“Visits are flat across the group.” The pack says so, and an operating partner reads it as one fact about twelve practices.
It’s twelve facts, and they don’t agree with each other, because none of the twelve counts a visit the same way.
Here’s the answer: medical practice portfolio reporting only works when every practice counts the same stops the same way.
Until a completed visit, a clean claim, and a collected balance mean the same thing everywhere, the rollup describes nothing that happened at any of them.
One number. Twelve different clocks.
Practice A counts a completed visit at check-out, Practice B counts it when the note is signed (clinical documentation finalized), and Practice C counts it when the charge posts (billable line recorded).
All three report visit volume monthly, the group rolls them up, and the resulting number describes nothing that happened at any of them.
Between a booked appointment and a zero balance, every practice runs the same seven stops: follow-up scheduled, note signed, charge created, claim out clean, denial worked, payment posted, patient share collected.
Every stop can be counted at one event or another, and every practice picked its own.
Related: portfolio oversight, appointment status, charge lag.
Every metric drifts the same way
Visit count is the obvious one, but the rest of the pack has the same problem, quieter.
Clean claim rate (share of bills accepted without a reject or return, or paid on the first try, depending on who you ask). One practice measures it at the billing vendor, counting what the vendor accepted, while another measures it at the insurer, counting what got paid. The first number runs higher every month, and it isn’t a better practice.
Collections. One practice books what the desk took on the day the desk took it, while another books it a month later when the deposit clears. Same money, two months, and the group’s collections line shows a swing that never happened.
Days in A/R (accounts receivable, average days billed money sits unpaid). One practice ages from the date of service, while another ages from the date of the last submission, so a claim sent again gets younger while the money gets older. Two practices with the same real backlog will show different ages.
Denial rate (share of claims the insurer refuses). One practice counts first-pass refusals and another counts only final outcomes, so the quoted rate can’t be compared until the event matches.
Each number is honest on its own terms, but none of the terms match.
Why this is the first casualty of a roll-up
A roll-up (the strategy of buying practices and combining them) acquires definitions along with the practices.
Every one arrives with its reports already shaped by whoever set up its system years ago, and nobody at the practice thinks of those choices as choices. They’re just how the number is run.
Standardizing the report template doesn’t fix it, because twelve practices can fill in the same spreadsheet with numbers that still mean twelve things, and the pack looks more comparable than it is.
Until the definitions match, the group can’t do the one thing it exists to do: tell which practice is drifting.
What to standardize first (simple set)
Three decisions, made once, for the whole group.
- Pick the event. For each of the seven numbers, name the single event at which it is counted, at every practice. A completed visit is counted when the charge posts, or when the note is signed, but the same one everywhere. A clean claim is measured at the insurer. A collected balance is booked the day it is taken.
- Measure upstream (early in the chain, before the claim goes out). The numbers that predict next quarter sit there: charge lag (days from visit to billable charge), late charges, unbilled care, unsent claims, unsigned notes. Define those five the same way and produce them at every practice, from records each practice already holds.
- Produce them from the records, not from the practice. A number the practice reports about itself is the practice grading its own work. A number produced the same way from each practice’s own system, by the same method, is a measurement.
Start with visits this week, then charge lag, then clean claim rate. One metric at a time beats a full rebuild that never finishes.
How a bad definition creates a bad decision
A site looks “behind” on visits because it counts at charge post while peers count at check-out, so you staff volume when the real stop was notes.
A site looks “best” on clean claims because it measures vendor acceptance while peers measure insurer pay, so you copy the wrong playbook.
A collections swing appears after one site books desk payments a month late, so you open a collections project. Nothing changed at the practice: the number had two definitions inside one column.
An aging report built on the latest submission date makes every resubmitted claim look young, while the money is older than the report says. The group compares backlog on a clock that resets itself.
Write the definition in one sentence
For each metric you lock this week, write one sentence anyone can read cold:
“A completed visit is counted when [event] at every practice.”
“A clean claim is counted when [event].”
“Days in A/R ages from [event].”
If the sentence needs a footnote, the definition isn’t locked yet.
What happens after visits are locked
Drift becomes visible. When a completed visit means the same thing at every practice, the practice whose visits are quietly falling is visible in the week it starts, not the quarter it surfaces.
The review changes shape: a variance arrives with the cause attached, because the number that moved is a real number and the step it points at is a real step. The conversation becomes a decision about who fixes which step this week.
Benchmarking becomes honest: the practice at the top of the collections table is at the top because it collects, not because it counts at the desk.
That’s when portfolio oversight can assign owners instead of arguing about whose spreadsheet is right.
Why definitions beat dashboards
A dashboard with forty tabs and no shared event is twelve stories with charts, while a short list of shared events is one instrument.
You can build the dashboard later: definitions first, tools second.
What to do this week
- Pick one metric: completed visits.
- Ask each of three sites what event they count.
- Write the one-sentence group definition.
- Rebuild last month’s visit total under that definition at those three sites.
- Compare. The gap between old pack and new count is the noise you have been managing.
Expand to charge lag and clean claim rate once visits hold for two weeks.
Reporting is the dictionary. Oversight is the assignment.
Reporting aligns definitions so numbers mean the same thing, and oversight assigns the action when a stop breaks.
You need both: reporting without oversight prints green averages, and oversight without reporting staffs the loud site for the wrong reason.
Common pushback, and the short answer
“Our EHRs are different.” Definitions are events, not software. Start with the same questions, and systems catch up.
“We already have a consolidated pack.” A pack with twelve clocks isn’t consolidated. It’s stacked.
“This will take a year.” One metric at three sites takes an afternoon, and you expand from there.
Start ugly. Stay consistent.
The first shared count will look worse than the old pack at some sites.
Keep the method fixed for a month before you decorate the dashboard.
What this means for you
Before you compare two practices, ask what each number means at each one. If the answer differs, the comparison is noise, and every decision built on it inherits the noise.
Pick the event, measure upstream, and produce the numbers from the records the same way everywhere. Then the pack turns from twelve stories into one instrument.
Grab 30 minutes with us. Prep nothing. You will see which of your practices’ numbers mean the same thing, and which only look like they do.
Questions people ask
Why don’t the numbers from my medical practices match?
Because each practice counts the same number at a different event. One counts a completed visit at check-out, another when the note is signed, and a third when the charge posts. Clean claims get measured at the billing vendor at one site and at the insurer at another. Each number is honest, but the definitions were never lined up.
Does a standard report template fix mismatched numbers across practices?
No. Twelve practices can fill the same template with numbers that still mean twelve things. Comparability comes from naming the event each number is counted at, the same event at every practice, and producing the number from each practice’s own records the same way.
What should a medical practice group standardize first?
Start with completed visits: pick the one event at which a visit counts, and use it at every practice. Then add charge lag (days from visit to billable charge) and clean claim rate. From there, lock all seven numbers along the money path, plus the early ones like unsigned notes and unsent claims, which predict next quarter.
How do I know a practice is drifting rather than just different?
Only after the definitions match. Once a visit, a claim, and a collected balance mean the same thing everywhere, a practice whose number moves against its own baseline is drifting. Before that, a move can be a real change or a difference in counting, and the pack can’t tell you which.
Can the practices’ own systems produce comparable numbers?
Yes. Every practice already records the events. The work is producing the same measurements from those records by the same method at every practice, rather than accepting whatever standard report each system happens to run.