Begin with a concession, because the argument is not honest without it. The proposition that continuous physiological monitoring produces better outcomes than periodic monitoring is, for most conditions, unproven. It is not disproven; it is unproven, which is a different and more interesting place to stand. We hold the proposition as a conviction we are working to establish, and we think stating it that way is the only way to be believed when it is established.
The one domain where continuous has clearly won is glucose. Across randomized trials, continuous glucose monitoring beat periodic fingerstick testing on glycemic control and, more consistently, on time-in-range and hypoglycemia — the harms that happen between the readings a periodic method would have missed.[1][2] The margins are real but modest, and even here the benefit depends on someone acting on the stream. Continuous glucose monitoring is a genuine precedent. It is not, by itself, proof that the pattern generalizes to blood pressure, cardiac signal, or the multi-parameter picture of a deteriorating patient.
The counter-evidence deserves equal air. The two largest, most rigorous trials of heart-failure telemonitoring — TELE-HF and BEAT-HF, together enrolling more than three thousand patients — found no reduction in readmission or death.[3][4] They are the honest ceiling on any claim that more-frequent data automatically saves lives. But read closely, they indict something more specific than frequency. In TELE-HF, 14 percent of patients never used the system and only about half were still using it near the end.[3] In BEAT-HF, engagement was decent and the signal chosen — daily weight — simply may not warn of decompensation in time.[4] The trials that worked point to the same lesson from the other side: TIM-HF2, which did reduce mortality, differed from the null trials chiefly in that a clinical team monitored the data around the clock and adjusted therapy promptly.[5] And on the general hospital ward, a 2024 meta-analysis of seven randomized trials found continuous vital-sign monitoring did not significantly beat intermittent checks on mortality, ICU transfer, or length of stay — the cleanest "not yet proven" citation outside glucose.[6]
Put the evidence together and a hypothesis sharpens. The active ingredient in monitoring is not the sampling rate. It is the response loop — whether the signal reaches someone, or something, positioned to act before the event. Continuous data collected and unread is worthless, which is what the null trials mostly measured. Continuous data feeding a workflow that acts is the condition the successful trials shared. No trial we know of has cleanly isolated continuous versus periodic while holding the response loop constant, outside glucose. That is the experiment the field has not run, and it is precisely the one worth running.
Here is why value-based care is where the question gets answered rather than merely argued. Under fee-for-service, no one is paid to find the deterioration earlier; the incentive is to bill the monitoring, not to prevent the admission. Under a risk contract, the economics invert: an avoided admission is money kept, which means the party carrying the risk has, for the first time, a reason to fund the response loop that makes continuous data worth collecting. Value-based care supplies both the motive to prove the hypothesis and the setting — a defined population, measured outcomes, a captured cost — in which it can be proven. The proof and the payer finally coincide.
We are candid about the burden this places on us. Some of the figures the industry cites for monitoring's benefit rest on weak designs — an uncontrolled cohort reporting large blood-pressure drops that no control group isolates, a false-positive rate drawn from a 37-patient study.[7] We will not build the case for continuous monitoring on numbers like those, because the readers who matter will take them apart, and should. The case has to be built on evidence generated the right way: a defined population, a continuous signal feeding a real response loop, measured against that population's own periodic-monitoring baseline, with outcomes and cost as the endpoints. That is a study, not a slogan, and it is the study the category needs.
The implication for a risk-bearing provider is not "adopt continuous monitoring now because it is proven." It is not proven, and we will say so until it is. The implication is that the frontier where prevention gets more effective runs through continuous signal joined to an intelligence layer that acts on it — and that the organizations positioned to generate the evidence, on their own populations, are the ones who will know first whether the frontier pays. We intend to be in the room where that evidence is made. Until it is, this remains what we have called it: an argument we are working to prove.
Sources
- Martens T, Beck RW, et al. (MOBILE), "Effect of Continuous Glucose Monitoring on Glycemic Control in Type 2 Diabetes Treated With Basal Insulin," JAMA, 2021 — jamanetwork.com
- Maiorino MI, et al., "Continuous glucose monitoring … time in range … hypoglycemia in type 1 and type 2 diabetes," Diabetes Care, 2020 — diabetesjournals.org
- Chaudhry SI, et al. (TELE-HF), "Telemonitoring in Patients with Heart Failure," NEJM, 2010 — nejm.org
- Ong MK, et al. (BEAT-HF), "Effectiveness of Remote Patient Monitoring After Discharge of Hospitalized Patients With Heart Failure," JAMA Internal Medicine, 2016 — jamanetwork.com
- Koehler F, et al. (TIM-HF2), "Efficacy of telemedical interventional management in patients with heart failure," The Lancet, 2018 — thelancet.com
- Bowles T, et al., "Outcomes in patients receiving continuous monitoring of vital signs on general wards: a systematic review and meta-analysis of RCTs," Digital Health, 2024 — journals.sagepub.com
- HHS Office of Inspector General, "Additional Oversight of Remote Patient Monitoring in Medicare Is Needed," Sept. 2024 — oig.hhs.gov
