Internal medicine carries the widest panels in outpatient care, and the least structure around them.

The RealityA panel is a list somebody owns

⚠️

The Problem

The list is wrong, so nobody uses it.

Patients who left, patients who never established, patients attributed to a physician who retired two years ago. One bad outreach call is enough to make staff stop trusting the report.

🔧

The Solution

One attribution rule, written down.

Decide in writing what makes a patient yours, apply it in the EHR rather than in a spreadsheet, and review the exceptions monthly instead of rebuilding the list each time.

✅

The Resolution

A worklist staff will work.

Outreach goes to people who are genuinely overdue, the list shrinks as it is worked, and the same data supports quality reporting without a second build.

What a patient panel actually is, and why yours is wrong

A patient panel is the set of patients for whom a named clinician holds continuing responsibility. AHRQ treats panel management as a facilitation activity in its practice facilitation handbook, which is a useful framing: something the practice does continuously, not a report it runs once. The definition sounds obvious until you try to express it as a database query, at which point four decisions appear and most practices have never made any of them.

The four decisions nobody wrote down

How recently must a patient have been seen to still count? Two years is common, three is defensible, five is a mailing list. What kind of visit counts: a physical, any office visit, a nurse visit, a phone call? Who owns a patient seen by three different clinicians in the group? And what happens when a physician leaves?

Answer those four in writing and the report becomes reproducible. Leave them to whoever builds the query and you get a different panel every time, which is precisely why the list has no credibility on the floor.

Empanelment: deciding who belongs to whom

Empanelment is the assignment step, and in most EHRs it is a field that was populated once at go-live and never maintained. Someone imported a legacy provider column, the values were partly wrong then, and nothing has corrected them since.

There are two workable approaches. Assign explicitly, using the primary care provider field, and accept that it needs a maintenance routine. Or assign by rule, attributing each patient to the clinician who delivered the plurality of their qualifying visits in the lookback window. Explicit assignment reflects what the practice intends; rule-based assignment reflects what actually happened. Run both once and compare them, because the size of the disagreement tells you how much maintenance the explicit field needs.

Whichever you pick, the field has to live in the EHR and not in a spreadsheet on a shared drive. A panel maintained outside the record cannot drive a reminder, cannot filter a report, and quietly becomes a second source of truth that disagrees with the first. If the assignment logic is being rebuilt anyway, it is usually worth rebuilding the workflow around the list at the same time.

Finding care gaps without buying another system

Most practices reach for a population health module before they have exhausted what the chart already does. Care gaps are, mechanically, a set of intersections: patients with a condition who lack a result inside a window. Every certified EHR can express that. What stops it is usually dirty structured data rather than missing functionality.

Three failures account for most of it. Diagnoses live in the note text instead of the problem list, so the denominator is wrong. Outside results arrive as scanned PDFs, so the numerator misses patients who did have the test elsewhere. And inactive patients were never marked inactive, so the gap list inflates with people who are not coming back.

Fix those three and the built-in reporting usually becomes adequate. Buy a module first and you will be running the same dirty data through a more expensive engine. In our engagements, practices that clean the problem list before they shop for software tend to conclude they did not need the software.

Registry reporting: turning a list into a worklist

A report is something you look at. A worklist is something someone works. The difference is ownership, a disposition field, and a way to see progress.

Registry reporting is where those come together. The same denominators that answer a clinical question also feed quality measures, so a practice that has defined its panel carefully has already done most of the quality reporting work it would otherwise repeat in January. The reverse is also true: practices that build measure reporting first, in isolation, usually discover their measure denominators and their clinical panels disagree, and then have to reconcile two definitions under a deadline.

Practical shape of a worklist that gets used: sorted by how overdue the patient is rather than alphabetically, capped at what the assigned staff member can realistically work in a session, and marked when a patient has been contacted so the next session starts where the last one stopped. Some of that outreach resolves without a visit at all, and some of it is better closed with a virtual visit than with a letter.

Where chronic care management billing fits

Panel work and billing overlap more than most practices realize. Medicare pays for a set of non-face-to-face care coordination services under the Physician Fee Schedule, and Medicare care management services sets out the current fact sheets and billing guidance for them. The eligibility criteria and the documentation requirements are specific, and they are not something to infer from a webinar.

The connection worth noticing is structural. The patients who qualify for chronic care management are largely the same patients a well-built panel surfaces as overdue. A practice that has already defined its panel, cleaned its problem list and staffed an outreach queue has built most of the operational base these services require. One design supports both, provided the documentation requirements are read directly from the source rather than assumed.

We are deliberately not quoting payment amounts or eligibility thresholds here, because they change and a stale figure in a blog post is worse than no figure. Read the current guidance, then decide.

What to fix first

If you are starting from nothing, the order matters more than the pace.

• Write the attribution rule. One page. Lookback window, qualifying visit types, tie-breaker, and what happens when a clinician leaves.
• Clean the problem list before anything else. Every downstream denominator depends on it, and no tool compensates for it.
• Mark inactive patients inactive. Tedious, unglamorous, and the single change that most improves how the list is received.
• Name an owner for the worklist. By name, before the first list is generated, not after the first week of nobody touching it.

None of this is specific to one vendor, which is the point. The work is the same in every major ambulatory platform, and it sits inside our internal medicine EHR and IT support rather than beside it. Because outreach lists move protected health information into spreadsheets, phones and mailings, the design also belongs in your HIPAA risk analysis, not only in the clinical plan. Most of what follows sits inside a broader internal medicine practice management engagement, because attribution disputes are rarely only technical.

One caution on vendors. A panel management EHR internal medicine company that sells you a dashboard before asking how your patients are attributed is selling you a mirror. Ask what happens to the numbers when the attribution rule changes, and ask before signing.

Panel management in internal medicine: common questions