JointCoach uses the phone camera to capture objective movement data during every session — no wearable, no extra staff. She reviews a 3-minute dashboard and sees a stalling trend at week 3 instead of hearing about it at week 6. Objective accountability makes adherence visible — so a clinician can act before recovery stalls, not after — and Medicare reimburses the monitoring, billed under her own NPI.
The joint-care journey — you're on the tracking step
Collagen, tendon, bone and muscle each remodel on their own trajectory, and those trajectories run in months — sometimes longer. An episode of care runs in days to weeks. The two clocks have never been aligned, and the gap between them goes unattended by design.
Symptoms improve before tissue matures. So care tends to end when the patient feels better — which arrives earlier than when they are better. The remodeling finishes unobserved, load returns before capacity does, and a share of those patients come back as a second episode that starts from scratch.
That gap is what JointCoach is built to sit in. Between-visit monitoring is the one instrument that runs on the tissue’s clock rather than the billing clock — and RTM is how the time spent watching it gets paid for, under your own NPI.
That's it. No extra apps, no wearables, no subscriptions to manage.
Hold your phone in front of you. JointCoach measures your joint angle and counts your reps. No wearable needed. Video never leaves your device.
Your surgeon and PT see your adherence, range of motion, and pain scores — without needing to call you. They can adjust your program between visits.
Medicare and most commercial plans reimburse remote therapeutic monitoring. Your surgeon's practice bills it like any covered service — your plan's standard cost-sharing applies, and supplemental plans often cover it in full.
The same patient. The same PT program. Different outcomes — because the surgeon could see what was happening at week 3.
6 weeks of PT. Compliance unknown — the patient self-reported at each visit. At week 6 the patient returns with stiffness and limited range of motion. The surgeon assumes the condition, not the exercises. A second PT episode starts from scratch.
Weekly movement quality scores visible in the dashboard. At week 3 the surgeon sees adherence has dropped to 30% and ROM has plateaued. She sends a message through the portal before the next visit. The patient gets back on track — ROM returns to trajectory by week 7, instead of a second PT episode starting from scratch.
Three steps from your surgeon's referral to your progress dashboard.
After your procedure, your care team sends your exercise program to JointCoach. You get a link or a QR code at the clinic.
Use your phone camera to track each session at home. JointCoach measures your range of motion and counts your reps in real time.
Your surgeon and PT see your progress between visits. They can update your program without waiting for your next appointment.
Remote therapeutic monitoring (RTM) is a CMS-recognized care category. Your care team bills under their own NPI — your plan's normal cost-sharing applies, and many Medigap and supplemental plans cover it fully.
Cold-therapy units, braces, TENS units, and home-safety gear qualify as-is. A recumbent bike, pool membership, or massage can qualify too — with a Letter of Medical Necessity tied to your diagnosis. Paying pre-tax is roughly a 30% discount on your whole recovery list.
Remote therapeutic monitoring (RTM) was built for PT — but until 2026, you needed 16 or more days of data per month to bill the device supply code. CMS changed that with the 2026 Physician Fee Schedule.
You evaluate the patient and send them home with an exercise program — a printed handout, an email, or a demonstration you hope they remember.
The same program delivered digitally, so the patient can follow it and the data comes back. Your team reviews it monthly, checks the movement, and makes contact.
Not for the exercises — for the monitoring and the clinical time spent reviewing it. That is the whole billable idea.
Most explanations stop at step 3, and that’s where the two hard parts live. The billable portion requires documented clinical review time and at least one live interaction every month — and under the CY2027 proposal that review must be performed by staff employed by your practice, not a vendor’s. So this is not a billing code stapled onto work you already do. It is a service line with real clinician time in it. Worth running, worth being honest about. The numbers below are what it pays, and what CMS is proposing to change.
You can't hire your way out of the coming MSK wave. So stretch every licensed hour you already have.
US PT programs graduate only enough to meet 67% of current demand — and 51% in the West. Even if every developing program opens, new graduates close a fraction of a 34,000-and-growing shortage (Childs & Benz, Physical Therapy J. 2022, on BLS data). Meanwhile 127M Americans carry a musculoskeletal condition and the population is aging. RTM is how one therapist covers more patients between visits — asynchronous monitoring, reviewed monthly, billed under general supervision. It isn't extra revenue. It's the delivery model for a workforce that can't scale fast enough to meet demand.
Every camera-tracked session at home generates structured data — joint angle, rep count, session duration, pain score — that qualifies for RTM billing under 2026 CPT codes. SpryPT white-labels JointCoach for clinics that want a turn-key setup.
New for 2026: CPT 98985 — a short-period device supply code. Patients with as few as 2 qualifying days of monitoring per month now generate a billable device code. A patient who completes just two camera-tracked sessions of their prescribed program per week now clears the threshold. The 16-day threshold still applies to the longer device code (98977).
Prescribe through SurgeonValue or your own EMR. Patients exercise at home. JointCoach captures the data. Your team reviews monthly. CMS pays under general supervision.
The RTM and RPM line items in that report — 98980, 98985, 99457, 99458 — are this. SurgeonValue identifies what’s sitting uncaptured in your panel. JointCoach is the infrastructure that captures it.
One piece of honesty about that number: it is not revenue you forgot to bill. Most practices bill no RTM at all — in WebPT’s 2025 survey only about 7% of small and mid-size practices offered it. It is revenue that requires actually running the monitoring, with your own employed clinical staff — which is what the CY2027 proposed rule would require, and what this is built to do.
| CPT | Description | CMS 2026 |
|---|---|---|
| 98975 | RTM device setup & patient education (once per episode) | $21.71 |
| 98985New 2026 | Device supply, MSK/respiratory — 2 to 15 days of data | $51.44 |
| 98977 | Device supply, MSK/respiratory — 16 to 30 days of data | $51.44 |
| 98979New 2026 | Treatment management, 10 to 19 min of qualified review | $26.39 |
| 98980 | Treatment management, first 20 min | $54.11 |
| 98981 | Treatment management, each additional 20 min | $41.04 |
2026 CMS Physician Fee Schedule, national-average reimbursement. Rates vary by locality. RTM and RPM cannot be billed for the same patient in the same month. A synchronous interaction (phone, video, or in-person) is required each calendar month. Reimbursement confirmed for PT and physician billing under general supervision.
Split your caseload the way CMS splits the codes: patients who log 16+ tracked days a month bill the standard stack; patients at 2–15 days bill the new short-duration stack.
12 × standard stack (98977 + 98980 = $105.55) + 8 × short-duration stack (98985 + 98979 = $77.83) + 5 × setup (98975 = $21.71)
Get your case-mix result as a one-page summary — plus the CY2026 code table and the setup checklist a clinic actually needs to start billing.
One email with your numbers. No list, no drip unless you ask for it.
Rates from the CPT table above (2026 CMS national average; varies by locality). Management add-on 98981 not included — treat these numbers as the floor. These are 2026 values, and three CY2027 developments should change how you model them (see the calendar). One: the proposed rule would revalue the device codes downward. Two: it would pay for RTM only when the monitoring is performed by clinical staff employed by the billing practice — not contractors — so a vendor-staffed program does not produce this revenue in 2027. Three: CMS is separately soliciting comment on collapsing all 17 RPM/RTM codes (98975–98981, 99453/99454, 99457/99458 and the rest) into four new G-codes (GRPM1–2, GRTM1–2), following an OIG recommendation to tighten remote-monitoring oversight. That last one is a comment request, not proposed regulatory text — but it means the code set underneath this calculator is under active reconsideration. Comments close September 14, 2026. Model these as current, not permanent. Context: in WebPT’s 2025 State of Rehab Therapy survey, only 6.6% of 2–5-provider practices and 7% of 6–16-provider practices reported offering RTM at all — for nine out of ten clinics this line item is currently $0.
Three are final. Two are still proposals — labeled that way below. None are pilots a clinician can wait out. The clinicians who do well under them will be the ones who started the record early, because an evidence base, unlike a software license, cannot be purchased in December.
Episode-based payment became mandatory for 740 hospitals. The procedure is now priced together with everything around it — including the recovery that used to disappear from view at discharge. The clinician who can show what happened after the OR holds the data the episode is scored on.
The surgeon side of TEAM →The Ambulatory Specialty Model puts the individual clinician managing low back pain at up to plus-or-minus 9–12% of Part B revenue. No opt-out. Not the hospital’s score — hers. And the score is built from data she either has or doesn’t.
How ASM scores the clinician →Medicare has proposed paying clinical software directly, with new codes that would pay tech-enabled platforms for chronic MSK care anchored on baseline PROMs. Still a proposal — but it shows exactly where the payment rails are being laid: monitored, outcome-anchored records.
Where the payment rails lead →CMS has proposed revaluing the RTM device codes downward and paying them only when the monitoring is the practice’s own employed staff, for established patients. Still a proposal (comments due Sept 14, 2026) — and it reshapes the outsourced-monitoring market, not this: JointCoach already runs under the treating clinician’s own NPI and their own team. Model the 2026 rates as current, not permanent.
Read the CMS proposal →In the outpatient and ASC setting — where joints are migrating — THA/TKA patient-reported outcomes become fully mandatory in 2031, affecting FY2032 payment. (The inpatient version is already tied to FY2028.) Miss the pre- and post-op windows, or fall under the 50% response rate — 45% in an ASC — and the OQR penalty is a 2.0-point cut to the payment update. Outcome capture stops being a differentiator and becomes the floor.
Why monitoring is the half with evidence →JointCoach is one way to build that record early: monitored recovery, reviewed and signed by the treating clinician, documented under her own NPI. The point isn’t the tool — it’s having the outcome data before the calendar starts asking for it.
RTM pays a fixed amount per patient per month — the standard stack works out to $105.55. The tissue’s clock, though, runs for months. So the question isn’t whether continuous monitoring is good medicine. It’s whether the cost of doing it stays under a number that never moves.
Send every session to a server and your cost rises with exactly the thing you’re being paid a flat rate for: more patients, more reps, more months. The better it works, the worse it pays. That is a business that breaks precisely when it succeeds.
On the phone, the compute is already bought and paid for — by the patient, years ago. Pose estimation runs on hardware sitting in their hand, and the marginal cost of the ten-thousandth rep is the same as the first: nothing. Fixed payment, fixed cost. That is the only shape in which a nine-month monitoring window is a business rather than a subsidy.
Those are the same sentence, and that is not a coincidence — it’s the whole architecture. The limitation everyone cites about on-device AI, that the models are small and narrow, simply isn’t a limitation for this job. Monitoring never needed a frontier model. It needed the same measurement, taken reliably, for months.
Apple is pushing this hardest — every recent iPhone ships a neural engine, and Apple’s own on-device models run free on it. That trend is a tailwind, not a dependency: JointCoach’s measurement runs on the phone today regardless of whose model wins. And the line stays where it is — on-device means on the device. Anything that ships a patient’s data off the phone for processing, including a vendor’s private cloud, brings back both the cost curve and every question about who is holding protected health information and under what agreement. Here is exactly what does and doesn’t leave →
JointCoach runs Google MediaPipe pose estimation locally on the patient's phone. The model identifies joint landmarks frame by frame and calculates angles in real time. When the session ends, only structured data is transmitted — rep counts, peak range of motion, session duration, and reported pain level. The underlying video is discarded on-device immediately after processing.
The gap this closes is one the profession names itself.
In a survey of 514 U.S. physical therapists, 97% used standardized outcome measures and 89% said performance-based tests drive their clinical decisions. But asked whether those measures are “administered in a standardized way throughout my profession,” only 48% agreed — the lowest-rated item on the entire survey. The tool is trusted; the consistency is not, because a goniometer read by hand varies from clinic to clinic and clinician to clinician. A camera angle computed the same way on every phone does not. That is the point of measuring range of motion this way: not a new number, but the number they already trust, made comparable across every session and every site. Kristal A, Gaunaurd IA, Morgan SJ, et al. Use of standardized outcome measures among physical therapists in the United States. PLoS One 2025;20(8):e0330528 (open access).
Peak joint angle per rep, session max, and trend over time. Knee flexion, hip abduction, shoulder elevation — any joint the PT designates as the primary tracking target.
Rep count, set completion, session duration, days active in the month. The numbers that determine whether 98977 (16+ days) or 98985 (2-15 days) applies.
0-10 numeric pain rating and perceived difficulty captured after each session. Logged to the RTM record for the treating PT's monthly review — satisfies the treatment management review requirement.
Video-based motion analysis is not a novelty waiting on adoption. Roughly half of physical therapists already report using it in practice, and it carries real methodological standards — anatomical landmarks, sagittal and frontal capture, high frame rates, joint angles and cadence measured frame by frame (Tripodi et al., Int J Osteopath Med 2024;51:100698). JointCoach doesn’t introduce a new instrument. It changes who holds it and when — the same measurement, in the patient’s hand, on the days between visits, which is where the recovery is actually won or lost.
And the limits, because they matter: 2D video is less accurate than a 3D motion-capture lab. Kinematics are one piece of a clinical picture, never the whole of it, and the evidence supporting any specific change is stronger for some problems than others. So JointCoach reports what it measures — angle, reps, cadence, adherence, pain — and stops there. The instrument takes the measurement. You make the call. That division isn’t a limitation we’re admitting; it’s the entire design.
On a checklist, non-adherence and fear-avoidance look identical — the exercise just doesn’t get done. In the movement data they look different. A patient who reaches range on most exercises but consistently skips or guards one — smaller range, slower, done far less often — is often avoiding it because they expect it to hurt or harm. That protective loop drives a large share of post-op disability (Alaiti et al., Braz J Phys Ther 2025), and catching it early is where recovery is won or lost. JointCoach surfaces the pattern for your monthly review; you decide what it means — a fear to work through with graded activity, or a real problem to evaluate. It is a signal to explore, never a diagnosis, and never a substitute for ruling out pathology.
Your patient is getting the other half of this in plain language — why hurt isn’t the same as harm, why the instinct to protect the joint becomes the trap, and why steady movement paced by your plan rather than by how much it stings is what breaks the loop. That is the lesson; this is the instrument that tells you whether it took. See what your patient reads →
They are related, but they do different jobs. RTM is the part Medicare reimburses — and the part JointCoach automates.
This year WHOOP wired medical records into its AI. Anthropic shipped Claude for Healthcare, with one-tap connectors for labs, records, and Apple Health. Your post-op patients are already asking an app whether their knee is “on track” — and getting an answer no clinician reviewed, no one is accountable for, and no plan will pay for.
JointCoach is that same objective, made accountable. The movement data runs past the licensed clinician who prescribed the plan — she sees the trend, she is on record for the response, and the monitoring is reimbursed under her own NPI. Same convenience for the patient. The one thing a consumer coach structurally cannot add: a clinician who is accountable for the read.
Purpose-built MSK wearables — muscle-activation patches, motion sensors — are starting to arrive, and that is good: more objective signal. But a sensor doesn’t get reimbursed. RTM does — and RTM is the clinical layer, not the device: a treating clinician who reviews the trend, objective evidence, and the CPT machinery that bills it. JointCoach is that layer. Today it needs no hardware at all — the phone camera captures objective range-of-motion and adherence. As better sensors arrive, the billing, the evidence, and the accountability don’t change. The same holds for whose hardware it is: as implant and device makers bundle their own recovery apps, each one covers only that maker’s patients — but your panel spans vendors, and one neutral layer covers all of it. The gadget is the door; the reimbursement layer is the building.
The research on unmonitored home exercise programs is clear — and it explains why RTM exists. Educational information only; follow your care team's instructions.
And the evidence is specific about which half works.
A 2026 meta-analysis of 55 randomized trials (5,889 patients) found that digital tools improve self-care monitoring — a moderate, real effect (SMD 0.49). The same review found they barely move behavior or medication adherence (SMD 0.06 across 17 trials — statistically indistinguishable from zero). So JointCoach does not claim to make a patient adhere. It measures whether they did, and puts that in front of the clinician who can act — which is the half the evidence actually supports. Longhini J, et al. Effectiveness of Digital Health Interventions to Improve Self-Care in Patients With Chronic Diseases. J Med Internet Res 2026;28:e88708.
A patient at 80% adherence to an under-dosed program still doesn’t get the outcome. That distinction is the whole difference between a product that reports engagement and one that reports whether the prescribed dose was actually delivered.
Post-op protocols carry their own targets, set by the surgeon and the treating therapist. The principle transfers even where the numbers don’t: a monitoring system should measure against the dose that was prescribed — frequency, volume, and whether the work is still hard enough to matter — rather than against whether the patient opened an app. That is the difference between an engagement metric and a clinical one.
However your clinic wants to operate RTM, JointCoach is the data layer underneath — and your own clinician always owns the billable monthly review. Under the CY2027 Medicare proposal, RTM must be performed by the practice's own employed staff, not an outside vendor, so every model below keeps the review inside your practice. Each one trades administrative lift against cost — never clinical ownership.
Three products, one recovery. Learn what to expect, track every step, and let your surgeon's team see the data.
Patient education. Everything your surgeon wishes you knew before and after joint surgery — free, evidence-based, no login required.
PT-based remote therapeutic monitoring. Home camera tracking with MediaPipe pose estimation. The data layer that generates RTM billing.
The practice OS for orthopedic surgeons. PROMs, encounter capture, billing intelligence, prior auth. Where the surgeon sees the panel.
A patient learns what to expect on JointClass, tracks the prescribed protocol on JointCoach, and the data flows to the surgeon's SurgeonValue dashboard — where the PT or covering clinician bills RTM. Education → monitoring → practice. One loop.
Tap your week and see what matters most right now — what's normal, what to watch, and what JointCoach tracks at this stage. Always follow your care team's specific plan.
Patients: try the camera demo now — it runs on your phone, no signup, and the video never leaves your device. Then ask your surgeon or PT to refer you. Clinicians: add RTM through SurgeonValue or the SpryPT white-label program.