There is a version of accountability that arrives once a year, survives, and is forgotten until the next one. Then there is the version that never stops. American health insurers have just been moved from the first category to the second, and the shift is quietly one of the more interesting governance experiments running anywhere in the world.
The programme under review
Medicare Advantage covers more than thirty million older Americans through private insurers. The government pays each insurer monthly, and the amount rises with how ill the member’s medical records show them to be. Fair in principle: sicker people cost more to look after. Vulnerable in practice, because the payment follows the paperwork rather than the patient.
For years, the paperwork drifted. Insurers built substantial operations to re-read old medical charts looking for conditions that could be added to a member’s file, each addition worth money. Congressional advisers now estimate the resulting overpayments in the tens of billions of dollars annually. This spring, three federal audits found that 81 to 91 percent of certain sampled high-risk diagnosis codes lacked proper support in the underlying records, and a major Medicare Advantage insurer agreed to pay 117.7 million dollars to settle federal claims about how its diagnosis records had been assembled.
From event to rhythm
The government’s response was not a single dramatic intervention. It was a calendar.
Audits of payment year 2020 began in the early months of 2026. Payment year 2021 followed in May. Further payment years are already scheduled through 2027, roughly one every three months, each covering a different year of past payments. The audit workforce grew from around forty reviewers to approximately two thousand certified medical coders. Insurers now know, years in advance, roughly when their turn arrives, and the published RADV audit timeline for payment year 2021 and the years that follow has become required reading in compliance departments across the industry.
That predictability sounds like a gift to the audited. In practice it is the opposite. A single annual audit can be survived by a well-drilled team working weekends. A rolling programme cannot. When the next review is always within a quarter, the only viable strategy is to be permanently ready, which means the internal habits have to change rather than the crisis response.
Why continuous beats occasional
Environmental regulators learned this lesson decades ago, which is why continuous emissions monitoring replaced periodic inspections in most serious regimes. Occasional inspection measures how well an organisation prepares for inspection. Continuous review measures how the organisation actually behaves.
The healthcare version is producing the same effect. Insurers describe rebuilding their internal review processes around a quarterly rhythm: sampling their own records the way federal auditors would, testing whether every documented condition traces back to a real clinical visit with supporting evidence, and correcting what fails before anyone officially asks. The ones doing this well report that the audit itself has become uneventful, because the work happened months earlier.
There is one discipline that separates the serious programmes from the theatrical ones, and it is worth naming because it applies to accountability systems everywhere. A credible internal review corrects in both directions. It adds conditions that were genuinely present but never properly recorded, and it removes conditions that were recorded but cannot be supported by evidence, even though removal costs the organisation money. Federal investigators have been explicit that programmes which only ever find errors in their own favour are the pattern drawing the harshest scrutiny.
The transferable question
Britain has no Medicare Advantage, but it has plenty of systems where public money follows recorded data: NHS allocation formulas, outcome-linked payments, regulatory self-reporting across energy, water, and finance. The American experiment offers a clean test of a question every such system eventually faces. Is your accountability an event or a rhythm?
Event-based accountability produces organisations that are excellent at being audited. Rhythm-based accountability produces organisations that are excellent, or at least honest, all the time, because there is no gap long enough to drift in. The difference is not a matter of stricter penalties. It is a matter of frequency.
The American insurers currently living through their first rolling audit programme did not choose this. It arrived after fifteen years of drift and a nine-figure settlement made the case unanswerable. But the design lesson is now visible for anyone building an oversight system from scratch: the schedule matters more than the severity. An audit that might come next quarter changes behaviour in a way an audit that comes once a decade never will, however fierce it is when it finally arrives.
