be someone that gets food stamps or that lives in a food desert and relies on the food bank, or people that have significant behavioral needs. But behavior is something that we could all improve on. If we just had better diet or exercise, we could impact our health more than any pharmaceutical intervention.
Gordon: So I can see the benefit in terms of health outcomes from identifying that somebody who lives in a food desert may have inadequate access to the kinds of nutritional support that would be beneficial to them. But as I’m thinking about the likely consumer of this information in participants is, let’s say, a hospital or a health system that’s looking at improving the health of a population for whom it’s caring and supporting. It begs the question, how do they connect with an intervention? The health systems how to they begin to figure out how to distribute food and solve food desert problems?
Kallol: I think one of the ways health systems and hospitals are starting to do it as an example is they’re integrating what they’re calling social determinants of health through questionnaires or through self-reporting within the health record when the patient shows up. What we’re stating is we already have that data and it’s in the form of utilization data that’s available from these different resources. So, step one is for us to be able to connect that data and show it to the hospital system and show the overlap and services. Really using utilization as a proxy for complexity as well as for the social determinants of interest to the hospital system to show what’s upstream impacting the client’s health and causing them to come to the hospital or to the emergency room today.
Gordon: And this data, you’re talking about the information you’re getting from social service agencies about interactions and events?
Kallol: That is correct.
Gordon: That’s also fascinating because, as you’ve described before, this information that’s not necessarily shared is interesting even if it’s just utilization data, because I suspect you’re unmasking the fact that a person is integrated with a particular kind of agency or service, plus the number and frequency of those types of services may have some bearing on risk and the sorts of things you might do about them. Is that true?
David: Exactly. Can you imagine seeing a patient where we started at the highland system, and instead of seeing someone for their first visit, you see them with all the utilization data they have? Not the one questionnaire that shows they live in a food desert that’s a cross-sectional, one point in time assessment, but actual utilization for the last six months, year, two years. Would you take care of that patient and think of their needs differently if you saw that they had had multiple visits to a food pantry or a homeless shelter, or if they had been to a recovery center or a place for mental health care? Would you look at them differently and try to make sure that their needs are a part of your care plan?
Gordon: And so to help facilitate this, you guys have developed not just the consumption of this information but also appropriate dissemination of this information that coordinates across multiple different agencies and intervention possibilities. Could you describe a little bit of that?
Kallol: Sure. So step one in the process is the analysis of the data and this overlap analysis that helps us bring the different datasets together. From that, we’re able to identify cohorts of patients that we would want to have a very targeted intervention with or the system you want to have a targeted intervention with. During that intervention process, we start off with developing a care plan that can be shared across multiple agencies, and is a single unified care plan that’s shared, which is showing the values of the patient. From there, we’re going into the needs and the barriers of the client, which help drive the goals and action steps required to achieve those goals or overcome those barriers.
That care plan is packaged inside of our health record system that multiple agencies have access to, depending on roles and depending on what kind of an agency they are. If they are a social agency, they are viewing slightly different data from a health care or a medical agency that’s coming in and seeing the health record information.
Gordon: Are you setting different permissions based on HIPAA requirements, that sort of thing?
Gordon: Got you.
Kallol: So depending on whether we are dealing with a covered entity or a non-covered entity and depending on what the patient or the client has consented to, we’re going to show different pieces of information.
Gordon: So you guys have developed a technological platform that consumes information based on partnership agreements with permissions. You can consume certain data and then share that again with appropriate permission so the agencies can log in and see which individuals are touched by other agencies, what are their conditions, what are their needs. Is that the premises actually?
Kallol: Correct. And it also enables a much more holistic view. What we’re seeing as an example from the analysis is, a patient has at least, on an average, five and a half to six different social agencies that they touch (when analyzing the data from the Homeless Management Information Systems). In each of these agencies, they are probably going to have a case manager who is creating some kind of a plan for the client or patient. What we’re seeing then is this: Patients have five or six different case managers, yet none of them are able to connect with each other or are able to share information, and we end up providing care to the patient in a silo versus reinforcing a common message to the client.
If I can see that they haven’t gotten their ID yet, the next time I see that patient, can I tell them, “Hey, it looks like you are trying to set up an ID with this agency, how is that coming along?” It provides a very different interaction with the patient when we were able to have a collective approach to providing care and when we’re able to target the care based off of their values or what’s important to them versus having a very problem-based approach. “Here are my ten different conditions, how am I going to come and try and solve that for you today?”
David: Looking at care from a problem list is often quite off-putting to the client or patient who see themselves as failures, failures of all these systems. They can’t get food, shelter, clothing, ID, etc. They can’t get their mental and physical or behavioral health needs met. And this multiple agency need really defines the complex patient. What often happens is there’s no accountability for all these agencies that may see someone hundreds of times to get an ID. They can’t get housing until they get an ID and yet, all these agencies may be seeing the same person for that one thing.. Our system tracks that and says, “Hey, this health system is looking at this issue and the substance abuse needs are being cared for by this entity” so we can all be on the same page.
Gordon: So, if you would, give me two things: One, an anecdote of a particular person who’s served by this and how that worked. And then second, talk about the kinds of outcomes improvements you’ve seen in Houston.
David: This first one is a patient that had bipolar illness and had HIV. He was seen in the ER almost everyday. We kept trying to take care of him and meet him in the ER, meet him on the street, or wherever he was, and, until we began to ask the question, “Who and what is most important you?” We kept almost arguing about how to address his needs. When we finally asked him who and what was most important to him, he said that his dog was most important to him. So, we still wanted to get him into primary care, but through his lens, we really changed the approach.
We said, “Well, what would have you feeling better about the care of your dog?” He said, “Well, sometimes, I’m going into the ER and I have to tie my dog up on a tree on a way in.” So we said, “Well, what would you need?” And he said, “Well, I’ve been offered housing, but I can’t get housing because they require you not to have an animal.” And so we said, “Well, he’s clearly been a service animal to you, we could codify his role in your life as a service animal. Would you be all right with that?” That’s how we were able to get him housing. Then finally, after eight years, he got Social Security and disability benefits, so he went from being a non-paying customer in the ERs to having benefits. Even with those new benefits though, we found that his ER visits didn’t change.
We went back to the drawing board and we said, “How is your dog doing?” And he said, “Well, the dog’s okay, but now I leave him in my apartment when I go to the ER.” And we said, “Well, you were thinking of going into primary care, what happened with that?” And he said, “Well, I don’t want to be one of those people that sits in the clinic all day and then they end up giving me medicine that makes me slobber,” meaning antipsychotics, and we said, “Well, who knows you the best? How you feel about things?” Well, certainly, he’s the expert. So we said, “Well, why don’t you entrust us to work with the side effects of the medicine and see what we can do?”
He finally agreed to going into primary care, which has integrated behavioral health, and reduced his ER visits to one visit in that next year. From hundreds to one visit. Just that shift of looking at things from his lens made all the difference.
Kallol: And from the outcomes in the last three years of our work, on the cohort of clients we’ve taken into our intervention, we’ve been able to avoid about 5.1 million emergency room visits and have been able to reduce those emergency room visits by about 52 percent. We also track things like quality of life improvement, using metrics like the DLA-20, which is an assessment tool they use, and we’ve shown to improve our client’s quality of life by about 16 percent on average. Over time, we’ve been constantly improving our model of care coordination, as well as the analysis and the system. We are now starting to track additional metrics, like what percentage of resources have we been able to connect clients to in the community, what percentage of their goals have they been able to complete.
Again, remembering that this is a six-month intervention and we’re not going to complete everything, but we’re setting them on a target, on a path to succeed in completing those goals.
David: For example, nationwide, there was a nursing study that examined Medicare patients with serious physical illnesses. They were able to, through the most basic care coordination methods, reduce cost by 30 percent. And what they did is they reduced that cost merely by informing the site of the care needed at these other places. There was no real shared record system. Well, what would happen if they actually could track that information? You referred someone for an appointment, did they ever make it? We never know if someone makes the appointment until now. But this system will help track what was successful and what continues to be a failure so you know if the top need is to refer people and get them from the ER into primary care homes, why isn’t that being tracked?
Gordon: You guys are now part of a medical school in Houston, is that right?
David: Yes. The University of Houston College of Medicine.
Gordon: Tell me about that. Why that affiliation, and how does it work?
David: This is a really exciting period in Houston because it’s our third medical school and the fifth in a 60-mile region, but it’s also the first to focus on indigent care needs. As our goal, we want greater than 50 percent of our class to be coming from the neighborhoods we seek to serve. And we’re going to be integrating the care with social determinants so it really provides the opportunity to start improving care and to start measuring what we call “evidence-generating practices.” This is a really exciting opportunity in Houston and in underserved care in general, to begin to apply what we know, which is that most of the impact in health care is not from the medical domain but from the social domain through these social determinants of health.
We’re designing evidence-generating practices that work at the grassroots level in the communities through community health workers to integrate the data so providers not only have the full picture of the client and the community in front of him, but can also work with community leaders to identify what their hopes are and what they would like to address as a community. One of the missing elements in health care is this value-based proposition of what does the person want, what does the community want?
And so really, from a training perspective, it is best to start off with care that is team based, where one piece of it is having visits in the home with what we call household-centered care where we go into the home with a team of medical students, pharmacy students, nursing and even health law students. How can we do this team-based care in a way that is much more holistic and works with the social problems?
Gordon: That is just an incredible step forward, I think, in terms of medical education and how it can holistically serve. I think about where this would be interesting for hospitals and health systems insurers and others throughout the country. I know there’s still a significant amount of uncompensated care delivery and concerns about how that works. And when I heard you guys describe the Patient Care Intervention Center, what you were doing there, it resonated with me in terms of taking the high road and doing the right thing for very stressed individuals in a way that serves them and helps serve the greater good as well. I’m thrilled to hear that this becomes part of the education cycle. That’s terrific. Thank you both, Kallol and David for your time and for the great work that you’re doing. Do you have any parting thoughts?
David: We’re really excited you think this is of interest. We’ve been working for four years trying to convince people that social determinants are the fundamental missing element in medical care today and so we’re really excited about your interest in where it may go.
Gordon: Well, Dave and Kallol, thank you very much for your time.
Kallol: Thank you very much.
David: Thank you.
Gordon: For Inside Angle, this is Gordon Moore. You can find more podcast episodes at www.3mhisinsideangle.com.