Value-based care (VBC) was promised to improve how American healthcare is delivered. Listen to earnings calls today and health plans report that 60–80% of members are in value-based arrangements, yet the total cost of care keeps rising. I’ve spent 30 years in healthcare data and analytics. I started in the staff-model HMO days at Coventry, led the first actuary team supporting the hospital system Highmark was building, and for the past two years I’ve been Chief Actuary at Arbital Health, where we build AI solutions for payers and providers. I’ve seen VBC from both sides.
I recently moderated a Becker’s Healthcare webinar, The VBC Reckoning: A Payer’s Honest Look at Wins, Failures, and What Comes Next, with Niall Brennan (Chief Data & Analytics Officer, Horizon BCBSNJ) and Ali Khan, MD (Executive Advisor at Valtruis, former Chief Medical Officer for Aetna’s Medicare business). We covered where VBC has and hasn’t delivered, why programs stall, the data problems underneath, and what comes next with AI.
When I asked whether VBC has delivered for payers, both panelists landed in the same place: it depends.
Niall put it plainly: “at best we could say that it’s been mixed.” He pointed to the Medicare Shared Savings Program (MSSP) as evidence that VBC can bend the cost curve. He cautioned that success there doesn’t automatically carry over, because Medicare is a price setter while private plans are frequently price takers. Quality and consumer experience have improved in places, but “as long as VBC operates on a fee for service chassis, there are always going to be challenges.” Even when certain services are reduced, if prices on the remaining services rise at an unsustainable rate, managing trends remains a struggle. Those unit-cost increases are driven significantly by provider consolidation, including hospital mergers and hospital acquisitions of outpatient and ambulatory practices. As Niall stated, these are pressures that “largely exist outside of the structures of value-based care agreements.”
Ali offered the other side. Where arrangements are full risk and capitated, and organizations have every incentive to change operations, he said “value-based care has shown outsized results.” The problem is that much of what gets labeled VBC is upside-only contracting that drives no real operational change.
Ali described the market bluntly: “there’s 31 flavors, like Baskin-Robbins, of value-based care arrangements, and the vast majority of folks are living in sidecars.” These are upside-only arrangements or very weak ACO structures that large health systems often simply write off. The missed opportunity, he argued, was a gradual transition into meaningful downside risk, which is what actually prompts a delivery system to change how it operates.
Ali traced part of the drift to well-intended policy. The shift from per diem to diagnosis-related group (DRG) payments removed much of the payer–hospital alignment around length of stay. After the ACA, organizations found that the first lever for making this a profitable business was revenue optimization through risk adjustment and quality gap closure. The result was “the rise of this quality industrial complex” focused on Healthcare Effectiveness Data and Information Set (HEDIS) measures and care gap closure, without the same force behind total cost of care. As he put it, “we can point to great quality measures… We can’t point to misery reduction in American healthcare or better clinical outcomes necessarily consistently.”
Niall noted that claims processing in the United States runs on “thirty, forty, fifty-year-old technologies.” Early on, providers were asking, “What do you mean I can’t find out about my final performance until a year after the performance period is over?” My own view as an actuary is that we can’t expect providers taking risk to wait until July for a first look at the first quarter. Providers need to have the early financial view of results that payers have if they are truly being asked to accept any substantial downside risk. Operational data such as admits, discharges and transfers should ultimately be available, and a financial reconciliation of VBC results can reasonably be a monthly exercise.
Much of the data needed to run VBC, including incentive payments and quality components, sits outside claims systems. Niall said this information is exchanged across multiple spreadsheets that are manually interpreted, a situation “ripe for error.” Ali added the infrastructure that sits on top: complex attribution, sub-capitation and other contracting mechanics, financial reconciliation, and prospective monitoring as close to real time as possible. He summed up the state of things: “payer data is usually held together with spit, hay, and bale wire.”
When I asked Niall which single link in the chain he’d fix, he said: “we’ve got to get away from spreadsheets for the stuff that doesn’t flow through the claim system.”
Niall framed AI as a glass half full and half empty. It can improve operational efficiency, stitch data together in ways that were once too expensive, and improve the customer experience. But “incentives trump just about everything in healthcare.” The early evidence shows providers using AI-enabled tools such as ambient listening to increase coding. Providers describe this as correcting historical undercoding, but the cost flows to payers as higher reimbursement and ultimately to consumers as higher premiums. Ali described the risk of “bot versus bot” arguments over revenue cycle management, coding intensity, prior auth appeals and denials.
For AI to be deflationary it has to be deployed by parties whose incentives align with what it’s trying to do. He sees the most promise in patient activation, navigation, accompaniment, action planning, task completion and proactive outreach. Those are the actions that prevent urgent care visits and hospitalizations and get durable medical equipment to someone in real time. These actions aren’t reimbursable under fee-for-service Current Procedural Terminology (CPT) codes today, and he argued they shouldn’t be. He called this “our next best chance, maybe our last chance” to move toward population-based payment.
The engagement gap shows the stakes. In commercial and Medicare Advantage plans, funded care management engagement is often only 5% to 15% of the population, and value-based contracts commonly assume a 30% engagement rate after the second year. Ali asked whether the same spending could raise engagement to 60% or 70%: “That could be transformative.” But it requires an incentive realignment the industry isn’t seeing today. Niall agreed that real deflation would require “a fairly fundamental reimagining of how healthcare is delivered in the United States.”
An audience member asked: if the infrastructure can’t support it, why keep pretending it will work? Niall cautioned that even perfect infrastructure won’t make a program succeed if provider incentives aren’t aligned. But Ali’s answer was clear: it can be done. Today much of it is done manually, and few plans have end-to-end infrastructure for attribution, contracting, performance evaluation and financial reconciliation, which slows everything down.
Ali pointed to companies like Arbital “spending time on actuarial cleanup and visibility” as part of moving past those barriers. That’s the work we’re focused on: taking the spreadsheet math off your plate, so payers and providers can see where their contracts stand in real time, while there’s still time to act.
Want to hear the full conversation? Download the complete webinar recording here.