The panel is the denominator
Risk scores, recapture rates, quality rates, per-member costs, and loss ratios are all measured per attributed member. If the panel is wrong, every one of them is wrong, and nobody finds out until the plan reconciles.
Guide · Attribution
Patient attribution: who is on your panel, how they got there, and why the list moves.
A working guide for practice administrators, analysts, and data stewards: what attribution is, how a payer roster and a claims-based method each decide it, why the panel changes every month, the three lists every practice has, what a wrong panel does to everything downstream, and the monthly routine that keeps it right.
By Jonan Portal Alvarez, Chief Executive Officer, Alara Healthcare (management services) · Updated September 10, 2026 · About 8 minutes
The definition
What attribution is
Attribution answers one question: which provider is accountable for this member? Every panel count, every gap list, every per-member figure, and every dollar of a value-based contract starts from that answer. It is also the determination people dispute most, because it rarely matches the list of patients a practice believes it has.
The answer comes from outside the practice. Under a Medicare Advantage or commercial risk contract, the health plan names the provider on its monthly roster. Under an accountable care program, a computed method derives the assignment from where the member actually received primary care over a lookback window. Either way, the panel is read from files somebody sent you, it moves every month, and a member who leaves is learned about one roster late.
It is not the patient list
The scheduling system knows who walked in. The roster knows who the contract pays for. The two overlap, and the differences are where money and effort go missing in both directions.
It is not a performance measure
The panel grows and shrinks with the plan's assignments and the members' choices, not with anything the practice did. Reading panel growth as success, or churn as failure, misreads the file.
The methods
How members are assigned
| Method | Who decides | How it works | What to watch |
|---|---|---|---|
| Payer roster | The health plan. | Each monthly enrollment roster names, for every member, the provider they are assigned to. The member chose at enrollment, or the plan assigned them. The panel is read directly from the file. | An attribution you disagree with is a conversation with the plan about its roster, not a calculation to appeal. Open the roster month in question and the answer is there. |
| Shared savings assignment | CMS, from claims. | Beneficiaries are assigned to the accountable care organization whose participants delivered the plurality of their primary care over a lookback window, in a two-step method that looks at primary care clinicians first. Preliminary lists arrive during the year and the final list is reconciled after it. A beneficiary can also designate a primary clinician voluntarily. | The list you work from mid-year is preliminary. Members appear and disappear at reconciliation, and effort spent on someone who ends the year assigned elsewhere earns nothing under the contract. |
| Claims-based plurality | The program's rules, from claims. | The practice that delivered the most primary care over the lookback window is attributed the member, with a voluntary alignment option in some programs. Lists refresh on the program's schedule, often quarterly. | The lookback window decides everything. A member seen once eighteen months ago can still be yours, and a new patient seen twice may not be yet. |
The calendar
Why the panel moves every month
Every roster
Members arrive, members leave, and some change provider inside the plan. A member leaves by disappearing from the next roster, so the month someone leaves and the month it can first be counted are never the same month.
January
The plan year resets enrollment, and the panel changes more than in any other month. Open enrollment choices land, plans and products change, and a practice can gain or lose a meaningful share of its panel without a single visit having happened.
Mid-year
Retroactive terminations clear, deaths and moves catch up with the file, and computed programs publish updated preliminary lists. The panel you thought you had in March is not the panel the year will be paid on.
After the year ends
Computed programs reconcile the final assignment, and the plan's final adjustments restate months already paid. Only now is the year's panel the year's panel.
Two numbers describe the same panel, and both are right. Total members counts everyone attributed at any point in a period, including people who left partway through, so it agrees with the revenue and cost figures beside it. Active members is the point-in-time headcount, always smaller. The gap between them is churn, and the volatility rate says how many members turn over within a few months of arriving.
The reconciliation
The three lists every practice has
Put the roster beside the practice's own records and every member lands in one of three lists. The monthly reconciliation is working all three.
Attributed and seen
The members the contract pays for who are also patients of the practice. This is the panel the work is measured on: gaps, visits, recapture, quality. It should be most of the roster. When it is not, the other two lists say why.
Attributed and never seen
Members the plan assigned who have no visit on record. They carry a risk score, a set of quality measures, and a cost, and nobody has met them. They are the first outreach list of every year, because a member with open conditions and no visit is where value and care both go missing.
Seen and attributed elsewhere
Patients the practice treats whose contract assigns them to another provider. Their gaps do not count for you, their costs do not land on you, and their visits do not move your rates. Some of them want to be your patients and never told the plan. That is a conversation with the member and the plan, not a data fix.
- 1
Load the roster before anything else
The roster month is the unit. Load it, confirm the method note, and do not let gap work start on last month's panel.
- 2
Reconcile identities
The roster's member number, the Medicare identifier, and the practice's chart number rarely match on their own. Resolve the same person across them, and queue the ambiguous pairs for a human decision instead of guessing.
- 3
Map providers, not just members
The roster names a provider by the plan's own code. That code has to map to a clinician, a location, and a national provider identifier in the practice, or the provider panels are wrong even when the member list is right. A clinician who left keeps collecting members until the mapping is fixed with the plan.
- 4
Work the movement
New members go to onboarding outreach. Terminated members leave the queues, and their open items are closed as no longer attributed, not as done. Members who moved between providers inside the plan move between panels the same month.
- 5
Read the shape
Age and sex, the split by provider, and where members live decide staffing and scheduling. A panel that skews older, or clusters in one county, needs a different visit plan than the average suggests.
The consequences
What a wrong panel does downstream
When the panel is right
- Recapture denominators are built from the members who will actually be paid on.
- Quality rates count the eligible members the plan will score.
- Per-member costs and loss ratios reconcile to the plan's statements.
- Outreach reaches the people the contract is accountable for.
- Provider-level figures describe the clinicians who actually carry the members.
When the panel is wrong
- Coders chase conditions on members who left, and miss the ones who arrived.
- Quality effort lands on patients the contract does not score, while scored members go unseen.
- Revenue figures disagree with the plan's settlement, and nobody can say by how much.
- Provider dashboards blame the wrong clinician, or one who no longer works there.
- The year-end reconciliation becomes the first time anyone learns the size of the problem.
Tools versus operations
What Panel 360 does, and what it does not
Panel 360 does
- Reads the panel from the roster, one record per member per enrolled month, so it can answer both how many over the period and how many right now.
- Names the attribution method in force and stores the reasoning for every member's assignment, so any attribution can be explained.
- Counts new, retained, and terminated members each month, and the volatility rate.
- Splits the panel by provider, by age and sex, and by county.
- Feeds every other module the same panel, so the gap list, the quality rates, and the money agree on who is in.
Panel 360 does not
- Decide who belongs to whom. The plan or the program does. A disputed attribution is a conversation with them.
- Know that a member has left before the next roster says so.
- Make the practice's patient list and the roster the same list. It shows the difference and hands the three lists to the people who work them.
- Fix a provider mapping the plan has wrong. It shows the mismatch, and the steward calls the plan.
Before any gap work starts, Alara Healthcare reconciles the plan roster against the practice's own records, so the panel everyone works from is the panel the contract pays for. Risk adjustment services
Questions people ask
Questions about attribution and the panel
Our member count does not match the headcount from our enrollment system.
That is almost always a period-versus-moment difference. Total members counts everyone attributed at any point in the period, including people who left, because the money figures beside it count their premium too. Active members is the point-in-time headcount and is always smaller. Compare the other system's number with active members, and use the larger figure for anything financial.
Why does a member who left still show in the panel?
Because departure is confirmed by absence from the next roster. The month a member leaves and the month that departure can first be counted are never the same month. The record is closed with its end date rather than erased, so the history stays intact.
Can we appeal an attribution?
Under a payer roster there is nothing to appeal; the plan's file is the answer, and a member who wants to be your patient changes their assigned provider with the plan. Under a computed method the rules decide, and the levers are the primary care the member actually receives from you and, where the program allows it, the member's voluntary designation.
A patient we have treated for years is attributed to another practice. Why?
Either the member is enrolled in a plan that assigned them elsewhere and never updated it, or the lookback found more primary care somewhere else. Until that changes, their gaps, costs, and visits count for the other practice. Ask the member to update the plan, or under a computed method keep delivering the primary care, and the next list will show it.
What is a good volatility rate?
There is no target. The rate says how many members turn over within a few months of arriving, and it moves with the plan's enrollment cycle more than with the practice. What matters is what happens to the new members: whether they get a visit before they leave, and whether the terminated ones leave the queues so nobody works gaps on people who are gone.
How often should we reconcile the panel?
Every roster, which means monthly, and on the day the January roster lands, because that is the month the panel changes most. A reconciliation done quarterly means up to three months of gap work on a panel that has already moved.
Bring a roster and your patient list to the demo.
We will lay them side by side, build the three lists, and show what the panel the contract pays for looks like against the one you thought you had.