Beyond Wearables, Part 2: The Subtle Signals That Save Lives

May 21, 2026

Jennifer Leatherbarrow
RN, BSN, RAC-CT, RAC-CTA, IPCO, QCP, CIC

In Part 1, we sat with an uncomfortable truth: the residents most at risk in skilled nursing are the ones our traditional tools serve worst. Call bells, wearables, and bed alarms ask the resident to participate in their own rescue, and the residents who can’t or won’t are exactly the ones whose lives are on the line.

This post is about what happens when you remove that requirement. When the room itself is watching, what does it actually catch?

The honest answer: certain early signals exist below the resolution of intermittent rounding. They aren’t catchable at the bedside in real time — and that’s about measurement, not effort.

The crisis you see is rarely the first signal

This is the pattern we keep seeing on post-event review: the chart from the 48 hours before a hospitalization shows nothing acute. The decline was already happening — it just hadn’t crossed a threshold visible at three vitals checks a day. That isn’t a clinical judgment problem. It’s a measurement problem.

Vitals rounded once a shift give us three data points across 24 hours, and each of those data points has limits. Respiratory rate, the strongest early predictor of deterioration in older adults, requires a full minute of focused counting to be reliable — a standard the realities of a busy floor rarely allow for. Heart rate is typically taken at the wrist for thirty seconds, doubled to give us a per-minute value — accurate enough at rest, but not granular enough to show the small shifts that often precede a cardiac or septic event. Temperature is the most consistent measure we have, and even temperature trails the underlying change by hours.

Now imagine a resident whose respiratory rate has been climbing one breath per minute, every shift, for three days. No fever yet. No shortness of breath the resident can articulate. No oxygen drop on a spot check. To a CNA at the bedside, nothing looks wrong. To a continuous monitor, the trend line is screaming.

That’s the signal we’re missing. And that’s the signal contactless radar monitoring picks up while the resident is sleeping or in bed.

What the published evidence actually shows

This is no longer a theoretical pitch. Tapestry Remote Patient Monitoring has been studied at scale in real skilled nursing settings, and the numbers are doing what good clinical evidence is supposed to do — they’re forcing a conversation.

In a study of 612 residents across five Tapestry facilities, the contactless radar platform predicted 77% of hospitalizations an average of 5.6 days before they occurred (Tapestry Health, Radar Vitals). For respiratory and cardiac hospitalizations specifically — the two most common, and most preventable, drivers of SNF transfers — the system caught 89%. Five and a half days of warning is not a faster response. It’s a different category of care.

The other number worth pausing on: 92% of alerts were filtered out before reaching facility staff (Tapestry Health, Radar Vitals). Without that filtering, continuous monitoring becomes another source of alarm fatigue. With it, the alerts that do reach the unit are the ones that actually need a clinician’s eye.

What a “save” actually looks like on the floor

What does that look like in practice? A few patterns come up repeatedly across facilities running contactless platforms.

The respiratory creep. A resident with a history of CHF starts trending three breaths per minute above their baseline overnight. No alarm. No call bell. Their daytime exam looks unremarkable. The platform flags the trend on day two. The nurse listens to lungs more carefully, calls the provider, gets a diuretic adjustment, and avoids the ED transfer that would have happened on day five. The resident never knew anything was wrong, and that’s the whole point.

The silent infection. An elderly resident can’t mount a fever the way a younger patient does. What they can do is run a slightly elevated heart rate for 36 hours before anything else changes. A wrist-check at 7 a.m. shift change won’t catch a 6 BPM elevation. A continuous radar trend will. By the time the urine culture comes back, treatment is already underway.

The fall that didn’t happen. Repeated bed-exit attempts at 11 p.m., 12:15 a.m., and 1:30 a.m. don’t register as a “fall risk” on any incident report — until the fourth attempt is the one that ends with the resident on the floor. A platform that sees the first three attempts as a pattern, and flags the room before the fall, changes the math entirely. That’s a fall that was prevented, and prevented falls don’t make it into your QAPI data unless you capture them deliberately. (We’ll come back to this in Part 3.)

The deterioration nobody named. This is the hardest one to talk about, because it’s the one where the resident dies. Sleep architecture quietly disintegrates over a week. Nighttime restlessness goes up. Respiratory variability climbs. Standard documentation shows no acute changes. The continuous trend shows the slope. Having that slope visible is what gives the clinical team the chance to intervene early — and that earlier intervention is often the difference between a treatable change and a hospitalization.

These aren’t dramatic Hollywood saves. They’re quiet saves. They look like a resident who didn’t go to the hospital, didn’t end up on a vent, didn’t lose function, didn’t die. The wins are invisible by design, which is part of why this technology has been undersold for so long.

What the bedside can’t see, even when it’s looking

This is worth saying plainly. Contactless monitoring isn’t catching what staff fails to see. It’s catching what no clinician — no matter how skilled — can see when their attention is required across an entire unit at once.

The bedside catches a lot — clinical instinct picks up “something’s off” the moment a nurse walks into the room. The limit isn’t the instinct; it’s the math. Each resident gets a fraction of a clinician’s attention across a shift, simply because there are more residents than clinicians at any given moment. Continuous monitoring covers the hours that bedside attention was never going to reach.

A continuous platform fills that gap. It doesn’t replace the nurse’s instinct — it gives that instinct something concrete to react to. “Mrs. Lee’s overnight respiratory rate has been creeping up for two nights” is a very different conversation with the medical director than “Mrs. Lee seems a little off today.” One gets a treatment plan. The other gets watchful waiting until something obvious happens.

This is also why the alert filtering matters so much. A platform that does the math, separates noise from signal, and only escalates the changes that actually warrant action is the difference between a tool your staff trusts and a tool they tune out. Tapestry Remote Patient Monitoring’s filtering — escalating roughly 1 in 12 alerts to facility staff — is what makes the model sustainable on a real unit with real staffing.

The clinical and ethical case is not subtle

If you can see a resident’s hospitalization coming five days out, and you don’t have the tools to see it, you’re going to keep losing those residents on the timeline the disease writes for them. If you can see it, and you act, the timeline changes.

This isn’t an efficiency argument or a financial one. It’s a clinical one. Our residents deserve attention to the subtle changes that have historically only been visible to us once they became obvious — and by the time something is obvious, the window for early intervention has usually closed.

What’s coming in Part 3

In the final post, we’ll talk about what happens after the saves start adding up. Continuous monitoring data isn’t just a real-time alert engine — it’s a record of how your facility actually performs at 3 a.m., and that record is gold for QAPI, root-cause analysis, and the conversations you have with hospital partners and payers.

The story isn’t “we bought a sensor.” The story is “we can now show, with data, that our high-acuity residents are safer here than they would be elsewhere.” That’s a different facility on the referral list. That’s the closer.