A practice manager we work with keeps a sticky note on her monitor that says "January lies". She is right. Every January a practice gets a new fee schedule, a new set of patient deductibles, a new crop of plan changes from open enrollment, and a handful of payer policy edits that took effect on the first. None of it is visible in January. Claims for January visits are adjudicated in February, appeals and corrections land in March, and by the last week of March you finally have enough remittance data to say what actually happened.
That is why we do a formal first-quarter review with every billing client in the last two weeks of March, and why we think any practice manager should do the same even if nobody is asking for it. It takes an afternoon. It catches problems that would otherwise run for nine more months, and in our experience the problems it catches are almost never the ones people expected.
This year the January changes were bigger than usual. The 2026 Medicare Physician Fee Schedule set the conversion factor at $33.40 for clinicians who are not qualifying APM participants and $33.57 for those who are, an increase of just over 3 percent from the 2025 figure of $32.35. It also applied a 2.5 percent efficiency adjustment to the work component of most non-time-based services, and revalued practice expense by site of service in a way that favors office settings. Commercial payers made their usual January edits. Medicare Advantage plans exited some counties and patients moved. All of that is now in your remits.
Key takeaways
- Pull eight numbers for January 1 to March 31, 2026 and the same eight for the same window in 2025. The comparison matters more than the absolute value.
- Days in AR rising while the denial rate stays flat is usually the deductible reset and a front desk that stopped asking, not a billing failure.
- A commercial payer whose average allowed per visit dropped with no change in visit mix probably loaded the wrong fee schedule. Twenty repriced lines will prove it.
- Every number needs an owner and a next review date. A number nobody owns drifts.
The eight numbers
Pull each of these for January 1 to March 31, 2026, and pull the same number for the same window in 2025. If your practice management system cannot produce a number for a past window, use the month-end reports you saved last year; if you did not save them, this is the year to start.
| Number | How to calculate it | Where we like to see it |
|---|---|---|
| 1. Days in AR | Total AR divided by average daily charges over the last 90 days | Under 40 for most office-based specialties |
| 2. AR over 90 days | Dollars aged over 90 as a share of total AR | Under 20 percent; insurance AR over 90 under 12 percent |
| 3. Net collection rate | Payments divided by (charges minus contractual adjustments), on claims old enough to be resolved | 96 percent or higher |
| 4. First-pass resolution rate | Claims paid on first submission divided by claims submitted | 90 percent or higher |
| 5. Denial rate | Denied claim lines divided by submitted lines | Under 8 percent; under 5 is very good |
| 6. Charge lag | Median days from date of service to claim submission | 3 days or fewer for office visits |
| 7. Patient collection at time of service | Copays and known balances collected at check-in divided by amounts due | 85 percent or higher |
| 8. Average allowed per visit by payer | Allowed amount divided by visit count, for each of the top five payers | Should match your contracts, not drift |
Pulling the numbers without fooling yourself
Three of the eight are easy to get wrong in ways that make a bad quarter look fine. Days in AR should exclude credit balances; a practice carrying $40,000 in unapplied patient credits shrinks its AR on paper and hides a refund problem. Net collection rate has to be measured on claims that have had time to resolve, so use dates of service from four to six months ago, not last month; a net collection rate on fresh claims will always look low and mean nothing. And the denial rate should be counted on lines, not claims, and should exclude clearinghouse rejections, which never reached the payer. A rejection is a data problem. A denial is a payer decision. They have different owners.
Charge lag deserves its own note. Most systems report the average, and the average is pulled down by the hundreds of same-day visits. Ask for the median and for the count of encounters not yet billed after seven days. Ten unbilled encounters from a single provider is a provider who is not closing notes, and that is a conversation, not a report.
Reading the numbers against what changed in January
Days in AR up, denial rate flat. This is usually the deductible reset. January and February visits leave more balance on the patient side, and patient AR moves slowly. Check number 7 before you blame the billing team. If time-of-service collection dropped from 88 percent to 70, the front desk stopped asking or the eligibility check stopped returning the remaining deductible.
Denial rate up with CO-27, CO-26 or CO-31 codes. Coverage terminated, expenses incurred before coverage, or patient cannot be identified. These are eligibility denials and they spike every Q1 because plans changed on January 1. In 2026 we are seeing more of them than usual in counties where Medicare Advantage plans withdrew, because patients who did not pick a new plan defaulted to traditional Medicare and did not tell the office. The fix is a real-time eligibility check within 48 hours of every visit, not a batch check the week before.
Average allowed per visit down for Medicare. Expected for procedural specialties because of the 2.5 percent efficiency adjustment. Not expected for primary care and other E/M-heavy practices, where the conversion factor increase should show as a small rise. If your 99213 and 99214 allowed amounts went down, check that your Medicare Administrative Contractor has your correct locality and that claims are not being paid at the facility rate because of a place of service error.
Average allowed per visit down for a commercial payer. This one deserves the most attention. Payers load new fee schedules in January and sometimes load the wrong one. Compare five paid 99214 lines from March against your contracted rate. If they are all short by the same percentage, that is not noise. We cover the full variance method in a separate article this spring.
First-pass rate down, denial rate up, and the denials are CO-4 or CO-16. Modifier or missing information problems. Usually a January coding or claim scrubber change that nobody tested. Pull twenty examples and look for a pattern by CPT code or by provider.
Net collection rate down with everything else flat. Look at the adjustment codes. A biller who started writing off balances to a generic adjustment code to clear a queue will move this number before anything else does. Pull adjustments by code and by user for the quarter and look for one that grew.
A worked example
A four-provider family medicine practice ran this review last week. Days in AR had moved from 34 to 41. AR over 90 was flat. The denial rate was flat at 6 percent. Time-of-service collection had dropped from 86 percent to 64 percent. Average allowed per visit for Medicare was up slightly, as expected. Average allowed for their largest commercial payer was down 4 percent with no change in visit mix. Charge lag was unchanged at two days.
Two findings, two owners. The front desk had stopped collecting toward deductibles in January because the eligibility response from the practice management system had been changed to a summary view that did not show the remaining deductible. That was a settings fix, made the same afternoon. The commercial payer had loaded a 2024 fee schedule instead of the 2026 one on the January 1 amendment. The practice sent a spreadsheet of forty underpaid lines to the provider relations representative with the contract page attached. Whether the payer reprocesses without a formal project depends on the payer, but the ask was made in March instead of November, and the dollars at stake for the quarter were about $6,000 that would otherwise have closed as paid.
Notice what the review did not find. The billing team was not the problem. Both leaks were upstream of them, and both would have been invisible on a denial report because neither one produces a denial.
Questions we hear
Our system cannot produce last year's numbers for the same window. What do we compare against?
Use the closest thing you have: the month-end AR aging reports from March 2025 if they were saved, or the fourth quarter of 2025 as a baseline. A baseline from a different season is imperfect because of the deductible cycle, but it is better than nothing. Then save this quarter's one-page summary so next year is easy.
Which of the eight matters most if we only have time for one?
Average allowed per visit by payer, because it is the only one nobody else is watching. Denials generate work queues. Aging generates statements. An underpayment generates nothing at all, and it repeats on every claim until someone compares the allowed amount to the contract.
Should the review include the providers?
Yes, but with one page and three findings, not the whole spreadsheet. Providers respond to "your charge lag is nine days and the practice standard is three" and to "your Medicare allowed per visit went down because of a place of service setting". They do not respond to a KPI dashboard, and honestly, most people do not.
What to do this month
- Pull the eight numbers for Q1 2026 and Q1 2025 side by side. One page.
- For any number that moved more than 10 percent in the wrong direction, pull twenty example claims and find the pattern before you propose a fix.
- Reprice a sample of paid claims from each of your top five payers against the contract. Do this every quarter, not every renewal.
- Confirm the front desk can see the remaining deductible in the eligibility response and is asking for it.
- Write down who owns each number and when it is reviewed next.
- Save the one-page summary where next year's reviewer will find it.
If you would rather have someone else do the pulling and the pattern-finding, an RCM audit covers all eight numbers and a 90-day claim sample. Practices that use Revelrex for medical billing get this review as part of the monthly report.
