You kept your program payments instead of splitting them with a Care Transformation Organization, and you carry the care management in house to earn that. What a CTO would hand you — monitoring labor, device logistics, documented escalation, automated claims — CoachCare supplies without taking a share of those payments, funded out of remote monitoring reimbursement that your program payments do not offset.
Twenty-four-month projection for a remote monitoring program on a 2,500-patient high-risk cohort — the recommended Year-1 scope, not the full panel. Avoided admissions shown at the model's conservative 20% baseline; the realistic figure for this cohort is roughly double. Illustrative, modeled — verify against practice data. Avoided-cost figures are clinical and system value; they are not added to the practice margin.
Most practices we model have to be convinced that longitudinal care management is worth doing. You have been doing it, under a CMS advanced primary care model, without an institution's infrastructure behind you. Everything below is verified against a primary source.
Maryland's 2026 participating-practice list carries Annapolis Internal Medicine at 116 Defense Highway, Suite 400, with the CTO field reading “No CTO Partner.” Practices that do partner with one split their program payments with it. You kept yours. (MDH, participation as of 1 Jan 2026)
Annapolis Internal Medicine, LLC is a participant in an Enhanced-track Medicare Shared Savings Program ACO — the highest-risk track, shared savings and shared losses — in a physician-led, low-revenue ACO, since January 2025. (CMS PY2026 Shared Savings Program participant file)
Maryland's own program materials feature an AIM physician and an AIM care manager presenting to the MDPCP community. The clinical model is not a hypothesis here — it is five years of practice. (MDH CTO materials, 2021)
Eight internal medicine physicians and fourteen advanced practice providers — including gerontology and certified diabetes education capability — on a single campus, with a ninth physician joining September 2026. Physician tenure runs from 1991 to the present. One location, one workflow to configure.
Anne Arundel County holds 111,257 Medicare-eligible residents. Original Medicare enrollment rose 3.1% year over year while Medicare Advantage fell 1.4% — the fee-for-service population these codes bill on, and that the ACO attributes, is expanding. (CMS enrollment, Apr 2025–Apr 2026; penetration 19.8%, Jul 2026)
CoachCare's athenahealth integration is bi-directional and is the only care-management integration that generates claims automatically. With no CTO supplying billing support, that is the difference between a program that runs and one that stalls on paperwork.
Two things changed this year, and both cut the same way: Maryland's advanced primary care program was extended rather than retired, and the fee schedule finally pays for the partial months that make up most of real-world monitoring.
On 1 January 2026 the program was folded into the AHEAD Model as MDPCP-AHEAD, continuing through 2028, with AHEAD itself running to 2035. Roughly 460 practices and ~350,000 attributed Medicare beneficiaries statewide. The runway is a decade, not a year.
99445 and 99470 pay for 2–15 days of device data and 10–19 minutes of management. A patient who transmitted twelve days used to be unbillable. Now the partial month counts — and partial months are most months.
In the forecast on this page those two codes alone carry $288,710 of net reimbursement — 17.5% of the total. Two years ago that revenue did not exist.
The Preserve Telehealth Access Act of 2025, effective June 2025, removed time restrictions on telehealth reimbursement across Medicaid and commercial insurance. The delivery model underneath remote care is no longer waiver-dependent in this state.
And your Original Medicare population is growing — up 3.1% in the county last year while Medicare Advantage fell 1.4%.
You carry two-sided risk. New Part B claims for attributed beneficiaries raise your measured total cost of care, so a gross-revenue pitch is worth less to you than it looks. Here is the arithmetic — and the reason we are proposing a monitoring-led program rather than the larger number we could have shown you.
| 24-month, directional — recommended cohort | At 20% admission baseline | At 40% baseline |
|---|---|---|
| Part B claims added | +$1,646,725 | +$1,646,725 |
| Part A avoided (106 vs 213 admissions at $15,000) | −$1,595,585 | −$3,191,170 |
| Net change in total cost of care | +$51,140 essentially neutral | −$1,544,445 strongly favorable |
| Fee margin retained by the practice | +$702,033 | +$702,033 |
Enrolling your whole panel would generate a larger headline and a worse outcome. Monitoring a stable, well-controlled hypertensive produces a monthly charge against your benchmark and very little avoided utilization. Enrolling the cohort where admissions actually happen produces most of the clinical benefit for a fraction of the claims exposure.
That is the recommendation: heart failure, CKD stage 3b–4, COPD, and poorly controlled hypertension and type 2 diabetes first. Roughly 2,500 patients — about half the Medicare panel. Expand only where the data earns it. The panel-wide figure appears later on this page as a ceiling, so you can see it — not as a plan.
The avoided-admission engine assumes a generic 20% annual admission rate — a whole-population figure. A genuine heart-failure and advanced-CKD cohort runs well above that, so the model penalizes the very targeting it should reward:
| Baseline admission rate | Avoided | Avoided cost |
|---|---|---|
| 20% — model default | 106 | $1,595,585 |
| 30% | 160 | $2,393,378 |
| 40% — realistic for this cohort | 213 | $3,191,170 |
| 50% | 266 | $3,988,963 |
Even at the conservative 20% baseline the added claims are almost exactly offset. At a realistic 40% the program is net favorable to your benchmark by roughly $1.5M before a dollar of margin is counted. Your county's Medicare readmission rate is 19.9% and its emergency-department rate 483.5 per 1,000 beneficiaries, so there is real utilization to move. (CMS Geographic Variation, CY2024)
The restriction applies only to attributed beneficiaries. Attribution follows plurality of primary care, so you carry a Medicare population outside it for whom chronic care management is billable. Sizing that group is a discovery question, and we have put no number on it.
Chronic care management remains billable to Medicare Advantage plans and commercial payers even for attributed patients. Roughly a fifth of your county's Medicare population sits in Advantage plans.
PCM is not named in the restriction, but it sits in the same care-management family and may be treated the same way. We have excluded it rather than assume in our own favour.
A Care Transformation Organization hires and manages an interdisciplinary care management team on a practice's behalf, and takes a share of that practice's program payments for doing it. CoachCare supplies the operational equivalent for a chronic-care program — and is paid from the reimbursement it generates instead.
Priced at your own Maryland locality (Novitas JL, carrier 12302, locality 01) — not national averages.
| Program | Codes | What it pays for | MD rate |
|---|---|---|---|
| RPM — device monitoring | 99453 · 99454 · 99457 · 99458 | Setup, monthly device supply with 16+ days of data, and 20-minute management increments | $23.41 · $55.96 · $54.61 · $43.50 |
| RPM — short window 2026 | 99445 · 99470 | 2–15 days of device data; 10–19 minutes of management | $55.96 · $27.47 |
| CCM — chronic care management | 99490 · 99439 | Not billable on your attributed panel — the care management fee replaces it. Retained for non-attributed, Advantage and commercial patients | excluded |
| PCM — principal care management | 99426 · 99427 | Not named in the restriction but in the same family — excluded pending confirmation | excluded |
| TCM — transitional care | 99495 · 99496 | Post-discharge management within 7 or 14 days — the readmission window | not modeled |
For a practice carrying two-sided risk, avoided utilization is the first layer and fee margin is what pays for getting it. We have ordered these accordingly.
Every reading in every program routes through one escalation engine with defined thresholds, defined actions, and defined documentation. Your physicians see signal, not noise — and the billed time is defensible on audit.
Chest pain, new shortness of breath, stroke signs, syncope, worst-ever headache or sudden swelling reported during outreach triggers a 911 call while the patient is still on the line. If the patient refuses, they are routed to the clinic; if they cannot be safely directed, CoachCare activates 911. This policy supersedes any practice-specific escalation preference.
Actionable but non-emergent findings route to a specific, pre-agreed member of your care team — not a general inbox. You decide who, and for which thresholds, during configuration.
Readings worked to resolution are documented as an FYI in the chart with no interruption to your day. This is the route that keeps the other two credible.
RPM, CCM and PCM all route through the same logic. The rules do not change by program — only the codes do.
Any emergency-department visit or hospitalization in the preceding 60 days triggers a fixed three-touch sequence. This is the mechanism behind the avoided-admission figures, and it maps directly onto readmission performance in both your shared-savings program and the state model.
Medication reconciliation against the discharge summary, red-flag symptom review, and confirmation that follow-up is actually booked.
Adherence and tolerance check, device data reviewed against the pre-admission baseline, escalation if the trajectory is wrong.
Stability confirmation and hand-back into the routine monitoring cadence — the window where most preventable readmissions occur.
Your staff do not learn a second system. Enrollment flags, vitals, documentation and claims live inside athenaOne — and CoachCare is the only care-management partner that generates athenahealth claims automatically.
Qualified Medicare patients are enrolled by CoachCare's team, prompted by enrollment flags in your existing workflow. Enrollment status is visible in real time, and patients begin receiving services in under five days.
The billing engine creates claims automatically, eliminating the manual per-patient, per-month step. For a practice with no CTO supplying billing support, this is the operational difference that decides whether a program survives its first year.
Clinical escalations arrive as tasks in the system your physicians already work in, routed to the person you designate — not as email, not as a separate portal login.
athenahealth integration setup and per-patient integration fees are included in the Value Analysis at catalog rates; exact figures are confirmed in contracting.
The charts below model remote monitoring on the recommended 2,500-patient high-risk cohort at Maryland locality rates, with one CoachCare-funded on-site enrollment specialist. Illustrative and modeled — to be validated against your athenaOne chart counts.
A single-program build, so unique patients and active enrollments are the same number — no dedup is needed: 744 patients at month 24, reached by month 6 and flat from there. Illustrative, modeled — verify against practice data.
Net reimbursement, CoachCare fees, and the margin the practice retains. Month 1 carries the one-time implementation and integration setup.
24-month net reimbursement by code family, recommended cohort. Net is apportioned across codes in proportion to gross billed.
| Program | Net reimb. | Program cost | Practice margin |
|---|---|---|---|
| Remote monitoring — devices, data & management | $1,646,725 | $903,939 | $742,786 |
| Implementation, athenaOne integration & outreach | — | $40,753 | −$40,753 |
| 24-month total | $1,646,725 | $944,692 | $702,033 |
The cohort is what we recommend. Panel-wide is what the opportunity looks like if you eventually screen all 5,000 — shown so the ceiling is visible, not because it is the plan. Note how little it adds: screening the other half of the panel yields only 131 more patients, because the cohort already contains most of the monitoring-appropriate population. Both are priced identically and both are monitoring-only. Illustrative, modeled — verify against practice data.
| Configuration | Net reimb. | Practice margin | Margin % | Patients at M24 | Avoided admissions |
|---|---|---|---|---|---|
| High-risk cohort — ~2,500 screened Recommended | $1,646,725 | $702,033 | 42.6% | 744 | ~106–213 |
| Panel-wide — all ~5,000 screened Ceiling | $1,906,292 | $813,811 | 42.7% | 875 | ~123–246 |
$1.60M to $3.19M of avoided cost, depending on the cohort's true baseline admission rate.
Over 24 months, after all CoachCare fees, at a 42.6% practice margin — positive from month one.
About 5.5 FTE-years of monitoring, documentation and outreach labor absorbed by CoachCare.
Continuous objective data on the patients driving your utilization.
With twenty-two referring clinicians and a full-time enrollment specialist, your outreach capacity is roughly 141 referrals a month before the specialist adds anything. Against a cohort ceiling of 744 patients, that saturates in six months and then holds flat through month 24 — the census at month 12 and at month 24 is the same number.
That flat line is the useful finding: your constraint is the cohort definition and the consent rate, not outreach capacity. Adding clinicians or a second enrollment specialist would not move this number at all. The two levers that do are widening the cohort — deliberately, where the clinical and benchmark case supports it — and consent rate, which is why it is a slider below.
The sliders drive the same enrollment engine and the same per-patient economics as the Value Analysis. At the default positions it reproduces the recommended cohort exactly. Note how little the clinician and specialist sliders move the result — and how much cohort size and consent rate do.
Patients in active monitoring. Cohort eligibility is held at the high-risk profile throughout.
Avoided admissions here use the model's conservative 20% baseline annual admission rate. For a true heart-failure and advanced-CKD cohort the realistic figure is roughly double — see the sensitivity table above.
No capital, no hiring, no second system. The work in the first month is clinical agreement and configuration, not construction.
Pull the real cohort from athenaOne — heart failure, CKD 3b–4, COPD, and uncontrolled hypertension and type 2 diabetes — which replaces the estimate behind every number on this page. In parallel, confirm against the current payment specifications exactly where the care-management-fee offset begins and ends, so the configuration is right on day one rather than corrected later.
Agree monitoring thresholds by condition, name the clinician each non-critical escalation routes to, and confirm the emergent protocol. This is the session that makes the governance model yours — and the one your care manager should lead.
Bi-directional integration configured and tested: enrollment flags, discrete vitals write-back, escalation tasks, and automated claim generation.
The CoachCare-funded specialist begins in your office, referral capture starts at the visit, and the first cohort is monitored and billing within days of consent.
Remote care programs fail on operations, not on clinical intent. Scale is the only real evidence that the operations work.
Across more than 400 managed conditions.
Providers committed to remote care excellence.
Successful in-market programs launched.
Care-plan coding and billing at production scale.
Plus more than 4 million care actions enabled.
From consent to monitored and billing, inside athenaOne.
Every number on this page is either a cited public fact or a modeled projection. This section says which is which, and what we could not establish.
The recommended configuration screens a 2,500-patient high-risk cohort: patients carrying heart failure, CKD stage 3b–4, COPD, or poorly controlled hypertension or type 2 diabetes. That is roughly half of an estimated 5,000-patient Medicare panel — consistent with chronic-condition prevalence in an adult internal medicine Medicare population, and an estimate, not a count. Of those 2,500, the model enrolls 744: 85% judged device-appropriate, of whom 35% consent.
The panel figure is derived from provider count and typical adult internal medicine Medicare panel density, then sanity-checked against the local market: Anne Arundel County carries 111,257 Medicare-eligible residents, so 5,000 is about 4.5% of the county's Medicare population — credible for one established practice of this size. The plausible range is roughly 4,000 to 6,300.
Because the model is ceiling-limited, both figures move the 24-month result close to proportionally. Pulling the real counts in week 1 is the single highest-value action in the roadmap, and every figure here is illustrative and modeled until that happens.
Every code is priced at the 2026 Medicare Physician Fee Schedule non-facility rate for your locality — Novitas Solutions JL, carrier 12302, locality 01 — resolved from ZIP 21401, not from national averages.
Reimbursement is modeled net, not gross: a 2.5% denial rate, 20% patient coinsurance with a 25% bad-debt assumption on that coinsurance, and realistic code-completion rates rather than assuming every eligible code is billed every month. Add-on codes are modeled at partial hit rates. The result is a net-collected figure, not billed charges.
Three pathways run in parallel: referral capture at the visit (22 referring clinicians, eight referrals each per month, 80% patient accept rate), one on-site enrollment specialist at 80 enrollments per month, and telephonic outreach at a deliberately conservative conversion. All three ramp over the first four months rather than starting at full productivity.
The consent rate is 35%. Device-appropriate eligibility is 85% within the recommended cohort — a deliberate condition-defined override of the 65% figure a panel-wide primary-care build would use, appropriate because the cohort is selected by diagnosis. The ceiling is cohort × eligibility × consent — 2,500 × 85% × 35% = 744 — which is why the census plateaus at month 6 and the month-12 and month-24 counts are identical. Enrolled patients attrit at 1.5% per month. Outreach capacity is roughly 141 referrals a month before the specialist is counted, so capacity is not the binding constraint at any point.
For the panel-wide ceiling, eligibility is derived from the bottom up rather than taken from a default, so the two scenarios stay consistent: the cohort's 2,125 device-appropriate patients plus roughly 15% of the 2,500 patients outside the cohort gives about 2,500 device-appropriate patients across the whole panel — 50% — and a ceiling of 875. That is why screening the other half of the panel adds only 131 patients: the cohort already contains 85% of the monitoring-appropriate population.
Avoided hospitalizations are modeled as RPM patient-months × a baseline annual admission rate × a 40% reduction factor, valued at $15,000 per admission. The model's default baseline is a 20% annual admission rate — a general Medicare population figure.
A genuine heart-failure and advanced-CKD cohort runs materially above that, which means the default understates the clinical value of targeting. The sensitivity range shown (106 to 266 avoided admissions across 20% to 50% baselines) exists so the assumption is visible rather than buried. For local context, CMS reports an Anne Arundel Medicare acute readmission rate of 19.9% and 483.5 emergency-department visits per 1,000 beneficiaries (CY2024).
Avoided-admission value is presented as clinical and system value. It is not added to the practice margin anywhere on this page.
Not asserted anywhere on this page: any quality rating; any existing remote monitoring program; any statement about the practice's finances, intentions, or performance under its accountable care organization; or any resolution of the care-management-fee question beyond the published language quoted. Where we do not know something, we have left it out rather than estimated it.