Nobody I talk to in health plan leadership argues that Risk Adjustment (RA) and Quality Improvement (QI) should stay fully separated. Operators, consultants, and vendors have been saying the two should integrate for most of the last decade, and the folks running these functions inside health plans agree with the thesis. What I notice, almost everywhere, is that the agreement has not produced integration. Most health plans still operate RA and QI as parallel functions with coordination layered on top, and the gap between what everyone agrees should happen and what actually happens is the thing worth looking at carefully.
The push toward integration has been well-intentioned. The case has always been that integration reduces member abrasion, reduces provider abrasion, cuts duplicative medical record requests from the RA and QI sides of the same plan, cuts duplicative review of the same clinical data, and lowers the overall operating cost of running two parallel outreach and documentation operations. Which is exactly what makes the gap between the agreed-upon benefits and the actual state of operations worth looking at. If everybody wants it and the case has been made for a decade, why has not it happened at scale?
The usual explanations (silos, misaligned incentives, legacy org charts, bad data infrastructure) describe real friction, but they stop short of the thing underneath. Genuine integration requires one of these two functions to give something up. Not in the abstract. In the specific. Budget line, headcount approval, provider relationship ownership, vendor contract control, workflow priority. The resources around each member are finite in ways that get explicit fast once a plan starts examining them. A 72-year-old member with multiple chronic conditions will answer only so many outreach calls before stopping. The PCP has twelve minutes with a member, and closing a care gap and documenting a comorbidity both have to fit inside those minutes. The care manager works one member at a time, and the attention spent on a chart review follow-up is attention not spent on a medication adherence check-in during the same call. Every one of those resources is a shared asset RA and QI both need access to, and most health plan leadership teams have not decided, out loud and on purpose, how to allocate them when the two functions want the same thing at the same time. The current regulatory environment is what turns that unresolved allocation into a financial problem. Plans that keep letting the trades get made by default, by whoever has the open headcount this cycle or the CFO's ear this quarter, are going to feel it in Stars results, RADV outcomes, and Medical Loss Ratio (MLR) performance by CY 2028.
This is Article 3 of a three-part series called Risk Adjustment in the New Regulatory Environment. Article 1 covered the plan-side buying criteria that stopped fitting the current environment, and Article 2 traced how the risk adjustment vendor market is splitting in response. This article picks up the organizational layer both articles pointed toward.
Why the RA/QI Separation Produced Results for a Decade
The separated architecture produced results for a decade because the two functions were measured on outputs that moved independently of each other. RA was measured on Risk Adjustment Factor (RAF) capture. QI was measured on Medicare Advantage Star Ratings across HEDIS clinical measures, CAHPS member experience measures, Health Outcomes Survey measures, and operational metrics. The data overlapped because both functions touched member records, ran outreach, and engaged providers, but neither function's success depended on the other's workflow.
Each function hired against its own measurement framework. RA hired coders, chart retrieval staff, prospective program managers, and analytics talent oriented to HCC capture. QI hired clinical quality staff, HEDIS abstractors, CAHPS program managers, and performance improvement talent oriented to Star Ratings movement. Vendor stacks, reporting lines, and budgets reflected the same split. The architecture made internal sense because the measurement frameworks themselves treated the two functions as if they produced independent outcomes.
I will keep this short because most readers lived it. The point worth sitting with is that the separation was not an accident of org chart inertia. It was a rational response to a measurement environment that treated the two functions as independent. That framework has been shifting since 2024, and the last five quarters of CMS action have reshaped it to the point where the architecture built against the old version produces operational cost and compliance exposure the old version did not.
What the Current Regulatory Environment Did to the Separation
The plan-side buying criteria that worked for a decade no longer produce the vendor selections health plans need in the current environment, and the risk adjustment vendor market is splitting around specialty depth because that is what the shift is now paying for. Those two shifts also reach inside the health plan and change the RA/QI relationship. The regulatory environment now measures the integrity of the operating process alongside the output metrics each function reports, which means three forces that used to act on RA and QI separately now act on them together. Each force is also where resource trades between the two functions already happen every day, by default rather than by design.
Member attention is the scarce resource, not member touchpoints. Members already get inundated by the health plan, the providers in their network, and the vendor partners those entities contract with. Postcards, phone calls, portal notifications, in-home visits, follow-up calls, reminder texts. Little of it is coordinated across RA and QI priorities or ranked against each other. The predictable response is that the member ignores most of it, which means the health plan loses access to exactly the information it most needs the member to provide, and both functions lose at the same time. The OIG February 2026 compliance guidance flagged in-home Health Risk Assessments as a specific risk area, which raised the cost of HRA programs that previously ran on their own cadence. Star Ratings pressure continues to push QI toward more touchpoints to move CAHPS and Health Outcomes Survey measures. Historically each function budgeted its own touchpoint strategy and the two rarely collided. The collision is routine now.
Provider engagement is the shared leverage point for both functions, and the separation hurts here first. QI wants providers to close care gaps and document screenings against HEDIS specifications. RA wants providers to document specificity and comorbidity against V28 requirements. Provider-sponsored plans close the loop on both agendas at the point of care because they control the clinical workflow directly, and health plans migrating RAF capture upstream through contracted provider relationships, which I wrote about when CMS ended the risk adjustment arbitrage era, are reaching for the same operating property through a different structural path. When the provider receives mixed signals, mixed scorecards, and mixed vendor interfaces from one plan, engagement erodes across both outcomes because incoherent asks produce incoherent response. The plan loses leverage on both measurement frameworks at the same time.
Documentation integrity is the audit exposure both measurement frameworks trace back to. RADV expansion to every eligible contract annually puts RA documentation under continuous scrutiny. Star Ratings methodology changes put HEDIS and other quality documentation under scrutiny that increasingly looks at how the plan documents, codes, and verifies clinical content against the standards it reports against. The direction NCQA is pushing HEDIS, toward Digital Quality Measures (dQM) sourced from structured EHR data rather than retrospective chart abstraction, compresses the QI-side workflow onto the same clinical data stream RA is already pulling from, which makes the parallel-operation answer harder to defend as the measurement moves. When the two functions maintain separate workflows with separate quality assurance protocols, defensibility gaps open in both audit directions because the plan has no single owner accountable for the documentation standard both audits measure against.
CMS has been consistent in signaling that the operating property of the plan matters alongside the output metrics. The Star Ratings methodology changes, the exclusion of risk-adjusted payment on diagnoses from unlinked chart reviews in the CY 2026 Rate Announcement, the RADV expansion, and the OIG compliance guidance converge on the same question for health plan leadership: are you running one coherent operation across revenue capture and quality measurement, or two parallel operations with the same roster of members and providers between them? Health plans that evolved into the two-function answer face a measurement environment that expects the single-operation answer.
Where the Separation Is Actively Hurting Health Plans Now
The compound problem shows up as operational failures that are common across health plans running the separated architecture. Each failure traces back to the same root: resource trades between the two functions happen in the normal course of operations without surfacing to leadership as the strategic decisions they actually are.
The first failure is in the member outreach. Separate RA and QI operations produce calls that each draw on the same finite member attention, often ask for information the member has already provided to someone else at the plan, and together weaken the member's confidence that the plan has a coherent view of them. The member gets a HEDIS gap closure call from a QI vendor on Tuesday, an HRA scheduling call from a care manager on Thursday, and a call from the plan's member engagement team the following week about a condition the RA side has flagged for provider follow-up. The calls rarely reference each other. The member hears the plan asking for different things through different mouths, with no indication that anyone inside the plan knows what has already been asked. Response rates drop, and more importantly, the care the member actually receives suffers because the gaps that would have closed through coordinated outreach stay open and the member's health declines in the places the plan was supposed to be helping them manage.
Competing provider scorecards are a second compounding failure. The average primary care provider in MA networks contracts with eight to twelve payers today, which means the five artifacts the one plan is sending (RAF scorecard, HEDIS scorecard, suspecting list, care gap list, VBC enablement scorecard) sit on top of the same stack of scorecards and lists arriving from the provider's other payer contracts, some as spreadsheets, some as portals, none synchronized with each other. The provider cannot operationally optimize for all of it and engages substantively with the few scorecards attached to the biggest contracts. The health plan loses leverage on both outcomes while still paying for every piece of measurement infrastructure it sends.
A third failure sits inside the chart itself. The medical coding and quality measure review workflows sit with separate teams at most plans. The two sets of standards exist for different regulatory purposes and the review work is specialized. What is missing in most plans is a feedback loop between the two review teams. Both teams are routinely surfacing things on the same chart that would be useful to the other side, often without realizing it. The RA review flags a documentation issue that would also affect a HEDIS measure QI is tracking on that chart. The QI review flags a gap that reflects a chronic condition RA would want captured. Even when plans have invested in centralized data repositories, the systems rarely flag or notify the respective team when one review finds something relevant to the other. The two reviews run against the same source document, surface overlapping but different gaps, and never get synthesized into one view of what the plan is actually looking at on that provider or that panel.
In-home assessment programs are the most expensive version of this. OIG named the shared operational interface as a specific risk area, and the agency's own 2024 analysis found that in-home HRAs and HRA-linked chart reviews drove nearly two-thirds of the $7.5 billion in risk-adjusted payments from diagnoses that had no corroborating care encounters, which is the pattern the 2026 guidance is responding to. When RA and QI priorities run through the same in-home visit without being reconciled at the plan level first, the clinician in the home is executing on two sets of priorities the plan may not have coordinated. The resulting documentation tries to satisfy both sets of standards simultaneously, and may create audit exposure when auditors review the diagnoses captured against the clinical content of the encounter. The vendor executing the visit is often not the problem; the gap is usually that the plan doesn't reconcile RA and QI priorities before the clinician enters the home.
None of these look like a vendor, staffing, or technology problem in isolation. Each one only becomes visible at the integrated operating level, and plans still reading them at the function level are responding at the wrong altitude. That is where this gets honest about what integration actually costs.
What the Integrated Operating Model Looks Like in Practice
The sketch below draws on fifteen years working with health plans across RA, HEDIS, Stars, and Quality, and on recent conversations with RA and QI leaders inside plans actively working on this problem. These are the decisions a health plan C-suite has to make if integration is going to happen deliberately, and each decision has a visible winner and a visible loser inside the operation.
The first decision belongs to leadership and is upstream of everything else. Health plans are businesses with finite resources operating across populations that often carry multiple chronic conditions, which means staying profitable is itself a prerequisite for continuing to serve those populations. The strategic levers available to leadership (MLR management, the timing and composition of RAF revenue, Stars QBP revenue, geographic footprint, product mix) run on different timelines and each one implies a different operational emphasis for RA and QI underneath. A plan leaning on QBP revenue for the next two years needs the Stars-adjacent QI work funded and staffed heavily, and the RA work has to be structured to support member engagement that also moves Stars measures. A plan leaning on RA revenue to stabilize MLR through the V28 phase-in needs the opposite emphasis for a different time window. The failure mode is leadership picking a strategic direction, naming it on an earnings call, and then expecting the operations to adjust themselves without being told what the adjustment looks like at the function level. The RA leader and the QI leader cannot translate a strategic direction into quarterly priorities without leadership telling them which RA and QI priorities matter most and which ones have to be deprioritized to make the direction operational. The leadership measurement framework that evaluates both function leaders on the integrated outcome alongside their individual function metrics is how this decision shows up operationally, because it signals to the function leaders what they are actually being judged on and over what cycle.
The second decision names who owns the member outreach prioritization logic. This is not a human sitting between the two function leaders reviewing every outreach, because that approach does not scale. The functional answer is a shared data layer with rules logic that integrates member-level RA priorities and QI priorities, flags conflicts, ranks the next-best outreach by expected value and member tolerance, and routes exceptions to human review when the rules cannot resolve them. Making that layer real requires one function to own the prioritization logic and the data integration that feeds it. The function that does not own it loses the ability to drive outreach to its own priorities on its own cadence. What the plan gets in return is a member who stops hearing the plan ask for three different things on three different days, which is what the CAHPS and Health Outcomes Survey measures are rewarding and what rising member abrasion is punishing.
The third decision aligns the provider engagement model and the scorecard. Ideally, one scorecard reaches each provider, updated frequently enough that the provider trusts the data is current, with RA and QI priorities reflected together and incentive structures aligned so the provider is not asked to choose between the two outcomes. This is the decision that unlocks the point-of-care integration provider-sponsored plans are already demonstrating structurally. For contracted provider networks, it requires rewriting the engagement contract, which means both the RA team and the QI team lose some of the specific levers they currently pull on providers independently. What the plan gains is a provider actually optimizing for both outcomes at the encounter, which is the only sustainable path to the RAF and the quality numbers the current environment is going to measure them against.
The fourth decision builds the review feedback loop between the coding and abstraction teams. The teams stay separate because the specialization is substantive and the regulatory purposes behind each review are different. What changes is that findings from either review get routed to the other team when they are relevant, through either systems logic or a defined human handoff. One function owns the handoff process and the system that supports it. The function that doesn't own it loses the ability to run its review in isolation from the other team's cadence. What the plan gets is one coherent view of what it is actually seeing on each provider and each panel, which is the thing most plans today cannot produce.
The fifth decision rationalizes the vendor stack toward platforms that span RA and QI data, workflow, and measurement. The plans making this work are not necessarily consolidating onto one platform, but they are retiring the duplicative point solutions that produce conflicting signals into the same provider or member. The cost is operational and political. One function's incumbent vendor relationship typically runs deeper than the other's, and consolidation means that function loses the vendor they built the relationship with. Health plan leadership that cannot absorb that political cost ends up paying for integration capability the plan cannot use, which shows up as margin drag across contract cycles.
What this requires is for the C-suite to make five decisions, each with a visible winner and a visible loser inside the operation, and then hold the consistency as the quarterly pressures that historically force reversion show up. That is the uncomfortable part. It is also the part that most health plan leadership teams have not been willing to do out loud, and it is why ten more years of agreeing that RA and QI should integrate have not produced integration.
What Will Be Visible in the Market by CY 2028
The gap between plans that have integrated their RA and QI operations and plans that have not will become visible in the market by CY 2028. Star Ratings will show it. RADV outcomes will show it. MLR performance will show it as cost per captured outcome diverges between the two groups. Member retention will show it in markets where member choice is actively in play. Sell-side analysts covering public MA plans are already asking about audit defensibility and documentation integrity on earnings calls in ways they were not two cycles ago, and the plans with integrated RA and QI operations will be the ones with cleaner answers.
Most health plan leadership teams have not connected the strategic direction they announced to the operational trade-offs it requires inside RA and QI. The trades get made at the function level in the normal course of operations, carried by whichever function has the most immediate pressure in that cycle, and the resulting incoherence shows up as workflow friction rather than as a strategic decision that was never translated. The health plans that make the translation work have leadership that carries the direction through to the function level, names which function is prioritized for the current strategic cycle and which is asked to absorb the trade, and holds the operation to those decisions as quarterly pressure shifts underneath them.
The vendor market is responding to this shift because the plans are responding to it, and the plans are responding because the regulatory environment is forcing it. That causal chain is why I wrote this three-part series. Platforms that span RA and QI data, workflow, and measurement are winning scope at CY 2027 renewals because that is what the plans adjusting to the new environment are asking for. Specialty point solutions that only serve one function are losing scope at the same renewals for the same reason. Health plans whose internal setup lets them use the integrated platforms will capture the full value of what they are already paying for. Health plans whose setup keeps the two functions apart will keep paying for integration capability they cannot use, which shows up as margin drag across contract cycles.
What the Regulatory Shift Did to Vendors, Plans, and the Operation Inside the Plan
The regulatory environment CMS has built is applying pressure at three levels of the system at the same time: on the plan-side buying criteria for RA vendors, on the vendor market composition itself, and on the organizational structure of RA and QI inside the health plan. The plans and vendors absorbing the pressure well are reading the three as the same underlying force expressed at different vantages, and the ones further along are the plans whose leadership has stopped treating RA/QI integration as an alignment project and started treating it as a resource allocation decision made deliberately, out loud, and with specificity about what each function is being asked to give up.
Ryan Peterson is the Principal of Upward Growth, a health plan market advisory firm that works with health tech vendors, investors, provider organizations, and management consultancies to build strategy around how health plans actually buy, operate, and make decisions. He has 15 years of health tech experience across go-to-market and strategy roles serving the health plan market. He also publishes the weekly Upward Growth newsletter at healthplanweekly.com, covering payor market trends, health tech go-to-market strategy, and CMS regulatory developments for a subscriber base of 7,800+ healthcare executives and investors.