Connected Health: What If the Device in the Field Could Actually Act?

The perception of physical assets is shifting. A case in point comes from the world of large energy utilities. An asset like a substation is no longer merely a network component — it evolves into a continuous source of real-time data. As a result, value today lies in orchestrating the information generated across an asset’s entire lifecycle. There are domains where this perspective has yet to take hold. Healthcare is one of the most striking examples. When a defibrillator in intensive care carries a compromised battery that no system has flagged, the consequences are of an entirely different order of magnitude than a neighbourhood blackout because lives are at stake.

From the Dominant IoT Model to the Edge-First Paradigm

The prevailing IoT model works as follows: a sensor detects, data is sent to the cloud, an algorithm processes it, an operator receives a notification. It has delivered results but carries a structural flaw that becomes untenable in high-criticality environments: the latency inherent in the data–processing–action cycle. In specific settings, that latency can make all the difference.

The edge-first paradigm inverts this logic. Intelligence resides in the device itself, in direct contact with the operational environment. An embedded software agent analyses in real time, detects anomalies, and acts locally within milliseconds. What changes is not the data pipeline but the point of decision-making which shifts from the end of the cycle to its very beginning.

Three Scenarios from the Healthcare

The hospital environment lends itself to numerous edge-first implementations. The first scenario unfolds in emergency care, around defibrillators. An embedded agent changes everything: it monitors charge and discharge curves, detects degradation in advance, and proactively generates a maintenance request calibrated to actual device conditions. Results from real-world implementations confirm the direction: a study published in the Journal of Medical Systems documented maintenance cost savings of up to 25%, with a payback period of under one year. The most significant paradigm shift, however, is not economic but clinical. An inspection of over 23,000 public defibrillators conducted between 2013 and 2017 and published in Scientific Reports found that nearly one in six devices was not ready for use at the time of inspection — due to depleted batteries, expired electrodes, or device malfunction. In a context where survival rates from cardiac arrest decline by approximately 10% for every minute without intervention, and where the immediate availability of a functioning defibrillator raises the probability of survival from 8–10% to as high as 50–74%, operational continuity of the device becomes a direct clinical variable.

The same logic applies to dialysis machines — a category almost entirely disconnected from structured IoT architecture. An agent embedded in the device monitors transmembrane pressure, blood flow, and dialysate conductivity, building a predictive model specific to that machine. The most compelling contribution, however, is systemic: when an entire fleet converges on a shared platform, correlating operational performance with clinical outcomes becomes possible. This is something structurally unattainable with isolated devices.

The third case is robotic surgery, where latency is not a matter of efficiency but of physical safety. A 2025 systematic review published in the Journal of Robotic Surgery established 200ms as the critical safety threshold: beyond that value, the desynchronisation between gesture and robotic arm response directly impacts surgical precision — a safety limit, not a design target, as high-precision systems are engineered to operate at 20ms or below. Cloud-centric architectures cannot guarantee that level of consistency. An edge node physically proximate to the operating theatre can.

From Device to Platform

These scenarios share a common thread: every time intelligence migrates to the data-generation layer, something larger than a single optimisation is created: an ecosystem node.

The IoTomic approach developed by BIP is grounded in precisely this premise. The model is structured across three layers: at the base, data generators — biomedical equipment, monitoring systems, and the energy infrastructure of critical wards; in the field and at the centre, a shared platform built on microservices architectures and open protocols; distributed across both the field and the platform, the agentic components and applications that act on the data collected. The underlying logic is not “one device, one system, one closed application” — it is that of a marketplace in which different operators contribute data and access services developed on the same shared infrastructure.

A critical enabler is software portability onto existing hardware. Open SDKs and documented APIs lower the barriers to adoption for third-party developers and operators, transforming already-installed assets into active nodes within an open ecosystem without requiring device replacement, and enabling virtually anyone to build compatible, compliant solutions on the same infrastructure. This marks the transition from the logic of cyclical modernisation to one of continuous ecosystem evolution: new capabilities delivered via software, new agents activated without physical intervention, value that accumulates on the platform rather than depreciating with each individual device.

For biomedical technology manufacturers, this inverts the revenue model shifting the value proposition from the sale of a single device to guaranteeing the operational readiness and continuity of an entire facility. Those who make this transition first will hold a structural competitive advantage. Those who remain anchored to the logic of the closed device risk becoming a commodity.

How Healthcare Facilities Must Reposition Themselves

By July 2025, the FDA had authorised over 1,250 medical devices incorporating embedded AI/ML, with 295 new clearances in that year alone. A simulation study published in Engineering Proceedings in 2026 estimates, across representative classes of hospital devices — imaging systems, patient monitors, analysers, sterilisers, and infusion pumps — a reduction of 18–35% in unplanned failures and 10–22% in energy consumption, compared with traditional maintenance schedules. Intelligence embedded in devices is no longer a future prospect: it is a regulatory reality that the market is already absorbing.

The question that will define the positioning of a healthcare facility in the decade ahead is not how many devices are connected. It is: what kind of platform do we want to become? The devices already hold all the information needed to keep the facility ready and resilient. What stands between that data and meaningful action today is a paradigm shift — one that is entirely within reach.