The KinematiX™ Platform: What It Is, and How It Measures You at Home

Why I built it

Twenty years of practicing physical therapy taught me the same lesson over and over: the biggest gap in most PT isn't the exercises, it's the measurement. A clinician watches you move once every week or two, forms an impression, and writes "improved" or "not improved" from memory. That's not a knock on any individual therapist — it's just what's possible without objective data. KinematiX™ is my answer to that gap: a platform that captures real movement data at every visit, so your plan of care is built on measurements, not impressions.

What KinematiX™ actually is

KinematiX™ is the technology platform behind Virtual PT Hawaii. It's not a separate app you download — it's the system running underneath your telehealth visits, your home exercise program, and your documentation. The clinical work is still done by me, a licensed Doctor of Physical Therapy. KinematiX™ is the infrastructure that makes a virtual visit as objective as an in-person one, and in the specific case of movement measurement, often more consistent, because a camera doesn't have an off day.

The part of KinematiX™ patients actually interact with is called Atlas — the motion capture engine. That's what's doing the measuring.

How Atlas measures you through a home webcam

During a movement screen, Atlas uses your device's camera to track the position of your joints in real time — the same underlying computer-vision approach used in sports biomechanics labs, adapted to run on a laptop or phone camera instead of a $150,000 marker-based system. As you move — a squat, a shoulder raise, a short walk — Atlas tracks joint angles frame by frame and turns that into range-of-motion numbers, gait speed, and movement-pattern flags I can review.

I want to be direct about the limits here, because I don't think a clinician should oversell a tool: a single 2D camera is genuinely good at some measurements (hip and knee flexion/extension, walking speed) and deliberately doesn't report others (rotational movement, fine ankle mechanics) because the accuracy isn't there yet. Atlas is built to say "I don't have a confident read on this" rather than guess. I wrote up the full accuracy breakdown, with real numbers and citations, on our accuracy page — worth a read if you want the unfiltered version.

Measured against the same academic benchmarks as an in-person exam

I don't want "objective data" to be a vague marketing phrase, so here's the actual comparison. Published inter-rater error for goniometry — the angle-measuring tool clinicians use in person — runs about 1.9–4.6° (Gallego-Izquierdo 2020 and related literature), and inter-rater agreement between two clinicians measuring the same joint can fall as low as 0.16–0.35 ICC for some joints. In other words: two skilled clinicians measuring the same knee, by hand, in person, often don't agree closely with each other.

For sagittal-plane hip and knee range of motion — the measurements Atlas is strongest at — it lands in roughly that same 2–5° error band. Not better than a skilled clinician with a goniometer, but in the same range, with zero drift between visits because there's no second rater to disagree with the first. For gait speed, markerless capture reaches ICC 0.97–0.98 against instrumented reference systems — gait speed is sometimes called the "sixth vital sign" in the research literature, with a known minimal clinically important difference of about 0.1 m/s, and that's precise enough to catch a change that size.

I'm not going to tell you Atlas beats an experienced clinician's exam across the board — it doesn't, and I wouldn't trust a platform that claimed it did. What I can tell you, with citations, is that for the specific measurements it reports, Atlas performs in the same real-world range published for in-person tools. The full breakdown — including what Atlas deliberately does not report because the evidence isn't strong enough yet — is on the accuracy page.

Why we don't look at one joint in isolation

Pain in one place is often driven by something happening somewhere else — a stiff hip changing how the knee loads, a weak core changing how the shoulder moves overhead. Physical therapists call this the kinematic chain: the idea that joints don't move independently, they move as a linked system.

Because Atlas captures multiple joints in the same pass instead of one angle at a time, I'm not looking at your knee in a vacuum — I'm looking at your knee, hip, and trunk together during the same squat or the same gait cycle. That doesn't replace clinical judgment; it's still me connecting the dots between what the data shows and what you're telling me about your pain. But it means the data I'm reasoning from covers the whole chain, not just the joint you happen to be complaining about that week.

From your initial eval through every re-check

Here's how it actually plays out across your care. At your initial evaluation, Atlas captures a real baseline — not "your knee feels stiff," but an actual measured range of motion, gait speed, or movement pattern, timestamped and stored. Every visit after that runs the same protocol, so the comparison is apples-to-apples: the same measurement, the same math, on the same joint, months apart.

That's what lets me show you — not just tell you — "your hip flexion moved from 82° to 94° over six weeks," with the two measurements sitting side by side. That progression, plus what you're telling me and what I see on the call, is what actually drives whether we progress your exercises, hold steady, or change direction. It's also what ends up in your documentation and, where applicable, supports the billing codes tied to your care — a record of real change, not a guess dressed up as one.

Try it yourself

You don't have to take my word for any of this. There's a free movement screen on our demo page — no email required to see a sample result. Next up: remote therapeutic monitoring — what it is, who's eligible, and how it fits into ongoing care between visits.