Continuous monitoring stops at the hospital door
In the days after cardiac surgery, a patient is one of the most closely watched people in medicine. Telemetry tracks every heartbeat. A pulse oximeter reads continuously. Nurses round on a schedule, vital signs are charted around the clock, and any drift triggers a response within minutes. The density of observation is the point: most problems are caught because something was always looking.
Discharge reverses that almost completely. The monitors come off, the patient goes home, and the next structured contact is usually a postoperative clinic visit one to several weeks out, with a phone call or two in between. Observation goes from continuous to a handful of scheduled snapshots. The patient feels better, which is reassuring, and is also exactly when a slowly developing problem has room to grow unnoticed.
This is not a criticism of discharge practice. Sending recovering patients home is correct, and keeping them inpatient longer carries its own risks and costs. The point is narrower: there has not been a practical way to keep watching the relevant trend once a patient leaves, so monitoring has effectively been all-or-nothing.
When complications actually present after discharge
The timing of the gap matters because some important complications are timed to land squarely inside it. Delayed pericardial effusion is the clearest example. In a cohort of 1356 cardiac surgery patients followed for six months, 6.2% required invasive treatment for a late pericardial effusion, defined as one presenting after the seventh postoperative day, with 4.1% reaching tamponade and 2.1% reaching pretamponade. The median delay was 11 days after the operation, and individual cases ranged out to 87 days [1].
Those numbers describe a complication that, for many patients, develops after they are already home. The same study noted that the most common subjective symptom was breathlessness, present in well over half of affected patients, and that the patients at highest risk were the younger, generally healthier valve-surgery patients, exactly the group likely to be discharged early and feeling well [1]. A complication that presents quietly, days to weeks out, in a patient who is no longer being observed, is the profile of a problem that gets caught late.
Delayed effusion is not the only post-discharge concern, but it illustrates the pattern: the risk does not end at discharge, while the monitoring largely does. The signal PulSentry follows, the respiratory variation in the pulse oximeter waveform that underlies pulsus paradoxus, is one objective marker that can change as a pericardial effusion develops. For a fuller picture of timing and detection, see our deeper piece on cardiac tamponade after cardiac surgery.
What the current follow-up cadence catches, and misses
The standard post-discharge cadence is built around scheduled assessment: a wound check, a symptom review, sometimes imaging, at a clinic visit set for a fixed date. That cadence is good at confirming expected recovery and at evaluating a patient who already feels unwell and calls in. It is structurally weak at one specific thing: detecting a slow, asymptomatic-until-late trend that crosses a threshold between two scheduled contacts.
A pericardial effusion that accumulates gradually may produce no decisive symptom until it is well advanced, and a patient may not recognize early breathlessness as anything more than the fatigue of recovery. By the time the next appointment arrives, or by the time symptoms are bad enough to prompt a call, the problem can be further along than it would have been if the trend had been visible as it developed. Echocardiography, the definitive test, is a point-in-time study performed when someone is concerned enough to order it; it is excellent at confirmation and not designed for continuous surveillance. We cover that distinction in why echocardiography misses post-surgical tamponade.
The miss, in other words, is rarely a missed test. It is a missed interval. The information that would have flagged the problem earlier simply was not being collected during the window when it mattered.
The cost of a readmission versus the cost of watching
The economics here are not subtle. An unplanned readmission after cardiac surgery means an emergency evaluation, often imaging, frequently a procedure such as pericardiocentesis or a return to the operating room, and another inpatient stay. The cost of one such episode is measured in many thousands of dollars, and the human cost, a frightening deterioration at home and a more invasive intervention, is larger still.
Remote monitoring sits at the other end of the scale. Weeks of following a signal from hardware the patient already owns is a small recurring cost, not a procedural one. The asymmetry is the entire argument: if watching the trend lets a deterioration be caught earlier, when the response can be a clinic visit and a smaller intervention rather than an emergency one, the math favors watching. This is the same logic payers apply when they reimburse remote patient monitoring under defined CPT codes, treating ongoing observation as cheaper than the events it can help prevent. The specifics of those codes are laid out in the 2026 Medicare RPM CPT code guide.
The honest caveat is that the size of any saving depends on the population and on how monitoring is deployed. A program that generates many false alarms or that monitors patients at very low risk will capture less of the available value. The point is directional and well supported: the cost of one avoided readmission dwarfs the cost of the monitoring that might avoid it.
How short-window remote monitoring fits the gap
The gap has a shape, and that shape suggests the remedy. It is bounded in time, roughly the weeks between discharge and the point where the highest-risk window for delayed complications has passed. It is bounded in what needs watching: a small number of objective signals that change when something is going wrong. And it sits with patients who are at home, mobile, and feeling reasonably well, which rules out anything tethered or burdensome.
Short-window remote monitoring matches that shape. It runs for a defined recovery period rather than indefinitely. It uses equipment the patient can manage without training, which for PulSentry means a standard pulse oximeter and no new device to buy or learn. And it is designed to be passive enough that a recovering patient can simply live their day while the trend is followed in the background. A clean, evidence-led overview of the recovery period for patients and families lives in our guide for patients and families.
The relevant evidence for the broader idea is encouraging. In a randomized trial of same-day-discharge surgery supported by one week of remote vital-sign monitoring, an outpatient pathway with remote monitoring was clinically comparable to standard overnight care while sharply reducing hospitalization days, with the explicit motivation being timely detection of postoperative complications after the patient went home [2]. That is a different operation from cardiac surgery, and it should be read as supportive context rather than as a cardiac-surgery result, but it demonstrates the model: a bounded window of home monitoring can substitute for some of what continuous inpatient observation provided.
What a signal-over-time approach adds to a weekly check-in
A weekly check-in is a snapshot. It answers the question "how is the patient right now" on the day it happens. A signal followed over time answers a different and complementary question: "what is the trend between the snapshots." For a complication that announces itself as a gradual change rather than a sudden event, the trend is where the early warning lives.
PulSentry follows one such trend. It computes a pulsus index, the ratio of the respiratory peak to the cardiac peak in the frequency content of the pulse oximeter waveform, and tracks how that ratio moves over hours and days. A single elevated reading is noise; a sustained, rising trend is a signal worth a clinician's attention. The value added is not a replacement for the clinic visit but a continuous thread running between visits, so that a developing problem can surface as a direction of travel rather than as a surprise at the next appointment. If you want the underlying physiology, our explainer on what pulsus paradoxus is walks through the sign and how it appears in the waveform.
Crucially, a signal-over-time approach is a supplement, not a substitute. It does not diagnose, it does not replace assessment, and it does not change the schedule of clinical follow-up. It adds visibility to an interval that has historically had almost none.
Aligning the clinical need with how monitoring is reimbursed
Two things have to be true for post-discharge monitoring to actually happen at scale: it has to help, and it has to be payable. The clinical need described above is the first. Reimbursement is the second, and it is more favorable than it used to be. Medicare and many commercial payers now reimburse remote patient monitoring through a defined set of CPT codes that cover device supply, data collection over a qualifying period, and the clinician time spent reviewing and managing the data.
That structure happens to fit a recovery-window model well. The monitoring period is bounded, the data are objective and continuously collected, and there is a clinician in the loop reviewing the trend, which is precisely what the codes are designed to pay for. The alignment is not accidental: the reimbursement framework was built on the premise that watching patients between visits is worth paying for because it heads off costlier events. A post-surgical recovery window, with a real complication risk that resolves over weeks, is close to the ideal case for that premise. The full mechanics, including the relevant code families and the time and data thresholds, are in the 2026 Medicare RPM CPT code guide.
None of this changes PulSentry's regulatory status. Reimbursement pathways and FDA clearance are separate questions. PulSentry is investigational and is being studied; the economic and coding context here describes the environment a tool like it would operate in, not a claim about its current authorization.
An honest view of the limits
A few caveats keep this grounded. First, no monitoring approach catches everything, and a signal-based one is no exception: it follows a specific physiologic marker, not every possible complication, and it will miss problems that do not move that marker. Second, the economic case is directional rather than guaranteed for any individual program; the savings depend on the population watched, the alarm burden, and how cleanly the workflow routes a concerning trend to a clinician. Third, the strongest randomized evidence for short-window home monitoring comes from adjacent surgical settings, and the cardiac-surgery-specific evidence base is still being built, which is exactly why PulSentry is described as investigational.
What survives those caveats is a clear and modest claim. There is a real, time-bounded gap in observation after cardiac surgery; some of the complications that matter present inside it; the cost of catching them late is high; and following an objective signal over time, reimbursed through existing remote-monitoring codes, is a plausible and economically sensible way to add visibility to that interval. The point of PulSentry is to fill that interval with a continuous, objective thread, not to replace the judgment of the team on either side of it.