+$8,241
more care management revenue in July than in May
1,336 care management patients billed in May, 1,187 in July
Care management revenue rose every month from May to July while the number of APCM patients fell from 952 to 749. The practice moved patients to CCM, which pays more per patient. Counting patients would call that a decline; counting revenue is the right measure, and Dr Shanklin was the one who said so.
July was the best care management month on record, up 10.9% since May
$75,435 in May, $78,645 in June, $83,676 in July, while APCM patient counts fell. Revenue per care management patient went from $56.46 to $70.49. Fewer patients, more per patient, more in total. Allowed amounts at the Rest of Texas fee schedule, before the 20% patient share; not collections.
Of everyone who could be earning you a care management fee this month, here is where they are. Each step loses people for a different reason, and only some of those reasons are fixable by you.
The 550 enrolled-but-not-billed is the number to look at first, and most of it is not a mistake: 418 were correctly billed CCM instead, since the two cannot both be billed in one month, and 271 are Medicare Advantage where payment depends on your contracts. What is left after that is 80 with no explanation, of whom 37 had no claim of any kind submitted in July. The other side of the funnel, the 745 not enrolled, is a conversation at the next visit for the 123 seen recently and a campaign for the rest; the Wellness visits and Before the visit tabs are where the practice acts on it.
The chart that corrected this report. Counting patients per program answers "which code did we use". Only revenue answers "are we better off", and here they moved in opposite directions.
Where the July increase came from. Level 3 APCM (G0558) went from 1 claim line in June 2026 to 224 in July 2026. It pays $61.65 a month more than level 2. That single change lifted revenue per care management patient from $56.46 in May to $70.49 in July, which is why total revenue rose even though 149 fewer patients were billed.
The same window by program rather than by dollars. APCM grew from nothing to 952 patients a month and then fell as CCM picked up. This is the view that produced the wrong headline, kept here because it is still the right way to see a substitution.
The substitution is real and it is visible patient by patient: 224 patients billed APCM in May were billed CCM in June, and 214 of them in July. CCM's base code pays $64.36 against APCM level 2's $52.27, so moving a patient across is worth more, not less. What the substitution does NOT explain is the 234 patients billed APCM in May with no care management billing of any kind in July - they did not move to CCM, they stopped. The rigorous subset of that is on the next tab.
eCW's PHM module reports 262 failed APCM claims for July and will not say which. Working from the roster and thirteen months of claims, here is the whole 832, sorted by whether anything is owed.
| Reason | Status | Patients | Value |
|---|---|---|---|
| Billed CCM that month instead - correct - the two are mutually exclusive | resolved | 418 | - |
| Medicare Advantage - depends on your contracts, which is an open question | verify | 271 | - |
| No consent date on record - fixable going forward, not billable retroactively | open | 63 | - |
| No claim of any kind submitted - not denied, not unpaid - nothing was sent | actionable | 37 | $1,135 |
| A claim exists but we cannot see its lines - report 361.02 has no CPT column | verify | 43 | - |
Our count does not match eCW's 262, and should not: theirs counts claim failures, ours counts patients with no claim. Different questions. The 37 and the 63 are named rows in the practice's work queue, one click from a telephone encounter; the 271 Medicare Advantage rows wait on the contract answer on the By payer tab.
A Medicare patient sat in your office in July with no annual wellness visit in the prior twelve months. It is a covered benefit with no coinsurance and no deductible, and it recurs every year for the life of the patient.
| Status | Status | Patients |
|---|---|---|
| No wellness visit in the visible claims history | open | 124 |
| Last one more than twelve months before the visit | open | 11 |
| Likely false positives from the short lookback - already deducted from the headline | verify | 20 |
| Gap already closed since the July visit - no value, stays off the call list | resolved | 20 |
135 open, less the 20 likely false positives, is the 115 on the front of this page, worth $16,041 at $139.49 each. Claims history starts 2025-08-01, so the twelve-month lookback is really eleven months. Patients whose only wellness visit fell in the missing month look like they never had one. We measured the equivalent cohort one month later and found 20, which is deducted above. Payments history back to 2024-08 removes the doubt entirely.
All finding types combined, attributed to the panel PCP rather than whoever rendered the visit. Unattributed means the roster and the exports name no PCP for the patient. The largest lines are the largest panels, not the weakest providers; a rate per patient is on the practice's dashboard.
| Provider | Patients | Value |
|---|---|---|
| Shanklin, Christopher L | 587 | $14,131 |
| Reed, Jubel K | 428 | $10,259 |
| Backlund, Amanda J | 248 | $5,877 |
| Fletcher, Jamie A | 245 | $5,468 |
| (unattributed) | 223 | $3,893 |
| Rosson, James D | 170 | $5,043 |
| Lust, Micah D | 103 | $2,835 |
| Cornwell, Darlene MARIE | 102 | $1,659 |
Whether Medicare Advantage pays G0556-G0558 is still an open contract question, and a large share of everything on this page turns on the answer.
| Payer class | Patients | Value |
|---|---|---|
| Traditional Medicare - the population an ACO is scored on | 1,406 | $35,099 |
| Medicare Advantage | 951 | $25,343 |
271 unbilled patients and 281 of the unenrolled turn on one answer. If your contracts pay these codes it is the largest line here. If they do not, those patients were correctly not billed and should stop being counted.
Read the way an ACO reads it: rates with their denominators on the page, and what each one feeds. The first three are MIPS quality measures scored from labs and vitals in the chart, so an A1c resulted in the office counts the day it is resulted; the claims feed the ACO works from cannot see the value. The rest are the reconciliation findings from the other tabs, as rates. 1,453 Medicare-class patients are in at least one measure denominator.
294 met, 117 gap, of 411 patients read as of 2026-09-06; 1 not yet read. Patients 18 to 75 with diabetes on the problem list; met when the most recent HbA1c in the year is at or below 9. No result in the year counts as poor control.
A result done in the office counts the day it is resulted. The ACO's claims feed sees the visit and not the value, so this is the list to attest.
328 met, 215 gap, of 543 patients read as of 2026-09-06; 1 not yet read. Patients 18 to 85 with diabetes; met when both an eGFR and a urine albumin-creatinine ratio were resulted in the year. The uACR is the one that gets missed.
Every metabolic panel carries an eGFR; the uACR has to be ordered. The gap list is an order at the next visit, not a workup.
1,132 met, 251 gap, of 1,383 patients read as of 2026-09-06; 3 not yet read. Patients 18 to 85 with hypertension; met when the most recent blood pressure in the year is below 140/90.
The last reading of the year is what scores. A high reading at the last visit is an open item until a lower one is documented.
1357 Medicare patients seen in July; 115 of them with no wellness visit in the prior twelve months, after deducting the short-lookback estimate.
Attribution runs through primary-care services, and the wellness visit is the one a Medicare patient reliably gets each year. A patient with no AWV is a quality miss this year and an attribution risk next year.
1737 enrolled of 2482 known eligible. The denominator is every patient with a claim in thirteen months, Medicare-class coverage, and two or more chronic condition groups on the eCW problem list.
Enrolled chronic patients cost less. Every qualifying patient outside a program is unmanaged utilization landing in the benchmark.
80 enrolled patients with no explanation for the missing claim, of 1737 enrolled; 37 had no claim of any kind. The other 752 unbilled were CCM instead, Medicare Advantage, or missing consent.
The practice did the work and nobody was paid. Leakage is the gap between the program a practice runs and the one the ACO can point to.
APCM level 3 was 0.1% of APCM patients in June and 29.9% in July. Level 3 requires QMB status.
A code's share of a program moving twenty points in a month is either a deliberate change or a configuration change nobody reviewed. An auditor sees the same chart a year later; the ACO should see it first.
23 billed at level 1 with two or more strictly chronic diagnoses; 222 billed at level 3 with no Medicaid-class coverage on record - a question for billing, not a finding of miscoding.
Level is a definition, not a judgment. Level 1 for a patient with three chronic conditions is money left behind; level 3 without QMB status is money that may have to go back.
Across 7,772 problem lists read from eCW, the whole panel.
E11.9 with documented complications is a risk score understated. Nothing in fee-for-service pays for the fix; the benchmark does.
The full-panel problem-list pull landed on 2026-09-05: 7,772 patients read from eCW, every account with a claim in thirteen months. Penetration is now measured against the whole panel rather than eCW's own queue, and diagnosis specificity across every problem list rather than the enrolled ones. Earlier versions of this page carried narrower denominators and said so. The same six measures, one row per practice, are at /reports/aco.
On the call you described half of Lisa's day: read tomorrow's schedule, open each chart, and type a reminder into the chief complaint so the provider does not forget. We found those notes in eCW, measured them, and built the same review to run every night. This is what it found for the practice's Tuesday, September 8 schedule, 12 providers, as a dry run: nothing was written to a chart. Dr Shanklin's own day: 19 visits, 9 patients with 11 open items.
| Item | Status | Patients |
|---|---|---|
| Care management consent not on the roster - Medicare, two or more chronic conditions, no APCM or CCM enrollment | actionable | 33 |
| Wellness visit due - none billed in the twelve months before the visit | actionable | 6 |
| HbA1c missing or above 9 - diabetic with no result in twelve months, or poor control | open | 1 |
| Kidney health evaluation missing - diabetic without both a uACR and an eGFR in twelve months | open | 4 |
| Blood pressure above goal - hypertensive whose last reading was at or above 140/90 | open | 2 |
| Already in the chief complaint - a person had been there first; left alone | resolved | 7 |
38 of 196 patients had at least one open item, 50 items in all, worth $2,475 in fees on the two that carry a fee. The rest are quality measures the ACO scores you on. Delivery is one telephone encounter per patient to the visit's own provider, the night before, in the same words a nurse would use. What it cannot do yet: diabetic eye exams, immunizations and mammograms, because eCW does not expose those to the app. Those stay with Lisa.
By provider, open items: Shanklin 11 open across 19 visits; Reed 9 open across 26 visits; Cornwell 8 open across 10 visits; Backlund 5 open across 16 visits; Bean 5 open across 22 visits; Lust 4 open across 18 visits; Fletcher 3 open across 20 visits; East 2 open across 17 visits; Fornari 1 open across 10 visits; Gordy 1 open across 4 visits; Hoffman 1 open across 19 visits.
What is new here, what was always available to you, and what would make these numbers firmer.
Already available to you without this: the pending enrollment queue, last wellness visit dates in the eBO reports, and the count of failed APCM claims.
Only reachable here: which APCM patients failed and why, split four ways, that the program peaked and turned down, and the nightly pre-visit review on the previous tab.
What has been verified rather than described: reading the chart through eCW's FHIR API for every resource the app is granted; writing a telephone encounter back and reading it again by its reference; the CY2026 fee schedule for your locality computed from CMS inputs; a full-panel pull of problem lists across every patient seen in thirteen months. Every view of these figures by a signed-in person is recorded.
Re-run monthly against trailing service dates, since claims lag date of service by weeks. Patients whose gap closes between runs are marked resolved rather than deleted, so what got fixed stays visible.
Sources: CCM Enrolled Patients Queue (1,737), CCM Enrollment Queue (993 pending), Detail Report Payments (223,252 rows, 2025-08 to 2026-08), 361.02 Billed Charges, eCW FHIR problem lists and coverage, CMS PFS CY2026 locality 0441299.