The intersection of human behavior and technology is a crucial part of improving healthcare. The big question is how to align the interests of stakeholder -- including doctors, patients, insurers, hospitals, and regulators -- when everyone has different goals and measures of success.
To dig into these complex dynamics, I asked one of the most prominent hospital executives in the country as my guest on episode 285 of the CXOTalk series of conversations with the world's top innovators.
Also: Medical records of high school students leaked
Dr. Rasu Shrestha is the chief innovation officer for University of Pittsburgh Medical Center (UPMC) and head of UPMC Enterprises, the investment arm of UPMC. It is a large payer-provider organization, with 80,000 employees and $20 billion in annual revenue. Dr. Shrestha also leads a Department of Veteran Affairs program, called the Open API Pledge, designed to improve sharing of electronic medical records data.
Given his position, Dr. Shrestha is uniquely qualified to address issues around using technology to improve healthcare delivery and economics. For that reason, the conversation is meaningful and worth your attention.
Watch a video of our full discussion embedded above and read the edited excerpts below. You can also see the complete transcript.
Dr. Rasu Shrestha: I build bridges. I work with a team across UPMC to build bridges between healthcare as we've known it in the past to healthcare as it needs to be. Bridges between the science and the medicine, pushing healthcare forward to entrepreneurship; to creating companies, to making commercial successes of innovations that have to be sticky. Bridges between computer systems, zeros and ones, AI algorithms, and human beings, patients, and clinicians, who are just human beings that need to use these solutions at the back end.
Also: Cancelled My Health Record data to be kept in limbo
Our goals are to move that needle forward, improve patient outcomes, [and] improve patient satisfaction, all of the things that a lot of health systems are very incentivized to do. But, our goals also are to innovate [and] create products, solutions, [and] companies that we would implement these technologies here at UPMC that we call the living lab. We would then commercialize with our partners, we would take out to market, and we would scale these solutions to impact healthcare at a much broader level.
Dr. Rasu Shrestha: When we talk about the struggles of interoperability and where we've been the last decade and a half in this journey of interoperability in healthcare, we've been just focused in on the specifics of interoperability around the electronic medical record systems. That's important. There is much work that needs to be done, but we see remarkable advancements that open up of these capabilities. At the back end, the Argonaut Project leads the charge around HL7 FHIR and creates capabilities for us to have specific data points and access points to data elements in ways that we just hadn't had access in the past.
Also: AI and health: Using machine learning to understand
There's been some progress but making that a reality we need to stop paying just lip service. I'm not looking at any one vendor; I'm looking at the industry as a whole when I address this challenge that we have. We shouldn't just pay lip service and say, "Yep, we've checked the box. We're doing FHIR," or, "We're doing interoperability." It's not about that. It's not about looking at just a handful of specific, say, FHIR-based resources that need to be opened up, but the entirety of what is required to enable a smoother care transaction to happen. To enable a specific set of workflows around Telestroke or transitions of care.
Dr. Rasu Shrestha: I vehemently believe that data science needs to be a new specialty in medical school. It is so important for us to have data science engrained in the practice of how we're pushing healthcare forward. Physicians, nurses, and other care providers need to be bought into this, pharmacists and others included.
How do you do this? That's the big challenge. Today, physicians, unfortunately, are an afterthought when it comes to health IT rollouts. I say this because I've seen this happen in front of me where, with the right intent in mind, with the right intentions in mind, you have an RFP process. You go out, and you say, "All right. Here's the best in class products around, say, a clinical information system, an ERM or a PACS System."
Also: Fitness app PumpUp leaked health data, private messages
You do all of the due diligence to decide perhaps the finalist, and you go and sign the contract. When you're about to go live with that solution, you say, "Hey, let's bring in our physicians. Let's have a go-live event. Let's put in some training. Let's buy them pizza. Let's create cheat sheets." That's not the time to engage physicians. That is way too late in the cycle to engage end users.
The approach that we're taking at UPMC and UPMC Enterprises is engaging end-users first. Before writing a single line of code, we sit down with physicians. We sit down with patients. We sit down with nurses. We whiteboard, we brainstorm, and we talk not about solutions they want but about their pain points.
The notion of design thinking mandates that you start first with empathy. You put yourself in their shoes. You sit down with them in their environment, not in a sterile, fancy room somewhere else in Silicon Valley, but really in the hospitals, in the trenches where they see those pain points experienced day-in and day-out.
Dr. Rasu Shrestha: There's a lot of hype around AI, smart dust, and blockchain. It's less about the technology. It's more about what the technology can do, in our case, to better the pursuit of healthcare.
Is AI important? Absolutely, it's important. AI has been around, however, for a long time. The set of capabilities to make AI sticky and scalable, which is the mantra that I've been harping on now for the duration of this talk, it's more real today than it's ever been before. Is AI an example of an innovation that would be important for us in healthcare? It absolutely is. Less because of the buzz surrounding AI, blockchain, smart dust, augmented reality, or anything else, but more because of how we can actually use that technology to effect change, how we can use technology to affect behavior, and how we can actually move the needle in the way that we're pushing care forward.
Also: No amount of alcohol is healthy, global study says CNET
Let me give you an example. Yes, AI tends to be a buzzword, and there's this notion of, "Hey. Run for the hills. The machines are coming," but AI used well could help in augmenting the care that we're providing to our patients. Around clinical decision support, for example. As a clinician, we're able to maybe synthesize a couple of streams of data, 10, 12, 2, 15 streams of data at any given point in time, let's say. But, computers, algorithms can synthesize and compute thousands, hundreds of thousands, perhaps millions of streams at a given point in time.
We can use the computational power of AI algorithms to look at vast troves of data, connect these silos of databases in a way that they've just not been able to be connected before. To give us insights we can use to make decisions as physicians and human beings. To connect with the patient that's sitting in front of me and to contextualize it to the specifics of what may or may not work precisely for that patient or that consumer... therein lies the perfect intersection of technology and use case.
Is AI the next big innovation? I'd say no, it's not. It's what you can build out of AI, the things that you can build on top of AI, blockchain, and augmented reality. Those things are the innovations that we should focus on.
Dr. Rasu Shrestha: The VA is about to pull the trigger on an electronic health record modernization mandate, that is significant. Significant not just regarding monetary value, upwards of several billion, but also significant regarding the impact that it'll have on the veterans and their families and, quite honestly, the industry as a whole.
What's happening right now is there are academic organizations, healthcare organizations, so UPMC, we're all in on this, and we have the likes of Partners, Mayo, Roche, Intermountain, and many others that have come together as the original signatories around this open API pledge.
My role is to pull together colleagues across the industry, to get them excited around this as the VA contemplates the specifics of rolling out a large, commercial solution, and making sure that we're able to continue to innovate for the sake of our veterans and their families. How do we use this as an inflection point to better the quality of the applications, the workflows, and the solutions that we'll be implementing at the VA. But also use it as an opportunity to get interoperability right once and for all.
Also: 3 health maintenance tips for frequent business travelers TechRepublic
The opportunity at hand is tremendous. I'm seeing a level of leaning in from the original dozen or so signatories and now more than 100 others that have pledged. If you go to the open API pledge website on VA.gov, you'll have an opportunity to get more information on this and join us in this pursuit.
Large organizations are committing to this. You have vendors, small, medium, and large entrepreneurs, systems integrators, organizations like HiMSS are pledging and saying, "We're behind this 100 percent. Let's do this." I'm excited about this and look forward to making it sticky for the industry as a whole.
CXOTalk offers in-depth conversations with the world's top innovators. Be sure to watch our many episodes! Thumbnail image Creative Commons by Mathew Schwartz on Unsplash.