Health plans play a critical role in advancing whole-person care and making healthcare more accessible, convenient and personal. ZS’s Kelly Tsaur, a principal in our health plans practice area, and Himanshu Arora, chief data analytics officer at Blue Cross Blue Shield (BCBS) of Massachusetts, met to discuss the innovations in artificial intelligence (AI), machine learning (ML) and predictive analytics that will help pave the way. Himanshu’s time in this role and his prior tenure with Health Care Service Corp. give Himanshu invaluable perspective on today’s state of healthcare and how health plans can best use their member data and provider relationships. The BCBS Association collectively covers one in three people in the U.S., so how its leaders leverage these capabilities to make healthcare delivery more responsive can be instructive to attuned national and regional carriers.
Kelly Tsaur (KT): What are your key areas of focus over the next 24 to 36 months?
Himanshu Arora (HA): The story of health plans, particularly regional health plans, for the next two to three years is going to be around four vectors. First, the pandemic upended the precarious balance between payers, providers, pharma and, to a certain degree, the medical devices industry. The delivery system continues to be strained by the aftershocks of the pandemic. A lot of this is on the human capital side. Not a day goes by when there isn’t a conversation which ultimately has its roots in the fact that there just aren’t enough physicians, surgeons, nurses and other healthcare staff. The second vector is navigating consumers to the right site of care and the right channel of care. The third vector is around care management. With nursing being as strained as it is, care management is becoming even more important. The fourth vector is equity. Equity has been on our radar, but it’s now formally a fourth pillar.
KT: AI and the proliferation of data promise a true 360-degree view of the member. Could you say more about the customer experience?
HA: The main idea behind focusing on the customer experience is to proactively navigate members with rising risks to the right sites of care and the right channels of care. This is different than what was done in the past because, oftentimes, rising risk would present itself and then we would figure out how to navigate care. We must get more predictive about that for a variety of reasons, some of which I mentioned around capacity constraints in the healthcare system.
The other part of that is the mental and behavioral health component and the sheer lack of capacity in that category of services. Now we are launching an effort to understand more deeply who is at rising risk for mental health and behavioral health services. And with the strong correlation the data has shown us, we are starting to look in a more intensive way at behavioral health as a leading indicator of medical care and other downstream care that’s needed.
KT: We often talk about getting the right care to the right person at the right time. Will this approach bring about the change the healthcare system needs, or is there more we need to consider?
HA: More and more of healthcare can be effectively delivered outside the traditional healthcare setting. I read somewhere that 93% of all COVID-19-related care was delivered either over video or through another virtual point of care—or somebody did self-testing at home and then consulted with a doctor or a nurse and went to a pharmacy.
What’s the vision? My aspiration is that the healthcare system, instead of being this thing that we go to, gets attached to us. It’s a combination of things such as augmented reality, virtual reality, internet of things, analytics and the data we are producing. So much of healthcare is about nudges at the right point in time and care delivery through the right mechanism at the right point in time so you don’t always have to wait to get an appointment. That’s where healthcare is changing dramatically and rapidly. The customer experience need is much greater because there’s a much greater chance now of consumers presenting for care or having to figure out on their own where to present for care. Oftentimes, those decisions are not optimal.
KT: The levers you’re talking about help to alleviate that supply and demand challenge, and I hear you saying that AI, technology and data create scale we didn’t have before. If we do it right, these will unlock real value. If we do it wrong, they create risk. Do Blues plans have an opportunity to solve this in a way the nationals don’t?
HA: Our advantage is that every BCBS plan is more connected at the community level than national plans. In fact, I raised your question with a new colleague who joined us from a large national health plan. His point was, “Look, we could never hope to have as good of a ground game, or connectivity, the deep relationships, the ability to influence or even understand the vectors.” We talk about social drivers of health, but where do those social drivers of health live? They live in the geographies that people are in. Our plans also are in tune with local employers and the employment market. We’re able to take a much more nuanced view of where we expect mobility across different populations, given what sectors and industries are coming versus which ones might be exiting. We have a better understanding of the ground game through a business and employment lens.
KT: If national plans compete on scale, process and a common playbook, the Blues win on their incumbent knowledge. How can you use AI to scale that expertise across multiple initiatives?
HA: For example, we’re in the middle of a pharmacy benefit manager (PBM) transition. No two PBMs are alike, and we’re trying to interpret what we have in place in terms of a business rule with our existing PBM and how should it translate into a business rule or logic for the new PBM. I wish I had a conversational AI tool. One expert on my team who’s in demand across several different areas is now 100% deployed on applying pattern recognition, ML and AI to create a conversational engine that lets a business expert ask or poll the conversational engine, as opposed to having a data scientist in the mix doing that. AI unlocks the local advantages and knowledge we have and enables us to better compete at scale.
KT: Clearly the Blues have the density, relationships and insight that ultimately translates into better judgment on how to apply AI, broader value propositions and use cases to capitalize on. What disadvantages do you see?
HA: With 34 different Blue Cross plans across the country, it’s very hard to get 34 different entities to operate as a cohesive whole. Most of these, from a capital investment intensity perspective, also have limitations. Whereas a UnitedHealthcare or an Aetna can afford to invest millions of dollars in creating advanced capabilities with the understanding that the return on investment will be down the line, for the general regional Blues, it is much more about creating longer-term impact even with smaller investment scales.
If we launch an initiative and it doesn’t deliver results right away, it is natural to expect some degree of focus getting diluted. So now I try and set that expectation further upfront to help allay the concern of “I thought this was six weeks, not six months.” But it’s important to show progressive outcomes quite often, almost every other week, in fact. We’re going to get on the other side of it with not just maturation of the tool but also with the adoption of one, two or three use cases. It will have to be, at least initially, a use-case-based approach. Once those projects start to get across the finish line, people start to see where it’s taking us. It takes time to build the factory, but once you start shipping product out the door, improvements and enhancements are easier.
The other challenge, at least for us, and I suspect for many others, is it’s not simply a matter of dollars, as important a constraint as that is. A lot of the limitations are around subject matter expertise and the growing trend around tenure and retirements in this space. Blue Cross plans generally have very long personnel tenures, and that’s something we take great pride in, but that doesn’t mean that every time long-time employees walk out the door that knowledge has to leave with them. Leveraging analytics, AI and ML to encapsulate that knowledge and keep it in-house, I think that’s one of the lesser explored opportunities. So much of the instrumentation of connecting different systems happens in a human-capital-intensive way, and that’s where the knowledge tends to get built up. We need to be able to connect systems in a way that is supported by a human, but the actual intelligence creation and deployment is done by the “machines in the back.”
KT: How have you responded to these challenges to advance your initiatives and build support for them?
HA: The challenge of untangling the non-functional, non-technical aspects has been my biggest learning. And so this is where the idea of building a coalition within an organization and oftentimes with your peers and collaborators outside of the organization is important. The way to do that is by delivering quick wins. All my internal champions are champions because we delivered something for them—and with them. We delivered it at the right time when they most needed it, and we didn’t let perfect be the enemy of good.
In many cases, we create scale by bringing people from the outside so that our internal subject matter experts have more opportunity to be the experts and not the experts and the doers. That model helps, but, from a cost perspective, it is not inexpensive. At the end of the day, this manifests in per member per month (PMPM) costs. For us to put something to the PMPM, we have to take something out from the PMPM, whether it’s medical costs or administrative costs. That cycle is one where I think the greatest area of opportunity is. What are the partnership models that can be created to co-invest in capabilities and then co-derive value from them over time as opposed to a fee-based, schedule-based or rate-based model?
KT: In evaluating potential partnerships, what advantages do the Blues offer?
HA: One of the things that makes partnerships more appealing with Blues is our member focus, sense of community impact and goal of advancing access, care and quality. A lot of this is driven by market share. Companies that want to create an impact in a certain geography or market can do that much easier by working with one Blue or a small coalition of Blues. We can choose to be focused on what we do best, taking the outcomes of functional capabilities and deploying them in a healthcare delivery setting, and letting the partner develop the technology or write the algorithm. It’s not a novel or a unique idea. Highmark Health is working with Google Cloud, and there’s experimentation happening with Microsoft and others. The challenge is that not all partnerships are created equal. Some companies you might partner with openly say, “Your margin is my revenue.”
KT: Are there other areas where it would be acceptable and, frankly, strategic to get a partner and develop a utility across several Blues? Several Blue utilities already have been built: NASCO for claims processing, Blue Health Intelligence for benchmarking, USAble Life for ancillary benefits and Prime Therapeutics for PBM services.
HA: The idea of building some sort of an analytics AI and ML utility across Blues has been in the air for years. For one reason or another, things don’t come to fruition. It feels like some sort of an external magnetic force, if you will, is needed to piece it together. In my mind, it’s not a matter of whether it will be done. Rather, it’s who does it first. I think it’s a given that something of that sort is going to happen in the next few years.
The opportunity so far has been thought of as a utility for the Blues has to be built by the Blues. And I’m more in the camp of a utility for the Blues must be enabled by the Blues. The underlying building blocks might be better, in my opinion, if they’re done by someone else. Why? Because design choices require neutrality and objectivity. How do you monetize the model when you’re talking about bringing capabilities and the intersection of lives under care? What does that mean in terms of who contributes what, in what way and how much? Those questions have run these projects into a wall in the past. Taking them off the table and encapsulating them with another entity that becomes the enabler has the most likelihood of success, particularly if we’re going to move forward in the next couple years.
KT: Looking ahead, what near-term changes could health plans lead to improve the future of healthcare?
HA: How to share data in a much more liquid way across organizations is going to be, in my opinion, the lifeblood of where healthcare goes and what we succeed in versus what we get innovated out of. We need to start to connect data wherever we can without worrying about the perfect platform. If I can start with 100,000 lives, I’ll start there. If I have to re-platform three, four years down the line, we’ll figure that piece out.
At this critical juncture in the healthcare industry’s shift from traditional fee-for-service to value-based care arrangements, BCBS acts as a valuable role model. Setting aside its more flexible ownership structure and lower sensitivity to earnings pressure, the ability for these companies to work across internal silos and large geographies to mitigate risk and create new business opportunities is instructive. The extent to which regional and national carriers watch from the sidelines or take inspired action of their own sets the stage not only for their success but also for the future of patient-centered care.