Jun 23, 2026 · 35 min · 18 segments
After a two-year hiatus, the PEPPER (Program for Evaluating Payment Patterns Electronic Report) is back. It’s now available to all acute care hospitals and critical access hospitals. During the next…
No entities detected.
Penny JeffersonHost
Cheryl EricksonGuest
Rose DunnGuest
Tiffany FergusonGuest
Frank CohenGuest
Ronald HirschGuest
Chuck BuckHost
Frank Cohen joins us now to wrap up part three in his three-part series, AI Encoding.

So over the last two weeks, what I've talked about is this chess match analogy and how practicing doing everything right at the chart level can still result in an extrapolated overpayment demand.

This week asked the question, so what do you actually do when the demand letter shows up? Now, traditionally, defending an audit has meant defending charts, right? The defense team pulls the records.

But if the audit was triggered by a statistical model that's looking at your data in that aggregate, winning on individual charts does not necessarily get you out from under the extrapolation.

Now, the sampling, the stratification, the extrapolation math, that machinery is doing most of the work.

I've served as a statistical expert in hundreds of these extrapolation matters over my career.

And the cases where providers prevail at scale are almost always those in which the methodology was challenged.

So the contractor's statistical methodology is almost always the softest target in the room.

Sampling frames get built sloppily, strata get defined in ways that violate the contractor's own protocols, and extrapolation math gets applied to populations that look nothing like the sampled claims.

And if your defense team is only arguing clinical merits, they're leaving the strongest ground untouched.

So if the audit was initiated by an algorithmic flag or by using AI, your organization has a right to know what flagged it and why.

Something was pointed at your practice, and there's a due process argument that you're entitled to understand what it examined.

If nobody has asked for the basis of the audit selection, that is a conversation worth having.

Frank Cohen joins us now to wrap up part three in his three-part series, AI Encoding.

So over the last two weeks, what I've talked about is this chess match analogy and how practicing doing everything right at the chart level can still result in an extrapolated overpayment demand.

This week asked the question, so what do you actually do when the demand letter shows up? Now, traditionally, defending an audit has meant defending charts, right? The defense team pulls the records.

But if the audit was triggered by a statistical model that's looking at your data in that aggregate, winning on individual charts does not necessarily get you out from under the extrapolation.

Now, the sampling, the stratification, the extrapolation math, that machinery is doing most of the work.

I've served as a statistical expert in hundreds of these extrapolation matters over my career.

And the cases where providers prevail at scale are almost always those in which the methodology was challenged.

So the contractor's statistical methodology is almost always the softest target in the room.

Sampling frames get built sloppily, strata get defined in ways that violate the contractor's own protocols, and extrapolation math gets applied to populations that look nothing like the sampled claims.

And if your defense team is only arguing clinical merits, they're leaving the strongest ground untouched.

So if the audit was initiated by an algorithmic flag or by using AI, your organization has a right to know what flagged it and why.

Something was pointed at your practice, and there's a due process argument that you're entitled to understand what it examined.

If nobody has asked for the basis of the audit selection, that is a conversation worth having.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.