Hi everyone, my name is Brad DeCiano. I am an associate professor in the Department of Physical Medicine and Rehabilitation at the University of Pittsburgh Medical Center, and I'm also director of the UPMC Adult Spina Bifida Clinic. And today we're going to talk about how to develop and implement a population health model. I'm going to use spina bifida as an example of a patient population that that can be done in. And then we're also going to talk about how to leverage data gathered from populations such as those with spina bifida to do research. More specific objectives for today's lecture are, number one, we're going to review some literature just very briefly on population health models relevant to spina bifida, and we'll hopefully by the end of the lecture be able to list strategies that can be used to improve patient outcomes in physiatry practices. And also, I'm going to touch on how we can leverage information technology to supplement population health programs. A little bit more about me. I'm also associate medical director at the Spina Bifida Association, and I serve as the chair on the Professional Advisory Council. We also will be releasing the new updated spina bifida health care guidelines this year, hopefully in May, and I served on the steering committee, and I'm also PI on the National Spina Bifida Patient Registry. I'll talk a little bit about that. We have a site, one of the sites, and at the University of Pittsburgh. I have specific conflict of interest disclosures listed within the presentation that are relevant to each study that I talk about. Just as some background, there are more than about 170,000 people in the United States who have spina bifida. Most of them are actually adults. Spina bifida is no longer a pediatric disability. We used to think of it purely as a condition that affects children, but now patients are living pretty long lives into their seventh or eighth decades. In fact, I have patients in my clinic who are in their 80s, and we call spina bifida the most common permanently disabling birth defect, and it occurs in about 7 out of 10,000 births. I'm going to first touch on two manuscripts that we published where we looked at national data sets on hospitalizations of adults with spina bifida, and to summarize those two studies, adults with spina bifida are on average hospitalized each person three to four times per year, and they have higher 30-day readmission rates than the general population, readmission rates up to about 50%. We see a lot of potentially preventable conditions in this population, which are causes for hospitalizations and also semi-atrogenic conditions. The most common being urinary tract infections, wounds, infections from things like catheters for the bladder, and also sepsis from those conditions, and these potentially preventable conditions account for more than one-third of hospitalizations of adults, and in just the two-year time period that we looked at in this study, potentially preventable conditions in this population are just related to hospitalizations, cost $364 million, and we're responsible for over 35% of the deaths that occurred in the hospital, and these potentially preventable conditions also occur most often under age 50 in this population, and we think that that's because during the time of transition from pediatric to adult care, a lot of patients lose access to health care services, and this is when they're most vulnerable. So there are different types of models of care in this population. There are clinics that are solely set up to take care of adults, like mine. These types of clinics are fairly rare. There's only a few in the United States, about five or six nationally, and some of these are medical home-type models. In a number of other clinics, adults continue to be seen in multidisciplinary clinics that are set up for children, and in those cases, adults continue to see specialists who are primarily pediatric-trained. That can be good because they're, they have good access to people that understand how to treat spina bifida, but the downside is that there are then adult-onset issues that come up that those pediatric specialists may not know how to treat. In the majority of cases, though, adults are generally discharged from their pediatric clinic around 18 to 20 years old, and they're not treated for spina bifida. In the majority of cases, though, adults are generally discharged from their pediatric clinic around 18 to 26 years of age, and they must find their own providers in the community, and guess what? There aren't any providers in the community because of lack of training. Most of the specialists who are trained to treat spina bifida are pediatric, and therefore generally already affiliated with the pediatric clinic that they were discharged from. So, this is potentially a good opportunity for physiatrists. I think that given our training, especially in the field of spinal cord injury medicine, it gives us the opportunity to potentially be a specialty medical home-type model for these adult patients who otherwise wouldn't have any care providers. There are also some transition models that are popular throughout the country, techniques and recommendations for transitioning patients from pediatric to adult care. Probably the most commonly used is Got Transition, and that can be a good resource if you're looking to set up a program to facilitate transition. It goes over how to develop policies, how to track patients, how to ensure seamless transitions, and has a pretty robust tool set of a lot of things that you can use for education and training of the patients. I want to just touch on a clinic survey that we did in the state of Pennsylvania. I partnered with the Pennsylvania Medical Home Initiative, and we sent out a survey to providers who are parts of medical homes throughout Pennsylvania. We did the survey in 2016. It was an online survey, and we asked a lot of questions about how comfortable people are in treating patients with spina bifida and what they knew in terms of resources that they used when they cared for these patients. We got responses back from 67 clinics, and the respondents for those clinics were pretty variable. They ranged from physicians to nurse practitioners and physician's assistants, nurses, and even other types of specialists, and the practices represented a wide array of private, academic, and other type of practices, and probably the most shocking thing that came about from the survey was that when we asked, does your practice follow or refer to the guidelines for spina bifida health care services, only 14 percent said yes. The remainder either said that they do not follow or refer to them, they're not sure if they do or not, or that they weren't familiar with that resource at all, and this is pretty representative, I think, of the medical community that there's not a lot of education out there about what is available and what are the guidelines to take care of these types of patients. So, I want to talk a little bit about my particular clinic at UPMC. Our clinic has outside funding from the Pennsylvania Department of Health, specifically the Bureau of Family Health, and this grant provides some care coordination dollars that allows us to provide the effort for the staff in our clinic, and I'll go over that in a second, but it also allows us to provide patient assistance funds, and these funds can be distributed to patients who meet certain criteria according to their poverty level, and the funds then can be used for things that their medical insurance doesn't cover. So, for example, shower chairs or mirrors to check their skin for a skin breakdown, or it may be able to be covered, be used to cover co-pays for medical visits, or even a babysitter to allow somebody to overcome a barrier to come in for their medical care. Our care team is made up of myself. We also have a physician's assistant. We have an RN who is trained in spinal cord injury care, spinal cord injury care. We also have an advanced practice nurse practitioner who has training in pediatrics, but her role in our clinic is to serve as a wellness nurse. We also have a dietician. We have a certified personal trainer who is specially certified to teach adaptive sports, and he actually himself is a Paralympic athlete. He won a gold medal in the Sochi Paralympics for sled hockey. We have a research assistant, and we also have a volunteer, and three total people out of our team actually have spina bifida. Our clinic is part of the National Spina Bifida Patient Registry, which I mentioned earlier. This is a CDC-funded program. There are over 20 spina bifida clinics in the country that participate in this registry, and some clinics also self-fund themselves to participate, and what this registry does is it allows us to enroll patients who are interested in being part of the registry and then pool their data together so that we can do studies that otherwise wouldn't be possible. Until we had the registry, most studies on spina bifida, especially on adults, were quite small. The registry is completely electronic, so we actually enter the information into an electronic database, but these are just some examples of some paper-based forms to give you an idea of what type of data that we do collect. There are over 20 medical areas that we collect data within. They range from demographics to topics like mobility, urologic care, bowel management, etc. The data is all housed. It is kept electronically at our within our own purview at each site so that only the people who have access to the data at each site are the people that know the patients. The CDC actually only receives de-identified data, and so patient information is protected. We published one of the first studies on mobility in such a large population of spina bifida, and it was the first mobility study to come out of the patient registry. And what this paper found by pooling data of thousands of patients together was that ambulation status seems to be predicted by a number of different variables. Those who seem to ambulate over time, those who seem to have preserved ambulation over time tend to be the patients with no shunt. They tend to be patients that have lower neurologic motor level, and they also tend to be patients that have never had surgery for contractures at the hip or knee. On the left, you can see some graphs. The one on the top is patients with myelomeningocele, the most severe subtype of spina bifida, and we stratified them by the level of ambulation that they are able to achieve and also the lesion level on their back in terms of their motor function. And this is the first time that we're really able to see large populations like this and look at stratification across these types of groups and also different subtypes. The lower graph looks at the non-myelomeningocele subtypes. Very few studies to date have really broken spina bifida down by the different subtypes that exist. So next, I want to talk about a wellness program that we implemented in our spina bifida clinic. We are part of an integrated delivery system, and part of our delivery system is UPMC Health Plan, which is our insurance provider, and they helped fund this particular program. It was also funded by a small pilot program that we have locally, and when I presented some of the national data to the health plan, they were pretty astounded at the cost that these patients have on the insurer aspect of their care. So that is what the impetus was for this study. And many of you may be familiar with the triple aim, and that was the concept behind this study. We wanted to see if we could decrease health care costs, improve population health, and we defined that in several ways, and we also wanted to enhance the patient's experience of their care. So our objective was to determine whether an evidence-based wellness program would have an effect on the triple aim, and our design was a non-randomized, non-controlled cohort study. Each participant enrolled for a total of two years. We ran the actual study over a three-year period. We just staggered enrollment, and the setting was our outpatient spina bifida clinic, both for spina bifida and also our outpatient spinal cord injury clinic. We called it the UPMC Center for Wellness for individuals with spina bifida and spinal cord injuries, and it had several components. One component is the care coordination component, and this was a free benefit that patients got through UPMC health plan if they were a member, and to receive those benefits, they simply had to sign a contract with that health plan to get the benefits. We also had a research component. They had to sign a research consent, and this was all under IRB, and within the research project, if they consented to having their data enrolled in our research study and they made progress throughout the wellness program, they received incentives along the way, so they could receive up to $25 in gift cards four times throughout the study for a total of $100. So our hypotheses were, again, focused around the triple aim. Our first hypothesis that we would improve function, mood, and quality of life, and those were our health measures. We also expected we would increase experience of care variables, and that we would have an effect on cost. We actually hypothesized that in year one, cost would increase. That's typically what you see in these kinds of programs when you start to uncover medical conditions that hadn't been treated yet, but we were hoping that by the end of the study we would have a positive return on investment and save money. What was the wellness program? Our actual intervention revolved around having a mobile nurse coordinator, and he actually met with participants in their homes, and a big focus of what he did was goal setting. So we had him sit down with patients and devise goals. They had to come up with short-term goals, long-term goals, and then maintenance goals that they felt that they would have to meet in order to maintain good health over time. He also provided education on preventive care measures, mostly targeted to those preventable conditions that we mentioned earlier, and the self-management program was evidence-based. We used clinical practice guidelines, and we also used validated screening tools, and there were ten different modules. Each of the different modules were topic areas of health, health care factors that we thought that we could positively impact with this wellness program, and these were also areas where we felt that there was potential to improve hospitalizations, ER visits, and also have an impact on cost. So those ten modules are listed here. Each module had three different areas. One was a patient education component, so we provided standardized educational material to each of the patients. So, for example, for skin integrity, we provided a lot of information about the importance of pressure relief maneuvers and how to protect the skin. We also had prevention algorithms, so we had certain techniques that we used to try to prevent skin breakdown, and then if a wound would occur, we had an algorithm for how we would handle that in terms of treatment. We screened 128 referrals from a UPMC health plan. Many of these patients screened out because they did not meet the screening criteria. One of the criteria was that they had to be a patient in our clinic naturally in order to be in the study, so a lot of them failed to meet criteria for that, but we did allow them to enroll in the study if they wanted to become a patient in our practice. We consented 69 individuals. Four were excluded after consenting due to failure to meet inclusion criteria or withdrawing, and then 65 completed their first baseline visit, and 52 completed the program in whole. We had 48 patients with spina bifida, 17 with spinal cord injury, and the mean age was about 39. At the end of the two-year period, all of our health and experience of care outcome measures significantly improved. The health outcome measures included the chart short-form, which is a measure of function, the Beck depression inventory, which is a measure of depression symptoms, the World Health Organization quality of life brief instrument, which measures quality of life. We also asked patients simply to rate their care and their care as on a Likert scale. We also measured experience of care using a validated instrument called the PASIC, and we gave them a quiz, which was basically information that they learned about their condition and preventive care. Our experience of care outcome measures actually also increased at year one, whereas all the measures also increased at year two, and what's really notable here is that by the end of even the two-year period or the 18-month period, patients were scoring almost 98% on their knowledge quizzes. Another measure that we looked at was utilization. We looked at how often patients were hospitalized or went to the emergency room. We did see decreases in these events. For example, the inpatient admission, for any reason, went from 50 at baseline to 54 and then down to 32 by end of year two. That was not a significant change. We did see a significant decrease in emergency room visits for any reason, starting at baseline at 112 events down to 101 and then down to only 53. That was pretty impressive, and what we found was we were more likely to affect emergency room visits for all reasons, not just preventable conditions, because as it turned out, patients were utilizing the emergency room for all types of care, and when we explained to them and provided them a way for them to communicate well with their providers, they realized that there were appropriate ways to use the health care system. So basically, we shifted a lot of care that was being done in the emergency room over to a more appropriate outpatient setting. In terms of cost, the overall cost of the program was a little over $150,000. That was for the entire three-year period. That included all the costs for the medical care. That includes the calculation of cost savings, and it also includes the cost of the program, so the cost of the nurse's salary. So if you divide the $150,000 over the three years, you can see that the cost overall would be less than what it would cost to have a full-time nurse. We did see an increase in cost related to preventable conditions in year one, as we expected, but by the end of year two, we did not see a positive return on investment like we expected. But we did see some trends in decreasing costs, and we think that if patients were enrolled for a longer period of time, we might see a significant reduction. But overall, we think that given the positive outcomes in the other domains of the triple aim, really, the cost of the program really was worthwhile. Now, I want to transition to another project where we developed a mobile health platform to help augment the type of services that we provided in the wellness program. What we found was one wellness nurse can't treat all patients in all geographic areas. It's really hard to physically visit all of those patients if they live far away. So, we wanted to see if we could develop a mobile health system that could help support the program, especially because a lot of our patients are very tech savvy. So, my disclosures here is that this funding was, for this research, was gathered over several years. We have a lot of funding from NIDILRR, from the Verizon Foundation, from the Craig H. Nielsen Foundation, and NIH. Also, I am an inventor of the software that I'm going to be discussing, but I don't have any financial interest in that technology. Our objective in the first clinical pilot study that we did was really to just determine the feasibility of using this system in patients with myelomeningocele. So, we had a one-year randomized controlled trial where the intervention group received a mobile health system, and the control group just received standard care in our Spina Bifida Clinic. We hypothesized that the intervention group would reach certain benchmarks for use of the system, that we would see fewer preventable medical events and lower utilization of hospital services and ER visits, and that we would see improved psychosocial and functional outcomes. We also thought that those who used the system more, the high utilizers, would have greater improvements in their self-management skill. This was what the system looked like when it was version 1.0, and that's what we used for this particular study. On the screen, you can see that the mobile health app that the patient opens has several modules, and we could customize the modules that the patient sees based on their own needs. The modules that we had available were for bowel, bladder management, skin integrity, mood, and medication management, and those modules had a number of different features. One is that they would provide reminders to the patients to do certain activities. It would provide an ability for the patient to report symptoms back to us. There were flow sheet functionalities that allowed the patients to track certain things, like the amount of incontinence they were having, so that we could see whether they were responding to treatment. There were surveys to gather symptoms, such as depression symptoms. The system would also allow you to upload photos for things like wounds, and then the patient could also securely message the clinician back and forth. This is an example of how an alert might appear. The patient opens their app. They see their suite of modules. A reminder appears to tell them to check their skin, just as a routine self-management task. If the patient notices that they do have a new wound, they can open the skin care module, find an anatomic body map, and click on the map to show us where the wound is. They can upload a photo and send information about the wound to us. On the lower part of the screen, you see the clinician interface. So, this is a web-based portal. The clinician has a dashboard, which triages all the patients and all the modules that the patients are using, and there's little color indicators to tell us whether the patients are having problems. When everything is okay, everything's green. When the problem comes in through the portal, the indicator turns red, and that gives the clinician the ability to know that there's a major problem. Also shown is an example of what a picture of a wound would look like in the portal, and the clinician can follow the wound over time to see if it's responding to treatment. Our first clinical trial was a small, randomized clinical trial. It was a small cohort. Our intervention group had 13 participants, and our control group had 10. This graph shows, after a one-year period of being enrolled in the study, the intervention group that used the mobile health app system had decreases in a lot of the preventable conditions and utilization measures that we hoped to impact, and the control group either stayed the same or had an immune response to the utilization measures that we hoped to impact, and the control group either stayed the same or had an increase in those negative outcomes. These were trends only. They weren't statistically significant, but we thought it was impressive that the trend occurred across all categories. The other take-home points from the study was that the patients actually did use the features within our system over a one-year period. We were really curious to see if the novelty would wear off and if patients would stop using it, but they stayed engaged over time. We also did find that the high utilizers had significant improvement in their self-management skill. The features that patients used most consistently were the reminders to do their bowel program, the functionalities that allowed them to manage their medications, uploading photos, sending secure messages, and filling out mood surveys. We also, like I mentioned, saw a trend toward fewer medical events and lower utilization, although that was not statistically significant. And we really saw less reliance on caregivers. We did measure how many hours they were using in terms of caregiver assistance, and those that used the system more had less reliance on their caregivers. In parallel, we recently conducted a similar study in a cohort of 40 patients with spinal cord injury. We're in the process of publishing that work, but the main findings from that study was that we did see a statistically significant reduction in urinary tract infections. The number needed to treat was just over two, meaning that for every two people that used the app, we prevented one UTI. We also saw significant trends, I'm sorry, trends that were not significant in reductions in emergency room visits and hospitalizations that were due to wounds and UTIs. So now we've redeveloped the system based on a lot of what we learned from these studies. So we're now on to version 2.0. We've also done a wide variety of accessibility and usability trials and focus groups. We've had about 200 people total that have participated in all of our different types of studies who have given us a lot of feedback and helped us to redesign the software. And we're planning on conducting a new randomized controlled trial on version 2.0 this upcoming summer. This is an example of what the version 2 looks like. You can see that we now have an interface for caregivers. So a patient can give access to a caregiver to share certain types of data that they decide. The caregiver can interact with the patient and also the clinician. And this could be for a family caregiver or it could be for a professional caregiver. We've also developed a lot of other features besides caregiver support. We have a personal health record, which is quite extensive and allows them to put in their a lot of their medical history, accessibility needs. We also have functionality to allow them goal setting since we found that to be so helpful in our wellness program. They can enter goals and keep track of how well they're doing and then receive feedback from their caregiver or clinician on how well they're doing. We have a reward system. So the caregiver and clinician can give thumbs up to the patients to tell them that they're doing a good job and for completing their self-management skills, they can earn points and those points can then be used to enhance the app with certain fun features like changing the way the background looks or adding moving images. We also have a very extensive educational database and we can deploy educational material to patients that is customized so that they only get information that's relevant to them and also when we want to deploy that information. We've customized the information in the modules to be very specific for people with disabilities. So for example, we have videos of exercises and these have been customized to be adaptive so that people with disabilities can do them. And that's all for today.

spina bifida

population health model

hospitalizations

wellness program

transitioning

mobile health platform

research