When the continuous-enrollment provision tied to the COVID-19 public health emergency ended in spring 2023, states began the largest redetermination effort in Medicaid's history. Over the following 18 months, tens of millions of enrollees had their eligibility reviewed. The headline number that should haunt every agency is this: across most states, the majority of disenrollments were procedural, not substantive. People lost coverage because of paperwork, not because they were found ineligible.
The difference between procedural and substantive denials
A substantive denial means the state evaluated someone's income, household, or category and determined they no longer qualify. A procedural denial means the state never got the information it needed to decide, so it closed the case by default. The person may well have still been eligible. Federal data from the unwinding showed procedural terminations running around 70 percent of all disenrollments in many states, with wide variation driven almost entirely by operational design rather than population differences.
That distinction matters because procedural denials are largely preventable. They trace back to undeliverable mail, forms that never reached the right person, confusing renewal packets, language barriers, and short response windows. None of those are eligibility problems. They are communication and process problems.
Why this lesson is urgent again
The unwinding was a one-time event, but the dynamics it exposed are about to repeat under different rules. H.R.1's Medicaid community-engagement (work) requirements take effect January 1, 2027, and they add a new layer of recurring documentation that enrollees must satisfy or claim an exemption from. Every mechanism that drove procedural churn during the unwinding, the bad address, the unread notice, the missed deadline, will now operate against a population that must prove activity or exemption on an ongoing basis.
Arkansas already ran this experiment. When the state implemented work requirements in 2018, roughly 18,000 people lost coverage in a matter of months, about one in four of those subject to the rules, before courts halted the program. Post-mortems found that most of those losses were not people refusing to work; they were people who never successfully reported their hours or did not know the requirement applied to them. It was a procedural-denial event wearing an eligibility costume.
The practical takeaway for agencies and plans is to treat the period before 2027 as preparation, not paperwork. The states that fared best in the unwinding invested early in address hygiene, ex-parte (automatic) renewals, multilingual outreach, and live human follow-up before a case closed. Those same investments are the highest-leverage moves available now. The unwinding was a costly tutorial. The agencies that studied it will keep more people covered when the next wave arrives.