Revenue cycle operations
Most billing teams have some version of a process for this. Here is what those processes catch, where they structurally break, and why the failure is almost never about effort.
There are three ways billing teams track Medicaid changes on closed encounters manually: periodic batch re-runs of eligibility, patient callbacks through financial counseling, and spot review of aged accounts before final write-off. Each one works. None of them scales to the full population of uncompensated encounters, because all three depend on a review happening to land after a benefit change occurs and before the filing window closes.
The problem is not that billing teams are not working hard enough. It is that the timing of a Medicaid benefit change is unpredictable, and a periodic process cannot reliably intersect an unpredictable event.
Almost every safety net provider we talk to is doing at least one of these. They are all reasonable. It is worth being precise about what each one is good at before talking about what it misses.
A list of self-pay or uncompensated accounts gets pushed back through the eligibility system on a schedule — monthly for some organizations, quarterly for most, annually for a few. It is cheap, it requires no new vendor, and it does find coverage. The limitation is that it returns current status, not status as of the date of service, and it only sees changes that happened before the run.
A counselor reaches out to a self-pay patient, learns that Medicaid was approved two months ago, and the account gets rebilled. This is the highest-yield method per account touched, and it also serves the patient. It is also the hardest to scale, and it has a selection problem: the patients whose encounters went uncompensated are disproportionately the patients whose phone numbers and addresses have changed.
Before an account is closed to bad debt or sliding fee, someone gives it a last look. In practice, because the queue is longer than the hours available, this becomes sampling rather than review — and the sample is usually chosen by balance size. That selects for the biggest accounts, which is rational, but it has no relationship to which accounts had a benefit change.
Follow one encounter through a quarterly review cycle and the structural problem becomes obvious. Nothing here requires anyone to make a mistake.
Eligibility came back with no billable coverage. The encounter moves to self-pay. Everything about this decision was correct on the information available.
The account goes back through eligibility. Still nothing. It gets marked as reviewed, which in most systems means it is now less likely to be looked at again.
An application filed in March is approved with an effective date that reaches back past January. The encounter is now billable. Nothing in the revenue cycle generates a signal, because nothing in the revenue cycle is looking.
The account may or may not be in this batch; many organizations scope the re-run to accounts from the last 90 days, which this one no longer is. If it is included, the eligibility response shows current coverage — which a staff member still has to connect back to a January date of service to realize it matters.
The change happened in a gap between reviews. No one did anything wrong. The revenue is gone anyway.
Shortening the interval helps, but it does not solve the problem — it only narrows the gap. Any periodic process has gaps by definition, and the change is free to happen inside one.
Timing is the harder problem, but volume is the one that makes the timing problem unfixable with staffing. Here is the arithmetic, using round numbers as an illustration rather than a benchmark — substitute your own.
Say a health center carries 8,000 uncompensated encounters in an average year. To catch a benefit change within 30 days of it occurring, each of those encounters needs to be examined roughly twelve times over the course of the filing window. That is 96,000 examinations a year. At even two minutes each — optimistic, since a meaningful check means comparing current benefit detail against a specific past date of service, not just reading a yes or no — that is about 3,200 hours, or a full-time role and a half doing nothing else.
And the yield on that work is low per touch by design. The overwhelming majority of those 96,000 examinations return nothing, because on any given check most Benefit Profiles have not moved. A human being asked to perform a task with a 2% hit rate, 96,000 times, will not sustain accuracy — and would be poorly deployed even if they could.
This is the real reason manual tracking fails. Not lack of diligence. It is a surveillance task with a low base rate and an unpredictable trigger, which is precisely the category of work humans are worst at and systems are best at.
Even a team with unlimited hours would still hit the third wall. An eligibility response tells you whether coverage exists now. These four movements are what actually make a past encounter billable, and none of them announce themselves in a yes-or-no answer.
Coverage existed all along and your check said yes — but the service category was excluded. When the scope broadens, the encounter moves inside the benefit without the coverage status changing at all.
Managed care assignment shifts, or a secondary payer appears. The patient still shows as covered. The party responsible for your January claim is now someone else.
The determination that closed your encounter was reversed, frequently with an effective date reaching back past the visit. The appeal happened between the patient and the state; your revenue cycle was never a party to it.
Coverage dropped for a procedural reason and was restored — often with no gap, which means the encounter was covered the whole time and your check simply caught the trough.
These are attributes of the Medicaid Benefit Profile, not of coverage status — which is why watching status, however often, will not surface them.
This is not an argument that your team should stop working accounts. Manual review beats any system on the complicated single account — the one with a coordination-of-benefits question, an unusual documentation history, or a payer dispute that requires someone to pick up a phone and exercise judgment. It is also the only method that serves the patient at the same time, which matters at a health center in ways that do not show up on a revenue report.
The mismatch is specific. Humans are good at depth on one account and bad at continuous surveillance across thousands. Asking your billing staff to do the second thing is what costs you the first.
A straightforward conversation about your facility and the encounters your team has already closed. Implemented by phone in under 60 minutes.
Contact us See how monitoring works5,000+ FACILITIES · 49 STATES + DC & USVI · NO EHR INTEGRATION · PAY ONLY AFTER REIMBURSED