How data science is shaping the evolution of healthcare systems
How data science is shaping the evolution of healthcare systems7 July 2021 | 13min
Data science will lead to a shift in healthcare business models, not through mergers and acquisitions, but rather through agglomerations, confederations and partnerships
The evolution of decision making tools as a result of data science will lead to a transformation in healthcare, with not only a focus on preventative and holistic care along the patient journey, but also more empowered patients
There will be many technological leaps in the next few years around data-driven healthcare and it is up to us within the healthcare system to transform it to its ideal state with an entrepreneurial lens
Dr. Atul Butte brings his experience and expertise in data science and healthcare to new heights. In the first part of the interview, Dr. Butte gave us an overview of the role of data science in today’s healthcare industry and how it will be used in medicine and healthcare.
In part two, Dr. Butte continues to share his insights around data science’s role in the evolution of healthcare systems. He speaks to the evolution of business models and decision-making tools as we work to provide better patient outcomes, a key measure of success within healthcare systems.
Data and the future of healthcare systems
HT: To make data useful to better patient outcomes, do you see the need for a new business model to access and use shared data that’s required to improve the outcomes of patients?
Atul Butte: I believe we’re going to see a lot of changes in business models in the entire healthcare enterprise over the next 5 to 10 years. Certainly, we’ve seen many changes in the last 5 to 10 years. There is still a push towards what is called accountable care in the United States, where healthcare systems, and delivery organizations, are going to be more responsible for the finances of how that healthcare is paid for.
For example, health systems would be paid per patient per month to take care of that patient, no matter what happens or what is needed. Obviously we’re going to need a lot of data to understand what is the right cost that we should be putting onto plans like that and what is the right care that should be delivered, but the push towards accountable care is not a matter of if it’s going to happen, it’s a matter of when it’s going to happen.
In the United States, more than 3 trillion dollars a year is spent on healthcare.1 A lot of it comes from one major payer, that’s the centers for Medicare and Medicaid. For elderly patients and the very poor, the government can cover healthcare and then the rest is private. We have the veteran’s system and then we have the defense department and then we have larger employers that cover their employees. Despite all the variety of ways that we pay for healthcare, there’s a general understanding that too much is being spent on healthcare and it’s up to someone to figure out how to spend less. We’re absolutely going to get to new business models and given what I’ve seen, more organizations are starting to come together to execute on those business models.
Changes in business models
Atul Butte: I’m not going to use a strict phrase like mergers and acquisitions, as these terms are more of a financial kind of commitment. Looser ones like agglomerations are going to make a little bit more sense, where health systems might agree to partner, maybe a larger one with several smaller ones, or some peer organizations working together. We’re going to see more of those business models used, and data sharing is going to be needed with those business models in a way that will enable them to be implemented.
For example, one large health system partnering with another large health system might decide that only one of them will do a certain set of procedures – for example, liver transplants – and the other might not. They might use data to choose which one does and which one doesn’t, perhaps to cut their costs, or to control expenses and quality of the care being delivered.
Data is going to help enable these new business models in the healthcare enterprise and the new business model is going to be about gaining scale through agglomerations, putting these confederations of like-minded institutions together, which are going to agree not to compete with each other.
The future of informed decision-making for healthcare professionals
HT: Given your expertise, both as a physician and data scientist, how does the future for physicians look like in terms of their responsibility and training as their professions will depend more and more on interpreting and using data?
Atul Butte: Medical education and medical training is constantly in flux and data is going to be driving some of the future lifelong learning that’s needed. Now here, when I think of data, I’m thinking a little bit more of a loose definition. Physicians are constantly bombarded with publications in top tier journals, such as JAMA and the New England Journal of Medicine.
Some of those publications are harder to interpret than others and it is our duty as medical schools to make sure physicians of the future know how to interpret data in publications, so that they can understand what is the right thing to do for patients. What type of trial was conducted? What were the actual results of these clinical trials? But the “bombardment” of publications on doctors is just going to continue to grow exponentially.
Broader than that, physicians of the future are going to have to better understand and explain risk. What could happen next if one choice is taken versus another? How do we communicate that risk to patients? For example, “if you don’t stop smoking, this might happen and then that might happen.” Then, based on all of that, how do we come up with the right next thing to do for a patient? The answer is precision medicine.
Physicians are going to be guided more and more with digital decision support tools, or decision support systems, because there are so many data elements now, computational tools are going to be needed to really help guide the physicians to figuring out the right choice from many.
Those decision support systems are going to be trained using data. For example, for this one patient in this exact condition, who were the last 100 patients like this patient? What did we do? What worked and didn’t work? We run clinical trials where we learn in randomized controlled trials what drugs might work or not, but we don’t have trials for everything. So, we’re going to be increasingly learning from our own experience, digitally captured from our own patients through their data. That’s what’s going to drive future decisions.
Understanding the benefits realized along the patient journey
Atul Butte: Even broader than all of that, one of the things I’ve noticed is that, after we get drugs approved or medical devices approved, physicians start to use these drugs, or they start to use these medical devices, but in general, physicians and health systems really never bother to study whether their patients are benefiting from these drugs or devices, in general. Yes, in some cases we might run post-approval studies or put in data to registries, but in general we’re really never seeing whether our patients are benefiting from these drugs or devices, in the same way patients did in the original trials.
That’s going to have to change. It should be our duty as physicians, as health systems, to ensure our patients are benefiting from these drugs and devices and procedures. The electronic health record system of the future is going to have to help physicians study their own patients to better understand: are the decisions they’re making actually improving care for their patients?
It’s a data-oriented world now and the future role of the physician is to be open to hearing and seeing and understanding all of that data.
Decision support tools for both physicians and patients
HT: As technology and artificial intelligence (AI) continue to advance decision support tools, how do you see these tools being used within the healthcare system by physicians?
Atul Butte: In general, we are still in the early days of how physicians are going to interrelate with electronic health record systems, in particular using decision support tools. First of all, ideal decision support tools are going to be built into the EHR system they’re using. If a physician has to switch to another tool or to pick up their phone when they’re using the computer, that’s going to make it much harder to actually get good uptake of those tools. In addition, decision support tools are going to need to get and gather as much data automatically as possible. For a physician to retype anything into those tools obviously makes it much harder to actually use the tool.
In general, physicians are still using electronic health record systems to document the care that they’re delivering to a patient. In general, that’s using the EHR after the fact, after the encounter. It might be during the encounter, but it’s usually as the encounter is ending. We have to start getting into a mode where physicians are going to be looking at the electronic healthcare system more and more before the patient arrives, to understand what the active questions and issues might be at hand when that patient does come in. In other words, the future EHR could start to predict and assist, showing all the tools and decisions that are going to probably be needed to be made during that encounter and all the support needed to help.
I think it’s very hard to use a decision support tool when the doctor is literally using the computer in front of the patient. As an extreme example there, can you imagine how a patient would feel if a doctor was literally typing in Google during the visit?
It’s going to be a challenge. The ideal situation is for physicians to use decision support tools before the encounter. But how do you even know what to do before the patient shows up? The tools are going to have to get much more intelligent and adaptable to the complex workflow around seeing a patient.
The future of decision-making alongside technology, AI and data.
HT: Do you think there’s some technological leap that could come and enable us to achieve these goals in an easier way? For instance, Tesla talks about solving the battery issue as a key step forward, so they’re focused on the battery issue. Is there something like that in this data-driven healthcare that you say, oh, that’s going to make such a huge difference?
Atul Butte: We are going to see many technological leaps in the next few years around data-driven healthcare. Probably the biggest one is going to come from the field of AI. We already have enough data now to be able to innovate new tools using machine learning and AI. I think we’re going to see more and more of those deployed in healthcare.
For example, there have already been major science publications, and now companies, that have shown that a computer looking at pictures of the retina can help diagnose diabetic retinopathy, with computers just as good, perhaps even better than ophthalmologists.2 Now that same kind of AI technology is being developed for radiology, computerized tomography (CAT) scans, cardiac imaging, and much more. I last counted more than 100 approved algorithms just in radiology. It’s still astounding to me that the United States Food and Drug Administration actually has a path forward towards approving AI-based tools, right now.
All of this has really just occurred in the past three years. This is brand new. We’re just at the beginning of AI-driven healthcare. It’s going to help doctors, but there’s also going to be a little bit of fear, meaning if the computational tools can do this, will physicians be compared to them? I mean as much as we can kind of laugh at that, to say with confidence that physicians are never going to go away, there is still that perception of a threat that AI could impact physicians. But this is the wrong way to think about it. Another way to consider it is whether future physicians using AI might be compared to those without using it.
AI is also going to help patients as well. I see a world where patients have more and more access to their own healthcare data. It could be in smartphones; it could be on computers. We are giving more and more of that data as health systems to patients and now it’s up to them and their computational tools to figure out what to do with that data. Right now, patients really don’t have medical decision support tools. I see a future of AI-driven, patient-facing decision support tools. We have enough data now to teach and train AI. I think we’re going to see an incredible future with AI.
Entrepreneurialism and the future of healthcare
HT: How are healthcare leaders and start-ups working together to transform the healthcare industry?
Atul Butte: As we are chatting, I am sitting right now in Silicon Valley. We clearly do a lot of innovating here in many fields, from personal computers to genetic engineering. I’m surrounded by entrepreneurs and idealists and idealism. When you look at healthcare itself, there seems to be a lot of room for improvement. I don’t think that’s disputable. One then naturally wonders, how much can one do to help innovate? We see the empowerment here in Silicon Valley that small teams can do incredibly big, useful, helpful things for the world. Now, to be clear, there are many external companies right now doing their best to try to fix the healthcare system, or to disrupt the healthcare system – disrupt meant as a positive term here. Disrupting for something better. We have companies like Amazon and Apple and Google and Walmart, major corporations, trying to assess the healthcare system and come up with a better state.
Now I wonder, can we within the system fix the system? That’s why I like to innovate. As a faculty member, as an academic, I love to create and invent and try to inspire my teams to write papers, sharing with the world what we do and then, in some cases, file the patents that need to be filed and start the companies to actually take it that much closer to patients and families. Now I know not every doctor is going to be wanting to do this, but I think for those that can think about entrepreneurism and innovation, this is a great time to be starting to think about how to fix the healthcare system from within. If we don’t do this from within, we know those outside the system are doing their best to fix it for us. I am hopeful that we within the system can be the leaders here, as healthcare systems, to show the world the right way to reach a better healthcare state for all of our patients and families.
It’s not every doctor, but I believe there’ll be teams, there’ll be healthcare systems that are investing in innovation teams, or research and development teams. We should be able to engineer our own solutions, not just buy them as healthcare delivery systems. That’s the part that excites me in my job.
Start-ups and healthcare system working together
HT: What are some of the best examples of healthcare systems and start-ups working together?
Atul Butte: When I look at the academic health systems that I’m in now, I’m at the University of California, San Francisco (UCSF) and the University of California. Even at UCSF we have incubated hundreds of companies creating thousands of jobs, right now in healthcare. You have to also consider an older start-up company named Genentech that came out of UCSF.
Academic medical centers have, from the very beginning, been able to create these inventions and get them out of the walls of academia into companies to help patients and families. I think it’s been a great tradition. It shouldn’t be taboo to talk about companies within academia. If we want to change the world, we can’t just keep writing papers about it as academics, we have to think of that next step. That’s why I try to practice what I preach, try to get my students and postdoctoral fellows to think about whether there’s a company in what they’re doing, beyond our academic productivity.
Atul Butte, MD, PhD is the Priscilla Chan and Mark Zuckerberg Distinguished Professor and inaugural Director of the Bakar Computational Health Sciences Institute (bchsi.ucsf.edu) at the University of California, San Francisco (UCSF). Dr. Butte is also the Chief Data Scientist for the entire University of California Health System and has authored over 200 publications, with research repeatedly featured in the New York Times, Wall Street Journal, and Wired Magazine. Dr. Butte was elected into the National Academy of Medicine in 2015, and in 2013, he was recognized by the Obama Administration as a White House Champion of Change in Open Science for promoting science through publicly available data. Dr. Butte is also a founder of three investor-backed data-driven companies: Personalis (IPO, 2019), providing medical genome sequencing services, Carmenta (acquired by Progenity, 2015), discovering diagnostics for pregnancy complications, and NuMedii, finding new uses for drugs through open molecular data. Dr. Butte trained in Computer Science at Brown University, worked as a software engineer at Apple and Microsoft, received his MD at Brown University, trained in Pediatrics and Pediatric Endocrinology at Children's Hospital Boston, then received his PhD from Harvard Medical School and MIT.
- US Centers for Medicare & Medicaid Services. (2020). Article available from https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/NationalHealthAccountsHistorical [Accessed June 2021]
- Padhy et al. (2019). Indian J Ophthalmol 67, 1004–1009