Contents
Sutter Health agreed to pay $90 million in the Sutter Medicare Advantage settlement announced in 2021, resolving allegations of unsupported diagnosis codes and failures to correct them. The Justice Department had announced a $30 million resolution in 2019. Behind those amounts was a problem a coding auditor would recognize: according to the whistleblower, the audit team found errors, but management restricted corrections to the billing side of the electronic record. The diagnosis in the patient encounter could remain untouched. Finding the error and getting it out of the payment system were different jobs, and the people assigned the first allegedly could not finish the second.
Sutter is a Sacramento-based healthcare provider whose medical foundations treated California patients enrolled in Medicare Advantage plans. The whistleblower was a risk-adjustment manager and certified coder at Palo Alto Medical Foundation. Her account puts this Medicare Advantage fraud case inside the provider’s audit department, where staff could see both the unsupported diagnoses and the limits placed on fixing them.
According to the Justice Department’s earlier settlement announcement, Sutter supplied diagnosis codes to its contracting insurers and received a share of their Medicare payments. The insurers passed those codes to the Centers for Medicare and Medicaid Services, which used them to calculate patients’ risk scores. More severe diagnoses generally meant larger payments to the plans, with a portion flowing to Sutter.
That is how a diagnosis becomes revenue beyond the bill for an office visit. Medicare Advantage pays plans a per-person amount adjusted for health status and other characteristics. Eligible diagnoses map into condition categories that carry payment weights. A diagnosis retained in the data can affect what the plan receives for that patient’s health care, and a provider’s contract can give it a financial stake in the result.
The 2015 paper Upcoding or Selection? Evidence from Medicare on Squishy Risk Adjustment, available through Duke University’s seminar archive, found that coding intensity increased with closer integration between insurers and providers. Its explanation of how the incentive reaches the doctor’s office is direct: insurers can tie physicians’ pay to the codes they assign instead of the visits and procedures they perform.
In Sutter’s case, the government described the financial connection specifically: unsupported codes allegedly inflated payments to both the plans and Sutter. In that account, the organization generating the unsupported diagnosis data also shared in the inflated payments. Taxpayers paid the inflated bill the government alleged.
According to the whistleblower, Palo Alto Medical Foundation’s five-person audit team found 383 errors across three conditions in a partial review of a single year. The team was examining a slice of one foundation’s coding, and its findings gave management a concrete reason to examine what had already been submitted.
The whistleblower alleged that the documentation supported active cancer in only 18 of 182 patients reviewed. The reported accuracy rate was 9.88%, against the 85% accuracy asked of physicians.
The whistleblower’s complaint challenged the relationship between the diagnosis submitted and the patient’s medical documentation. An active-cancer code represents a clinical claim about that patient. For the audit, the important comparison was between that claim and what the chart actually supported at the encounter.
The manager’s role makes the account useful to an insider. Checking codes was her work. She could identify the disputed entries, explain why the team questioned them and follow what management permitted afterward. The alleged failure came into focus when the audit findings met the correction process.
The whistleblower alleged that, on the afternoon before Thanksgiving in 2014, the team received instructions to make changes only on the billing side of EPIC. Changes to the patient encounter were reserved for physicians. That division of authority is the center of the billing-only correction problem in this case.
The whistleblower’s account says the clinical record fed the diagnosis information the contracting insurer later collected for Medicare. Removing a code from the internal billing file could leave it standing in the encounter record used along that path. The staff could perform the authorized correction while the underlying diagnosis remained available for submission.
The complaint also alleged that management had been told that changing only the billing file would not fix the inaccurate diagnosis data reaching CMS. That makes the instruction more consequential than a disagreement about which screen an auditor should use. In the whistleblower’s account, the people directing the work knew that the permitted change would not reach the relevant data.
The physician-only restriction added another obstacle. According to the complaint, auditors could not amend the encounter themselves. If the treating physician declined to make the change or had left, that route to correction closed. The team could identify the entry it believed was wrong but lacked authority to change the record that carried it forward.
For someone working in coding or compliance, that is a recognizable split between responsibility and authority. An audit team can be responsible for identifying unsupported diagnoses while depending on someone else to remove them. We examine that handoff with the insider: the finding, the instruction, who could act on it and which record actually changed. A correction count tells only part of that story.
Independent Health’s work with coding vendor DxID supplies a useful parallel. The government’s complaint quoted meeting notes describing a gap between signed diagnosis forms and the clinical record: doctors were signing DxID’s addendums while the diagnosis list in the medical record went unchanged.
There, the concern involved diagnoses added through separate forms. At Sutter, the whistleblower described corrections confined to billing. Both accounts make the same operational question worth asking: which version of the diagnosis did the organization change, and which version continued toward payment? An insider who understands the systems can explain a difference that a completed form alone conceals.
The whistleblower alleged that Sutter’s compliance department reviewed five of the audit team’s own encounters, found three errors by the auditors and used those findings to declare the wider concerns unfounded. Her account says the audit was stopped.
The direction of that review matters. The original concern was about unsupported diagnoses in patient records. The response focused on mistakes by the people who raised it. In the alleged sequence, that response brought the broader inquiry to an end, leaving the underlying correction problem unresolved.
According to her complaint, the whistleblower had also drafted a revised corrective-action policy assigning responsibility for ensuring that documentation supported the reimbursement received. The complaint therefore described both a written expectation and a response to actual findings. A policy promising accurate documentation sits awkwardly beside an instruction that allegedly prevents auditors from correcting the relevant record.
This is why what compliance did with an audit can matter as much as the audit itself. In this account, the findings reached management, staff explained the correction problem, authority remained restricted and the concern was rejected. The employee who saw that sequence can explain what an outsider looking only at the company’s policy would miss.
The Justice Department’s announcements describe different groups of claims. The earlier announcement left the government’s case against Sutter and Palo Alto Medical Foundation ongoing. The later one resolved claims concerning Palo Alto Medical Foundation and some claims the whistleblower had continued to pursue concerning other foundations.
On April 12, 2019, the Justice Department announced the first settlement with Sutter and four of its medical foundations. The government alleged that unsupported diagnosis codes inflated beneficiaries’ risk scores and caused Medicare Advantage plans to be overpaid.
That announcement expressly distinguished Palo Alto Medical Foundation. A reader who stops at the first settlement misses the part of the matter containing the Palo Alto Medical Foundation audit and correction allegations.
The August 30, 2021 announcement made the correction allegation explicit. The government alleged that, once Sutter became aware of unsupported codes, it failed to take sufficient action to identify and delete additional unsupported diagnoses. The allegation reached the organization’s response after it knew there was a problem, as well as the codes initially submitted.
The government had joined the whistleblower’s case for the Palo Alto Medical Foundation claims. It did not take over the claims involving the remaining foundations. The whistleblower continued pursuing those claims, and the later settlement included some of them.
The later settlement brought a five-year corporate integrity agreement with the Department of Health and Human Services Office of Inspector General. Sutter agreed to implement a centralized risk assessment program within compliance and hire an independent review organization.
That outside reviewer was to examine a sample of Medicare Advantage patients’ medical records and associated diagnosis data every year. The paired review goes directly to the case’s central concern: what the medical record supports and what the diagnosis data say about the same patient.
Seen alongside the whistleblower’s account, those obligations have a concrete purpose. The earlier alleged process allowed a change in one file to stand in for a correction that needed to reach another. Reviewing records and diagnosis data together examines both sides of that divide.
The Sutter case gives a coding employee, clinician or compliance reviewer a specific point of recognition: a documented concern can reach management and still fail to produce a correction in the data that matter. The employee may know why an encounter was flagged, who limited the changes or how a recheck became the reason to close the issue.
Those are connections we work through with a whistleblower. The clinical documentation, the route into the insurer’s submissions and the response to the audit each explain a different part of the alleged overpayment. The strength of the account comes from putting those parts together around actual decisions and records.
A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers. The value an insider brings here is knowledge of what happened between the finding and the supposed fix.
If you saw an audit find unsupported diagnoses and watched the correction stop short of the data sent to Medicare, speak to one of our whistleblower experts about where the correction stopped. We bring in Medicare Advantage fraud whistleblower lawyers who understand provider coding and failed corrections. We work beside you throughout the case, connecting what the audit found with who kept the diagnosis in the payment data.