Skip to main content

Forecast's predictive and expected allowed metrics

Know what you should be paid and when with predictive allowed and recovery metrics

Overview

If your account has access to the Forecast product tier within Revenue Diagnostics, you can see and report on a number of additional predictive metrics.

These predictive metrics go beyond reporting what a payer has already paid. They show what should be paid, how likely you are to collect it, and when to expect it. Built from patterns in your account’s own historical claims, these metrics estimate expected allowed amounts, the odds that a line will be paid or recovered, and the likely adjudication date.

Together, they help you spot trends, prioritize work by likely return, and measure collection performance against a realistic benchmark rather than billed charges alone. Most are available as worklist columns and Reports metrics, so you can use them to guide day-to-day work as well as high-level analysis.

Metrics at a glance

Metric name

Definition

Location in the app

Historical expected allowed

Rivet's calculated expected allowed amount based on trends in the account's historical claims, assuming no payment issues will occur.
​
Calculated as the 25th percentile of Allowed per unit on similar line items × units, capped at billed charges.

Worklist column, Reports metric

Likelihood to pay

Rivet's calculated percent likelihood that a line will be eventually paid in full, even if it was previously denied or underpaid. Line items still waiting on a payer response count as not paid, which can pull the value down for recent dates of service.
​
Calculated as # similar line items now Paid/Overpaid ÷ # similar line items total. When used as a metric, we calculate the average likelihood across line items in the queried data set.

Worklist column, Reports metric

Likelihood to recover

Rivet's calculated percent likelihood that a line will be recovered in full if a payment issue occurs.
​
Calculated as # similar line items ever Denied/Underpaid and now Paid/Recovered/Overpaid ÷ # ever Denied/Underpaid. When used as a metric, we calculate the average likelihood across line items in the queried data set.

Worklist column, Reports metric

Expected adjudication date

Rivet's calculated date of adjudication, where days to adjudicate is the number of days from date of service to the latest check date, counted only for line items with at least one remit.
​
Calculated as Date of service + median days to adjudicate of similar line items.

Worklist column

Hybrid expected allowed

A single expected allowed value that uses the line's Expected allowed (from its fee schedule) first, and falls back to Historical expected allowed when there isn't one.

Worklist column, Reports metric

Hybrid expected with issues

An expected allowed amount adjusted for payment issues that are likely to occur.
​
Calculated as Hybrid expected allowed × Likelihood to pay.

Worklist column, Reports metric

Potential recovery

The dollar amount that could be recovered by pursuing the full expected allowed amounts. It can be negative when the payer allowed more than expected.
​
Calculated as Hybrid expected allowed − Allowed.

Worklist column, Reports metric

Net collection rate

The share of the expected allowed amount that was actually allowed by the payer.
​
Calculated as Allowed ÷ Hybrid expected allowed.

Reports metric

Potential recovery %

The outstanding/recoverable amount as a percentage if the full expected allowed amount were paid on all claims.
​
Calculated as (Hybrid expected allowed − Allowed) ÷ Hybrid expected allowed.

Reports metric

How Rivet defines "similar" line items

For metric calculations that require looking at historical trends, Rivet compares similar line items, not whole claims. A past line item counts as similar when it:

  • Is in the same account,

  • Has had a date of service in the last 180 days (counted back from the latest weekly refresh),

  • Was billed to the primary payer

  • Matches the line being estimated on the fields of one of the groupings below (prioritizing the most specific group)

Rivet tries four tiers in order, from most specific to broadest. It uses the first tier with enough past line items that have an allowed amount:

Line items must match on

Minimum past line items needed

Payer, Rendering provider NPI, Procedure code, Place of service

20

Payer, Procedure code, Place of service

15

Payer, Procedure code

10

Procedure code only

30

Notes on matching

  • Payer refers to the payer name exactly as it appears on the claim file (837), not the mapped payer in Rivet. Two spellings of the same payer count as different payers.

  • Matching ignores upper/lower case.

  • Modifiers, diagnosis codes and billing provider are not used.

  • The same tier and group supply the line's Historical expected allowed, Likelihood to pay, Likelihood to recover, and Expected adjudication date.

How line item calculations roll up to the claim

  • The dollar metrics (Historical expected allowed, Hybrid expected allowed, Hybrid expected with issues, Potential recovery, and the estimated payer payment, patient responsibility and advance payment) are summed across the claim's line items.

  • Lines without a value are skipped, not counted as $0. If 3 of 4 lines have a calculated value, the claim total is the sum of those 3 - meaning it will understate at the claim level. The claim total is blank only when no line has a value.

  • Expected adjudication date for the claim is the latest of its lines' expected dates.

  • Likelihood to pay and Likelihood to recover only exist on line items. There is no claim-level version.

Refresh schedule for calculations

  • Similar-claim groups refresh weekly on Saturdays. Each refresh recalculates the percentiles, likelihoods, and median days from the latest 180 days of data.

  • Each line's calculation is made once, when the claim's first 837 is received. It uses whichever weekly groups existed at that moment and doesn't change afterward, even when the groups refresh or the claim is corrected. That one-time calculation includes Historical expected allowed, Likelihood to pay, Likelihood to recover, and the days used for Expected adjudication date.

  • Combined values (Hybrid expected allowed, Hybrid expected with issues, Potential recovery, Net collection rate, and Potential recovery %) update whenever the claim itself is reprocessed, (e.g. when a remit arrives or its expected allowed changes).

Why a value might be blank

Historical expected allowed, Likelihood to pay, Likelihood to recover, and Expected adjudication date fields may be blank due to the following reasons...

  • The account didn't have access to the Forecast product when the claim was imported into Rivet.

  • No similar-claim grouping had enough similar line items to produce a reliable calculation. This is common for new accounts, rare procedure codes, or payers with little recent history.

  • The line had a billed charge of $0 or less.

  • The line wasn't on the claim's first 837. Lines added on a later corrected or replacement claim aren't estimated.

  • The line is billed to a secondary or tertiary payer. Only primary is estimated.

Other scenarios include the following...

  • Likelihood to recover will be blank when none of the similar line items were ever denied or underpaid

  • Expected adjudication date will be blank when none of the similar line items have been adjudicated yet.

  • Hybrid expected allowed will be blank when both Expected allowed and Historical expected allowed are blank.

  • Hybrid expected with issues will be blank when Hybrid expected allowed or Likelihood to pay is blank.

  • Potential recovery will be blank when the line hasn't been adjudicated yet thus no Allowed amount exists or when Hybrid expected allowed is blank.

  • Net collection rate and Potential recovery % will be blank when no claim in the report has a Hybrid expected allowed.

Did this answer your question?