Driving Affordability in Healthcare with Data & AI (Recorded Webinar)
On August 13, we partnered with The Hive to host the first of our three-part series examining the role of data and AI in transforming the US healthcare industry. Part I: Driving Affordability in Healthcare with Data & AI (watch our video recording of the webinar) featured guest speakers with deep and varied expertise from across the industry:
- Kaushik Bhaumik (moderator) — CEO of Glide Health, a software startup and recent DV investment transforming the US healthcare industry by making the cost of care more financially transparent for both patients and providers (doctors / clinics / hospitals). Kaushik previously ran Cognizant’s $3B healthcare business.
- Patrick Spoletini (panelist) — VP and Senior Partner with IBM Watson Health Consulting with over 27 years of healthcare industry (payer and provider) and management consulting experience ranging from strategy, operations, and compliance to digital transformation, change management, and large-scale program management.
- Larry Bridge (panelist) — Partner at Bridgehealth Partners with more than 30 years of experience in healthcare, including leadership of several payer organizations, provider-based plans with integrated physicians, medical centers and hospitals.
The panelists spent the hour discussing the key drivers behind the escalating cost of care in the US and the role of data and AI in helping payors, providers, and patients make healthcare more accessible and affordable. Some of our key takeaways include the following:
- Why the US Spends So Much on Healthcare but Achieves So Little
Healthcare spending in the US is among the highest of all developed nations, coming in at an astonishing $3.5 trillion annually, which amounts to approximately $10K per US resident. Despite this astronomical figure, Americans have a lower life expectancy compared to other nations and higher degrees of affliction across a range of maladies. The irony of this truth — that Americans spend much more but aren’t markedly healthier than citizens of peer countries — was the motivation behind our moderator’s first question: Why does the US spend so much but achieve so little in providing sufficient care?
Panelist Larry Bridge outlined two root causes:
- The healthcare system in the US has evolved piecemeal, rather than being engineered
- The key components of healthcare are individual, private market companies whose incentives are often misaligned with the goal of keeping costs low
He went on to describe the US healthcare system as follows:
“Picture a large rock wall, made up of rocks of all different sizes, some big and some small, […] and a lot of mortar in-between trying to hold it together. Now compare that to a finely engineered brick wall.”
The US healthcare system is more like the former than the latter; comprised of private components with individual incentives that are often in conflict. Letting these components operate individually leads to fraud, waste, and abuse, estimated to amount to a dismaying 25% of America’s $3.5T annual spend.
2. The Impact of AI on Care and Cost
What role, then, can AI play in mitigating this waste? Patrick Spoletini broached this question by first describing two key developments that have empowered AI to have an impact on the quality and cost of care. Unlike in the past, the industry now has access to a wealth of data that simply wasn’t available previously. The gradual accumulation of more and more data points, along with the consolidation and convergence of large payors and providers, have produced substantial data sets that make AI enabled outcomes more reliable. In addition, data scientists now have the computational power to process and analyze this data.
The panelists next addressed how AI can reduce costs from both the administrative and care perspective. Patrick and Larry outlined a number of AI applications on the payor side, including the use of Machine Learning and AI to optimize costly and time-consuming processes such as authorization of care, denial analytics, and revenue integrity. Companies such as Glide Health, a recent DV and Hive investment, offer innovative solutions powered by AI to automate these processes, thereby reducing the cost of previously cumbersome, manual tasks. On the care delivery side, clinicians are using AI to come up with better diagnostics and more targeted therapies. While both panelists have historically seen payors employ AI to a greater degree than providers, Larry and Patrick also agreed that AI use on the clinical side has the potential to drive a lot of future value. According to Patrick,
I’ve seen AI been deployed in payors much longer than providers… It’s not as ubiquitous in the provider space, and there’s a lot more upside in the provider space than in the payor space.
Whether it be in AI-guided diagnosis or AI-guided treatment protocols, the panelists expressed enthusiasm for the future clinical use of AI to improve quality of care and reduce cost.
3. Challenges to Adopting AI from within the Healthcare Industry
Despite the promise of AI to optimize existing processes, its use has been met with some resistance from within the healthcare industry. One reason for this wariness may stem from an existing internal organizational culture that doesn’t rely on data-driven decision making. Patrick pointed out that even some of his most sophisticated clients who say they want to use data to improve performance don’t measure the success of internal initiatives via data, metrics, or KPIs.
“If you don’t adopt the attitude of ‘I want to be, today or in the future, a data driven organization,’ you don’t have the right culture to start with.”
Another point of resistance stems from barriers preventing the exchange of information. Data points in isolation have exponentially less predictive power vs. a more robust data set, and there is dramatic potential upside for an organization that is open to exchanging data both internally and externally. Patrick put it succinctly:
“You’ve got to be able to exchange data within your organization as well as outside your organization. You have to be intra-operable as well as inter-operable.”
To combat provider resistance to AI, Patrick suggested showing physicians facts and figures around improvements in AI-assisted outcomes, while also making this data easily digestible to mitigate physician fatigue.
On the topic of data privacy, both panelists agreed there was a need to balance the value of having more complete patient data with developing protocols and security that protects individuals. Going back to Larry’s earlier point about the evolution of the US healthcare system, Patrick highlighted the difficulty of developing secure systems within a network of separate, often misaligned entities. Positioning privacy as a business imperative, he suggested, rather than a compliance issue may be one way to incentive better data privacy protocols within organizations.
4. The Role of COVID-19 in Accelerating Technology Adoption
The panelists also touched upon the impact of the pandemic on the adoption rate of new technologies — a timely topic with important ramifications for the current and future state of the US. In terms of new technologies, one major trend observed by both Patrick and Larry was the promotion of preventative care and initiatives to push health care more into the community, in an effort to address patient ailments before they get to the ER. Both panelists were optimistic about the potential for telemedicine, wearables, and better data collection to enable more intelligent decision making among clinicians and practitioners interacting with patients well in advance of a hospital visit.
More specific to COVID-19, AI is also being used to expedite clinical trials and hasten the path to a set of treatments and vaccines for the virus. Instead of looking for how AI can create a cure, however, Larry emphasized the importance of stepping back and looking at the larger picture of how automation is improving the healthcare system as a whole. He summarized his point of view by saying:
“That’s the real promise of all this… it’s not that AI will solve any one thing, or that there’s an AI-enabled solution that will solve COVID-19, but if you look at the myriad of things that are going on out there to tackle this issue, AI seems to be making a lot of them bigger, faster, quicker to getting to results.”
The question, then, is perhaps not “how will AI cure COVID-19?”, but “how will AI improve existing processes to get us faster to a more efficient cure?” From that perspective, the future looks bright given the wealth of new AI-empowered innovation.
5. How the Startup Ecosystem Will Shape the Healthcare Industry
To round out the discussion, both panelists gave their opinions on whether the proliferation of startups applying AI solutions to problems in the healthcare space would lead to further fragmentation within the industry. While the number of new startups in healthcare offers great promise, the flurry of activity also threatens to exacerbate the ailments of an already decentralized system. In a thoughtful response, Larry suggested that AI-focused startups may avoid the pitfall of producing an excess of options that fail to integrate by looking to improve, rather than invent, solutions.
“AI is not about bringing something new to the market, as much as it is correcting what’s not working well. Let’s use this to address this problem. Let’s get a better clinical outcome. Let’s get a lower cost outcome. Let’s do this more quickly, more efficiently.”
Because AI tends to focus on optimizing existing solutions rather than creating niche solutions, the growth of startups in the space has the potential to improve healthcare without leading to dissolution. Patrick went on to suggest that M&A activity and future roll-ups would also prevent further fragmentation. Consolidation of targeted solutions will promote rather than hinder healthcare affordability by making larger organizations more effective while arming them with cutting edge technology.
These are just a few of the highlights from a very rich discussion. Overall, the session offered a unique and insightful glimpse into how AI is reshaping the healthcare landscape and addressing the fraud, waste, and abuse that currently plagues it.
If you were unable to attend the webinar, or if this coverage piqued your interest in the role of data and AI in improving affordability, you can find a link to a recording here.
Be on the lookout for Meetup invitations to Parts 2 and 3 of the series:
- Part 2: Preventative Healthcare & Lifestyle in our Sensorized World — scheduled for 09/10/2020
- Part 3: Personalization of Healthcare and Medicines with Data & AI — date TBD