How advanced analytics can help social determinants of health [PODCAST]

Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes

“Advanced analytics can provide the insights needed to understand social determinants and help develop interventions that assist patients in overcoming some of the challenges and adverse environmental and social factors that are barriers to healthier behavior. Moreover, analytics are needed to support the team-based approach to care delivery. With advanced analytics bringing together data on clinical and social needs, as well as health behaviors, providers and payors can enable effective care coordination and successfully implement value-based care models.”

Mike Dulin is a family physician and health care executive.

He shares his story and discusses his KevinMD article, “How advanced analytics can help social determinants of health.”

Reflect and earn 1.0 AMA PRA Category 1 CME for this episode.

---

Did you enjoy today’s episode?

Please click here to leave a review for The Podcast by KevinMD. Subscribe on your favorite podcast app to get notified when a new episode comes out!

Do you know someone who might enjoy this episode? Share this episode to anyone who wants to hear health care stories filled with information, insight, and inspiration.

Hosted by Kevin Pho, MD, The Podcast by KevinMD shares the stories of the many who intersect with our health care system but are rarely heard from.

Transcript

Kevin Pho: Hi, and welcome to the show, where we share the stories of the many who intersect with our health care system but are rarely heard from. My name is Kevin Pho, founder and editor of KevinMD. Today on the show we have Mike Dulin. He is a family physician and health care executive. He wrote the KevinMD article “How advanced analytics can help social determinants of health.” Mike, welcome to the show.

Mike Dulin: Thanks very much, Kevin. I’m really excited to be here. I’ve been a fan for a long time, so I’m excited to have made it onto the podcast.

Kevin Pho: Oh, thank you so much for listening all these years. We’ll get into your article in a little bit, but first off, can you share your story and journey to where you are today?

Mike Dulin: Sure. I have a little bit of an atypical journey into medicine. I actually started my career as an engineer, and I worked as a data scientist for a firm that manufactured computer chips for a number of years. I went from that (I didn’t have a lot of time to volunteer and do a lot of health care work) directly into medicine, and I was just shocked. In the computer chip manufacturing world, we used data every day to drive improved performance, and this was in the early ’90s. I was really shocked when I went into health care that we didn’t do the same thing.

So that’s really been my passion as a health care provider over the past 20-plus years: to try to use data and technology to advance the way that we deliver care, and particularly to make our jobs easier as clinicians.

Kevin Pho: Over the past year, of course, as physicians we have been inundated with more and more data. So tell me about the evolution over the last few years, about how that intersection with data and patient care has come to fruition.

Mike Dulin: Right. Well, I think the area of my article outlines part of my passion in this domain too, which is the concept of social determinants of health: how we start to understand them and use data to draw around social determinants to drive that improved performance. When I was training, in addition to not doing a very good job of using data and analytics, we very much didn’t understand how social determinants played a role in health outcomes.

Now, over the past decade, our knowledge in this domain has really deepened, where we understand that the majority of health outcomes are actually driven by health behaviors that are influenced by social determinants. The clinical delivery systems themselves only influence a small fraction of those behaviors. So, helping to use data and technology to both unearth and understand those social determinants and bring them forward, so that as health care providers we can understand them and make sure that we tailor our interventions, particularly behavioral interventions, in a way that helps us to improve patient outcomes.

Kevin Pho: And like you said, you write more about this in your KevinMD article titled “How advanced analytics can help social determinants of health.” Can you summarize the article and go beyond what you just said? Give examples, and show how these analytics can identify and perhaps help solve some of these social determinants.

Mike Dulin: Sure. In the article, I really tried to focus on some of the key behaviors that influence or underlie preventable health outcomes. Number one is tobacco use. We’ve known that for a long time. As providers, we have algorithms in place to help us understand if a patient smokes or not and to connect them to intervention. We have lots of different interventions in our pocket, like motivational interviewing and shared decision-making, the five A’s, those types of things. But one of the issues is that the societal pressures trump any of those interventions that we have, by far. So the number one lever is actually to do something like increase cost, so increase taxes. The other is to start to think about social behaviors, to make it more difficult, I guess, in some ways to smoke, and so social circumstances: if you could have a nonsmoking airport or a nonsmoking park.

The same holds true for behaviors underlying obesity, so the ability to exercise and the ability to access healthy foods. I think maybe this is an area of even more opportunity for us as health care providers. Often we can say, “All right, you need to change your behavior. You’re developing a chronic disease related to lack of activity or consumption of unhealthy foods.” But those changes in behaviors can be extremely difficult for patients, particularly if they live in a food desert, or they live in a community where there’s no access to parks and those types of things. So, understanding that first. If we try to apply an intervention as a clinician and say, “OK, what you need to do is exercise more. Here’s an exercise prescription. Go out and do it,” but then the patient doesn’t have a safe place where they can go and get exercise, or they’re unable to access the healthy foods, it can be extremely difficult, and actually end up causing worsening stress and difficulty for that patient rather than helping them to improve their outcomes.

I touched on this to some degree too, but I think these social determinants also play a role in the current pandemic. People use their social circumstances and their social networks to get information. So there may be misinformation within their social network, and that may underlie some of these behaviors that we’re seeing, like vaccine hesitancy or resistance to getting vaccines against COVID, or also an ability to change behavior, social distancing, masking, those types of things, because of that misinformation in the social network.

Kevin Pho: I’m a primary care internal medicine physician. So tell me how I can use, or my team specifically can use, analytics to help change some of my medical decision-making or influence what I do in the exam room. Can you give some real-world examples or stories?

Mike Dulin: Part of the idea of this next generation of primary care delivery is based on the idea that you have team-based care. So it’s not just an individual physician providing all of the services; rather, we have a full team that might include a social worker or pharmacist, others that can help with behavior change to a further extent than us, or a care manager or care navigator. But I think underlying that, you have to have a good analytic system to help you understand who has the ball at what time. How do you work together as a team? How do you interact with the patient? And then the other component of that is how do you tailor the intervention appropriately for that individual who’s coming into the clinic at that time, or who needs outreach of some sort, and how do you bring into context these social determinants of health?

The reason that the analytics are so important is because the data is so complex. You probably have seen this, Kevin. I got dumped data from, like, Apple Watches and stuff like that, right? And to me, as the physician, I had to be the point of integration, the person who interpreted all that data, brought it together, and helped to design a clinical intervention based on it. If you can imagine, we have all that data, patient-generated data, but now we also have all the social data. The complexity is beyond what an individual or even a team can do. We really need these analytic systems in place to help us understand what is the most impactful thing we can do and how we can tailor, particularly, this behavioral intervention.

I hesitate for a minute, and I’d say I think there are two components of medicine that we need to think about here. There’s clinical delivery that needs to be standardized, right? We should do the same things in certain conditions across the board. Preventative care: Everyone needs their colonoscopy at a certain age or time. Someone who’s admitted to the hospital needs DVT prophylaxis. Those things should be standardized and delivered in the same way across the board to help prevent health disparities. But the flip side of that are these behavioral interventions, where we really need to understand the patient’s context, where they’re coming from, and their social circumstances, so that we can appropriately tailor the behavioral interventions in a way that they’re more likely to be successful.

Kevin Pho: Are you thinking, as part of this team-based approach, which a lot of primary care physicians, frankly, are moving towards, that these interventions are going to be done by members of the team, perhaps a social worker or perhaps nurse outreach, and not necessarily in the exam room?

Mike Dulin: Yeah, I think absolutely. Again, I think the exam room may be designed best for clinical care, and we need to contextualize that in some way. If we’re providing clinical care that may not be accessible, like we’re prescribing a medicine that’s unaffordable, we need to understand that and pivot to make sure that we provide clinical services that do contextualize some of these social components. Just the clinical delivery itself as well: Transportation barriers can make it extremely hard for people to come in and get services. I used to work at a clinic designed to care for Medicare, Medicaid, and uninsured patients, but the clinic was two bus stops away from where most of the patients lived, so it became very difficult for them to get into clinical services.

So yes, I think that’s a key component: to think through how we provide asynchronous care. How do we care for patients who aren’t in front of us at that time? How do we reach out into the community and break down this wall of the clinic around us in a way that makes it easier for patients to get care when they need it? My feeling as a primary care physician is that it is vital to get these types of primary care services. If we just have a nine-to-five, “you have to come to my clinic to get those types of services,” they are inaccessible for a large component of the population without them going through these barriers: getting daycare, two buses, thinking through how they get into care. That is one of the silver linings to the pandemic, as we have finally seen some of these care delivery methods, like telehealth and asynchronous care delivery, start to get a foothold, take off, and actually be shown to be effective.

Kevin Pho: I think one thing that we’ve seen, especially during the pandemic, is how much staffing can certainly play a role in primary care offices. We have the Great Resignation, and in my own primary care office, we really have a skeleton crew right now. I think this is replicated in primary care offices across the country. Whenever doctors hear that we need to do more with data, it always feels like, “Oh, we have to do more. How are we going to get the resources to analyze this data? What’s the cost impact of something like this?” So can you speak about how a typical primary care office can implement some of these advanced analytics, keeping in mind things like cost and staffing?

Mike Dulin: Yeah, it’s tough right now. I don’t know that we have the solutions at our fingers now. I think a lot of this is thinking through how we have input into these new solutions that need to be developed and implemented. And I agree with you. I think for years, what we’ve done. And so I was one of the optimists. I came from the background of being an engineer, and I saw electronic health records as being a major solution to help us to be more efficient and effective and provide higher quality of care. Like yourself, I got to go through the tough times of late-night pizza parties entering clinical data into electronic health records and those types of things, three days of training to make the electronic health record work, kind of wrapping the provider’s behavior and training us to work around a lot of the limitations in the electronic systems that we’re using.

I think we need to change that paradigm, right? We need to think of these tools as ways of making us more effective and making our work easier. And as you mentioned, particularly now, there’s so much stress on the health care system. There’s so much difficulty, and it’s really hard work, and the tools need to work with us and make us more effective. I mentioned before the example of getting all this data in my inbox. You get faxes, you get PDFs emailed to you, you get things that are really data that’s really hard to penetrate, to understand, to bring forward, and to improve clinical effectiveness.

But the tools are there. These newer tools, the artificial intelligence and machine learning tools, can start to drill in. They can use natural language processing to start to pull out the important facts and data, and really, they can be built in a way that helps to make us more effective and efficient. They’re not going to replace providers by any means, but they can help to make us more efficient and make our jobs easier, and that’s what we should be pushing for.

Kevin Pho: We’re talking to Mike Dulin. He’s a family physician and health care executive, and he wrote the KevinMD article “How advanced analytics can help social determinants of health.” Mike, can you give us a road map for the next couple of years? How do you expect that intersection between data analytics and primary care to evolve?

Mike Dulin: Well, I think one of the first things: We’ve gotten a lot of good information now about our limitations around data, particularly in the space of social determinants, and also around some of the tools. In one of my past roles, I helped implement an AI algorithm in the acute care space to look at readmissions and help improve the way we deliver care in that space. One of my regrets when I think back on that is that we had a lot of excitement about the tool but didn’t really think through some of the bias in the data itself. Now we have really good publications and knowledge about how data can be biased and how it can be influenced by access to care, and certain groups and populations may have data that’s lower quality, or that may even be incorrect in their medical records. So we need to start to focus there: How do we collect the data in a way that helps to reduce the bias and helps us to really understand the social determinants of health?

I think the second thing is that we need to be much more patient-centered. We need to be transparent with patients, help them to understand how and why we’re using the data, and give them control over how that data is used. I think some of the new technologies that are emerging, like blockchain, will allow that to happen, so individuals can have control in terms of where their data is used and how their data is used, and see who’s accessing the data over time.

I think a lot of the tools, we see them on the consumer side all the time. We use tools like Amazon and Uber and Facebook that are able to manage huge amounts of data rapidly and provide information that’s tailored to the individual who’s using the tool at that time. Those same sorts of tools can be used in health care. We just need to first understand what the potential biases would be when they’re deployed, and make sure that we don’t replicate interventions that could create worsening disparities in health care, and then, again, deploy them in a way that’s patient-centered and provider-centric, so that we’re not retraining our physicians on how to use yet another tool that’s layered on all these other tools that we’ve learned over time. So I think the technology is there, and the knowledge about how to deploy it is there. We just need to be engaged and aware as those things develop over time.

Kevin Pho: And my final question: What are some of your take-home messages that you want to leave with the KevinMD audience?

Mike Dulin: Yeah, I think the key is that social determinants are extremely important. I think we know that, but taking that extra time to understand social determinants and how they can play a role in outcomes up front can actually save time in the long run. It’s kind of like asking that second open-ended question when you engage with patients. Getting the information about the social determinants can take more time, and again, maybe that’s part of the role of one of the team members on the primary care team.

I think that AI is here. The tools are here, they’re developing, and we need to not be scared of them and not think of them as replacements, but rather as tools that help us to become more efficient. As you mentioned, the primary care workforce is stressed. It’s under difficulty now. We need to take care of our primary care physicians and make sure that they have the tools that can make them more efficient and allow them to work in the space where they want to work, engaging patients and improving health outcomes. So I think those are the main takeaways.

I guess the last one is that we also need to understand the drivers of health disparities. It’s extremely important that we not build new tools on top of old systems that have created those health disparities and just make them worse, and I really do worry about that. We’ve already created systems in place that may worsen the current outcomes that we have. I think every time we look at outcomes across different groups or populations, we unfortunately see the disparities that exist in the U.S. health care system, and we need to create tools that don’t augment those in any way, but rather address them and help us to provide more equitable care that improves health outcomes for everybody.

Kevin Pho: Mike, thank you so much for sharing your time and insight, and thanks again for being on the show.

Mike Dulin: Thanks very much, Kevin. I really appreciate the opportunity.

Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes