Shared Financial Risk Isn’t the Same as Shared Responsibility

by Virtix Health Editorial Team | Sep 2, 2026 | Article

A few years ago, a health plan quality team came off a brutal flu season with what looked like a breakthrough. They pulled ACO performance data on members who had declined or missed the flu shot, cross-referenced it against hospital admissions and ED visits for anything influenza-related, and found substantial overlap. They were excited to bring it to their providers.

They brought it to them in February and asked them to act on it in September.

“That did not work,” Savannah Gonsalves, Director of Quality at Kaiser Permanente Nevada, told a panel of payer and provider leaders at RISE Qualipalooza. “You can have all of the tools at your disposal; the data, the metrics, the outcomes, but that alone is not a strategy.”

Value-based contracts have gotten more sophisticated. Upside-only incentives became shared savings; shared savings is becoming two-sided risk. The infrastructure underneath them mostly hasn’t kept up. Financial risk has been redistributed; decision rights, workflows, and day-to-day accountability have not. That is the core flaw in most value-based arrangements: shared dollars without shared control.

Megan Piotrowski, Senior Director of Solution Architecture at Virtix Health, sits in the middle of that gap, between payers designing value-based programs and the provider organizations expected to succeed inside them. Her read on why they underperform is blunt: “It’s usually not because we didn’t try hard enough or that we didn’t care enough. It’s that we weren’t set up with the right infrastructure from the beginning.”

A Contract Is Not a Plan

Why Contracts Don't Explain How Success Happens

Provider organizations enter downside risk arrangements, often called shared risk contracting, intending to succeed, then discover the contract says almost nothing about how. Performance expectations sit in appendices. Quality targets get negotiated in the boardroom and never translated into anything a clinician can act on. Compensation models stay exactly where they were even though the entire premise of the agreement is that behavior will change.

Savannah Gonsalves watched this exact situation play out. Her plan signed a value-based contract with its provider group and left the providers’ own quality metrics untouched. They were still being paid for quality just not the quality that moved the contract. “It literally took two years for us to be like, guys, we’re not moving the needle here, what’s happening?”

The diagnosis, when it came, was a compensation model that had nothing to do with the agreement. “If you’re working two different quality programs, then both of them are being done kind of half-baked.”

Value-based programs get built from the top down: plans and executives set the goals, and the people expected to deliver them; the clinicians closing gaps, fixing documentation, calling patients often never see the contract, let alone an incentive that reflects it.

Without a unified story about what matters most and why, a contract never becomes an operating model. It stays a legal document.

Four Places Accountability Quietly Disappears in Shared Risk Contract Healthcare

Accountability in shared risk contract healthcare rarely fails at the moment a measure slips. It fails earlier, in four predictable places:

  • Unassigned work. Patient outreach, pre-visit chart review, patient education, and member engagement because nobody owns them on paper, so nobody owns them when performance falls short.
  • Stale data. Claims lag, then review lags. Risk adjustment data can reach a provider a year after the encounter it describes. A quarterly spreadsheet tells you what happened, not what you can still change.
  • Competing asks. Quality pushes one list, risk adjustment pushes another. Both are right. At the point of care, they are noise.
  • Governance theater. Joint operating committees meant to solve problems become one-sided scorecard reviews. “The health plans are just showing your quality data and what you’ve done right or what you’ve done wrong,” one panelist noted. “No one asks questions from the provider side.”

None of those show up in a strategy deck. A provider who looks like they are failing a screening measure may be missing an EHR prompt. They may have lost the medical assistant who closed the loop after visits. “That might be an IT problem. That might be a data problem. That might be a member problem,” Gonsalves said. “Let’s find out why.”

Finding out why is the plan’s job too. Her team runs root cause analysis on the data itself by pulling charts and shadowing physicians, before telling anyone to try harder. “If you’re not performing well, I’m going to really help you understand why.”

Piotrowski’s version tells us more. Working with a group on documentation and quality measures, she found they had had staff turnover and that the new staff did not know they could upload more than one document to the system they were using. “It was little things like that that all of a sudden created an ‘aha’ moment, which made a huge difference when it came to gap closure.” The difference between those two conversations is one question. Effective partnerships replace “Why didn’t you do this?” with “What is getting in the way?”

Building the Foundation Before the Risk

The organizations that thrive under shared risk rarely begin with the contract itself. Instead, they build the capabilities that make success possible before meaningful financial exposure enters the picture. Roles are clearly defined, technology is embedded into clinical workflows, and providers receive practical support through tools like clinical documentation improvement (CDI) and pre-visit review, before they are expected to deliver against ambitious performance targets.

Sequencing Value-Based Contracts from Pay-for-Performance to Two-Sided Risk

“I don’t think you can just turn on a value-based contract, especially a two-sided one,” Gonsalves said. Her plan sequenced it: pay-for-performance first, to get groups fluent in quality reporting and regulatory programs. Then an MLR-based contract — one-sided risk in year one, with a target the group could actually hit. “You’ve got to have attainable goals, because if they’re going into a contract with a finish line that looks too far away, then the effort slows down.” Two-sided risk came last.

Panel moderator Dr. Shannon Decker, CEO of VBC One, put the logic plainly: “You’re not given a driver’s license just on the written exam. You have to go out and practice.”

Turning Data into Action at the Point of Care

Data is a critical part of that evolution, but only when it arrives in a form providers can use. Most organizations already have access to more performance data than they can realistically act on, so another dashboard is unlikely to change behavior. What proves more valuable is identifying the handful of opportunities that matter most for a particular provider at a particular moment, often just two or three priorities, and surfacing them within the electronic medical record while care is being delivered. Peer benchmarking can further strengthen those efforts. Knowing that a comparable practice is consistently achieving a target often creates a stronger incentive to improve than simply seeing the benchmark on a report.

One plan got concrete about it. Every month, each provider received their top three opportunities, and the data behind them surfaced in the EMR. Alerts were toggled per physician, off for the measures they were already strong on, back on if performance slipped. “Nobody loves alerts,” Gonsalves said, “but we got really smart with some of our alerts.”

Decker once presented to a room of more than a hundred physicians with stars marked next to the top performers, blinded so nobody knew whose was whose. They figured it out. “These are grown people chasing me around the room, wanting to know how they too could get a star.”

Extending Incentives Beyond the Physician

The same principle extends to incentives. Too often, they stop with the physician whose name appears on the contract, even though medical assistants, care coordinators, and front office staff are responsible for much of the day-to-day work that determines whether a value-based program succeeds.

Piotrowski has found that organizations achieve more consistent execution when recognition and incentives are shared across the care team, ensuring everyone is working towards the same outcomes and that health plan performance goals are reflected all the way to the point of care.

What Makes Value-Based Programs Succeed, Start with One Question

Building that kind of partnership does not require reinventing the contract. It starts with one question: for every performance measure, who is in a position to influence the outcome? Once those responsibilities are clearly defined, the rest of the model starts to fall into place.

Data becomes a tool for action rather than retrospective reporting, reaching clinicians in time to shape care instead of arriving weeks later as another spreadsheet. That means agreeing upfront on what information will be shared, how often it will be delivered, and how it will fit into existing clinical workflows rather than another standalone portal.

Incentives reinforce the same priorities on both sides of the contract. And governance stops reviewing scorecards and starts removing whatever is in the way, with both sides committed to naming obstacles and changing course when they appear.

At the end of the panel, Decker gave all four leaders a list; leadership education, analytics, clinical documentation, data integration, physician engagement, governance and made them pick the one thing provider groups should build today. All four picked data integration.

“We have too much,” Gonsalves said. “It’s analysis by paralysis at this point. Get it into an actionable spot.”

From Shared Risk to Shared Responsibility

Plans that keep treating shared risk as a financial arrangement will keep getting the uneven results and strained provider relationships value-based care has produced for a decade. Contracts do not change behavior. Operating models do and the people on the hook for the outcome have to be able to see it, in time to do something about it. That is the difference between sharing exposure and sharing responsibility.