A collections-based financial forecast for behavioral health starts from what a facility’s payers actually pay after denials and underpayments, not from the fee schedule they’ve contracted to pay — and the gap between those two numbers is usually the difference between a forecast that holds and one that doesn’t.
- A fee-schedule forecast models a ceiling, not a forecast: it shows what a payer has agreed to pay in the best case, not what actually clears after denials and underpayments.
- Modern FP&A models payer mix, denials, and collections lag together, rather than treating gross charges as a stand-in for expected revenue.
- Static annual budgets built once a year drift from reality fast when denial rates or payer mix shift mid-year, and a rolling forecast catches that drift months earlier.
- Denial rate and days in AR are financial model inputs, not just operational RCM metrics tracked somewhere else.
- A facility opening a new level of care or location needs a collections-based ramp, not a fee-schedule-based one, because early-stage claim volume ages and denies differently than a mature payer mix.
- Revenue Logic’s own financial modeling process builds forecasts from adjudicated-claims history, so the model already reflects what a payer actually pays before a single new claim gets submitted.
Why the Fee Schedule Overstates What a Facility Will Collect
A fee schedule states what a payer has contractually agreed to pay for a service — it says nothing about how often that same payer denies the claim, underpays it relative to the contracted rate, or delays payment long enough to affect near-term cash position. A financial modeling process built on fee-schedule assumptions inherits all three gaps silently.
Modern healthcare FP&A treats payer mix, denial patterns, and collections lag as inputs to the forecast itself, not as separate operational metrics tracked elsewhere. A forecast that only references the contracted rate is really modeling a best-case ceiling, then presenting that ceiling as an expected outcome.
That last comparison is the real argument for a rolling model over a static one. An annual budget locks in assumptions once and revisits them on a fixed schedule; a rolling forecast updates continuously as new collections and denial data comes in, catching a shift in payer behavior long before the next scheduled budget review would. A realistic spread puts roughly 25 percent of collections in the 0-30 day bucket, 60 percent in 30-60, 5 percent in 60-90, and treats the remaining 10 percent as bad debt that never arrives.
The gap between when a claim is submitted and when the payment actually clears, driven by payer processing time, denial-and-appeal cycles, and follow-up cadence. A financial model that ignores collections lag can show healthy revenue on paper while actual cash availability runs weeks or months behind that number.
What This Looks Like When a Facility Is Growing
A facility adding a level of care or opening a second location needs its model to reflect a collections-based ramp, not a fee-schedule one. Early-stage claim volume at a new location ages and denies differently than an established payer mix — credentialing delays and unfamiliar payer relationships both push that direction. Sizing reimbursement runway before break-even is where that ramp gets tested against real numbers.
This connects directly to a facility’s claims denial management data — per AHA’s Costs of Caring research, underpayment relative to cost of care is a persistent, measurable gap industry-wide, not a one-off billing error. A financial model that assumes a mature facility’s denial rate for a brand-new location will overstate near-term collections exactly when accurate cash-flow visibility matters most.
Per HFMA’s revenue cycle research, the organizations that model collections realistically — rather than assuming contracted rates convert to cash — are the ones that catch cash-flow problems months before they become a real liquidity issue.
Connecting the Forecast to the Rest of the Revenue Cycle
The best inputs for a behavioral health financial model come from the same data used to run the revenue cycle day to day. PayerLenz reimbursement benchmarking shows what a payer actually pays for a given service, which is a far better forecasting input than the fee schedule’s stated ceiling.
Facilities renegotiating a payor contract should update the financial model with the new contracted terms immediately, but should also keep tracking actual collections against those terms — a renegotiated rate is still just a ceiling until real payment data confirms it’s what’s actually clearing. tracking collections against contracted terms is what turns a renegotiated rate into a verified one.
This is the same modeling discipline behind Revenue Logic’s earlier piece on the fee schedule piece — a facility’s forecast is only as trustworthy as the collections data feeding it.
How often should a collections-based forecast be updated?
As a rolling model rather than an annual one — updating continuously as new collections, denial, and payer-mix data comes in catches drift from plan months before a fixed annual cycle would surface it.
Is a fee-schedule-based forecast ever appropriate?
It’s a reasonable starting ceiling for a brand-new service line with no claims history yet, but it should be replaced with a collections-based model as soon as real adjudication data becomes available — typically within the first few months of billing.
Should denial rate really be treated as a financial modeling input?
Yes. Denial rate and days in AR directly determine when contracted revenue actually becomes usable cash, which is the core question a financial model exists to answer.
For the broader pattern behind why fee-schedule models fail, see why financial models get this wrong.
If your financial model is still built on the fee schedule instead of real collections data, contact Revenue Logic and we’ll walk through what the numbers actually show.
- Collections-based forecasting built on real adjudicated-claims data
- Rolling updates as denial rate and payer mix shift, not a once-a-year reset
