Sibel Health

A sensor engineering company first, with genuine analytics layered on top.

The hard problem this company solved is physical. Producing soft, flexible epidermal sensors that adhere to the fragile skin of an extremely premature neonate, capture clinical grade electrocardiography and photoplethysmography without cables, and survive regulatory scrutiny for accuracy is materials science and signal engineering rather than machine learning. Eight clearances were granted for measurement, and measurement is what the company is bought for.

The intelligence is real and does consequential work. Platform material consistently describes artificial intelligence enabled analytics, the central station clearance covers alarms and alerts across multiple patients, positioning insights are derived rather than measured directly, and the maternal product carries an automated obstetric warning score that converts raw parameters into a prompt for evaluation. Turning continuous multiparameter streams across a ward into a manageable set of signals is analytic work without which the data volume would be unusable.

The balance sits differently from the imaging and radar vendors graded A here. Their sensors produce output no human can read, so inference is constitutive. A vital sign is directly interpretable, so this platform delivers substantial value before any model runs.

Graded B rather than higher because no algorithm performance figure is published anywhere, so the analytic layer is asserted rather than demonstrated. Ask what the analytics contribute over threshold alarms, and at what performance.

Alerting built on a published clinical standard rather than a proprietary score, which is the right choice and rarely made.

The maternal platform integrates an established obstetric early warning system, generating automated signals when prompt evaluation is indicated. That decision matters more than it appears. A recognised scoring system means the alerting logic is public, already taught, already trusted and already auditable, so a midwife receiving an alert knows what triggered it and can evaluate the reasoning. Most vendors in this index substitute a proprietary risk score that a clinician must accept without inspection, and this index has repeatedly noted the difference.

The wider oversight model is conventional and sound. The central station clearance covers alarms, alerts and multi patient monitoring, results reach clinicians who decide, and the software takes no clinical action.

What is missing is calibration, and for continuous monitoring the key figure is alert volume. Nothing publishes notifications per patient per day, false alarm rate, sensitivity for deterioration, time to notification or expected staffing ratio. Alert fatigue is the documented failure mode of ward monitoring programmes, and another vendor in this index publishes both alert volume and clinician engagement rate, which shows the disclosure is achievable.

European material describes smart alarming as a workforce multiplier allowing nurses to cover more patients safely, which is a supervision ratio claim carrying no published number.

Ask for alert volume per patient per day, false alarm rate, and the supervision ratio the platform is designed to support.

Measurement disclosure is precise down to individual parameters per sensor, while the analytic layer is named only.

The hardware specification is unusually complete. The chest sensor is documented as producing electrocardiography, skin temperature, heart rate, respiratory rate, body position and activity, and the limb sensor as producing photoplethysmography, oxygen saturation, pulse rate and skin temperature. The display application is described as showing data from both, and additionally from third party cleared devices for non invasive blood pressure and temperature. Practical detail extends to a five hour wireless charge cycle and a skin friendly adhesive intended to preserve skin integrity, which matters for neonates and for long inpatient stays.

The communication layer is named as a specific open standard rather than described as interoperable, which is a checkable technical claim.

The analytic layer is where disclosure stops. Artificial intelligence enabled analytics is stated repeatedly with no description of what the models produce, no accuracy figures, no operating points for alarms, no model card and no architecture. Positioning insights and smart alarming are named capabilities with no performance behind them.

One element is a genuine transparency positive and belongs here. The maternal warning system is a published clinical scale rather than a proprietary algorithm, so at least one alerting mechanism in the product line is fully inspectable by any clinician.

Ask what the analytics compute beyond threshold alarms, at what operating point, and with what validation.

The most consequential dependencies are visible, and the chain around them is not enumerated.

What can be established is more than usual for a device company. The sensors are the company's own, developed from university research and cleared in its own name, so the core hardware is not a rebadged component. The communication layer is an open published standard rather than a proprietary protocol, which means a customer can understand and independently implement against it, and reduces lock in to an undisclosed dependency. A major patient monitoring manufacturer is named openly as investor and co development partner, so the most significant commercial relationship is disclosed rather than hidden. Third party cleared devices are explicitly supported for additional parameters, so part of the measurement chain is deliberately open.

What is undisclosed is the rest. No cloud or hosting provider is named, no sub processor register was located, no contract manufacturer is identified for sensors produced at scale, and no component or silicon supplier is named.

Training provenance is unaddressed. Artificial intelligence enabled analytics were developed against physiologic data from somewhere, the company operates a clinical trials data platform and monitors patients across more than 20 countries, and nothing states whether or how those datasets relate.

The philanthropically funded work introduces a further question, since data gathered in research deployments across four countries operates under research governance that does not automatically extend to commercial model development.

Ask for the hosting provider and sub processor register, the manufacturing arrangement, and the provenance of data behind the analytics.

Extensive regulatory and deployment evidence across an unusual range of settings, without a published outcome study.

The regulatory record is itself evidence, since eight United States clearances and European Class IIb certification each rest on accuracy and safety data submitted and reviewed, and clearances spanning neonates, adolescents, adults and maternal fetal monitoring mean validation was performed separately for populations with very different physiology.

Deployment is named and geographically diverse. A Danish capital region selected the platform with a major monitoring manufacturer for hospitals across Copenhagen, a leading paediatric hospital in Chicago holds a strategic collaboration, and the maternal platform is deployed across India, Pakistan, Nigeria and Rwanda in a research partnership with a United Kingdom university and a global surgery research unit. Operation spans more than 20 countries, and a separate platform is used by pharmaceutical companies for clinical trials.

Philanthropic funding of 17.5 million dollars directed at low resource maternal monitoring implies external technical and clinical diligence by a funder with its own review standards.

What is missing is outcome measurement. No peer reviewed publication was located in this pass, and no study shows that deploying the platform reduces deterioration, mortality, length of stay or nursing time. The commercial argument rests on continuous monitoring detecting deterioration earlier on unmonitored general wards, which is plausible and unquantified here.

Ask for peer reviewed accuracy studies by population, and any outcome result from the Copenhagen or paediatric deployments.

One meaningful assurance stated, with the operational detail behind it absent.

The published claim is that platform data is synced securely to the cloud with full traceability for patient care. Traceability is a real and useful property rather than a generic one, since an auditable chain from sensor to record is what allows a clinician to verify when a reading was taken and by which device, and it is a requirement for the regulated device context this operates in. Compliance with both United States health privacy law and European data protection law is claimed alongside it.

Everything operational is missing: no encryption description, no retention schedule, no access control model, no deletion process, no data ownership statement and no position on whether collected physiologic data contributes to model development.

The dataset makes those questions weightier than usual. Continuous multiparameter monitoring across neonates, labouring mothers, hospital inpatients and home patients in more than 20 countries produces a physiologic corpus of substantial scale and unusual sensitivity, including fetal monitoring data that concerns two people at once.

The clinical trials platform introduces a distinct flow that is unaddressed. Data collected for pharmaceutical research operates under different consent and retention rules from clinical care, and nothing describes how the two are separated.

One pre emptive note: further clearances cannot move this grade, since device regulation governs safety and effectiveness rather than data handling. Only a published retention and ownership position will.

Ask what is retained and for how long, who owns it, and whether collected data trains models.

Both major regimes named against a specific product, with nothing contractual behind the claim.

The remote monitoring platform is explicitly stated to be compliant with United States health privacy law and European data protection law, and naming both is appropriate for a company operating across more than 20 countries. European Class IIb certification under the current medical device regulation additionally requires documented data governance as part of the technical file, so a governance framework demonstrably exists.

What is absent is everything a compliance officer would ask for next. No business associate agreement template, negotiation stance, execution requirement or subcontractor flow down position was located, and the published privacy policy is a brief page offering a contact address rather than a substantive statement of processing purposes, retention or rights.

The scope of the claim is also narrower than the product line. Compliance is asserted for the home monitoring platform specifically, and nothing states the position for the hospital central station, the clinical trials platform or the maternal system, which operate in different contexts with different data flows.

One population deserves particular attention and gets none. Neonatal and maternal monitoring generate data about patients who cannot consent, in the maternal case covering two individuals through a single system, and no material addresses consent, record ownership or retention for either.

Ask for the agreement template, the retention position, and whether compliance claims extend across the full product line.

Two dedicated passes located no information security certification, no attestation, no trust centre, no penetration testing statement and no vulnerability disclosure policy. The published privacy policy is a brief page offering a contact address.

One distinction needs stating because this vendor's technical record invites it. The open medical device communication standard the sensors are cleared under is described as supporting secure communication between devices, and that is a communication protocol property rather than an organisational security attestation. It says nothing about how the cloud platform is secured, how staff access is controlled, or how incidents are handled. This record credits it on interoperability and not here.

The surface is substantial. Body worn wireless sensors, bedside hubs and a cloud platform operating across more than 20 countries constitute a connected medical device fleet inside hospital networks and private homes, and device authentication, firmware update practice, pairing security and decommissioning of sensors that leave with discharged patients are all unaddressed.

The buyer profile makes the absence more surprising rather than less. European Class IIb certification and a national capital region deployment both imply demanding security review, so a posture certainly exists in private and is simply not published where a prospective buyer could begin diligence.

One pre emptive note: further device clearances cannot move this grade, since device regulation and information security are separate regimes. Only an external attestation, or documentation available under agreement, will.

Ask whether an information security certification is held or in progress, and how sensors authenticate and receive firmware.

Eight or more United States clearances across distinct populations, European Class IIb certification, and an unusual third pathway.

The clearance count is substantial and, more importantly, structured. A seventh clearance announced in March 2025 covered alarms, alerts and a central station, and the maternal fetal platform cleared in April 2026. Separate authorisations cover neonates including extreme prematurity, adolescents and adults, and maternal fetal monitoring, which means accuracy was demonstrated independently for populations whose physiology differs fundamentally rather than extrapolated from one.

European Class IIb certification under the current medical device regulation was granted in June 2026. That framework replaced a lighter directive with materially heavier clinical evidence and surveillance obligations, and a number of legacy monitoring products did not survive the transition.

The third pathway is genuinely uncommon in this index. A cough monitoring sensor has been accepted into a federal qualification programme as a drug development tool through a clinical outcome assessment route, which is a regulatory process for establishing that a measurement is fit to serve as an endpoint in drug trials. That is a different kind of validation from device clearance and opens a different market.

The interoperability certification is also regulatory rather than marketing, since the sensors were cleared operating under an open medical device communication standard.

What is not published is the clearance list with numbers and indications.

Ask for the full clearance list with indications, and the status of the drug development tool qualification.

The company addresses the exact measurement bias that this index graded a direct competitor down for ignoring.

Pulse oximetry has a well documented and clinically serious failure mode: optical measurement through skin overestimates oxygen saturation in patients with darker pigmentation, a discrepancy severe enough to have prompted regulatory review and to have contributed to delayed recognition of hypoxia in patients whose outcomes were already worse. Any optical vital signs sensor inherits that exposure.

Sibel states it has validated the accuracy of its pulse oximeter across a wide range of skin tones. That is a direct answer to the central bias question for this technology, and it is corroborated by context rather than standing alone: the work sits alongside philanthropically funded development for low resource settings and deployments across India, Pakistan, Nigeria and Rwanda, which are exactly the populations where an unvalidated optical sensor would fail hardest.

What holds it short of the top is that the claim carries no data. No accuracy figures by skin tone category are published, no bias study is cited, and validating across a range says nothing about whether performance was equivalent across it. No other subgroup analysis was located, no fairness framework, and no external audit.

Other exposures go unaddressed, notably sensor adhesion and signal quality across body habitus, and performance differences across the neonatal and maternal populations the platform separately serves.

Ask for oximetry accuracy by skin pigmentation category with sample sizes, and whether the skin tone validation extends to heart rate and respiratory rate.

Nothing contractual is published, for monitoring that replaces an existing observation regime in populations where failure is severe.

A dedicated pass located no service level agreement, no accuracy warranty, no uptime commitment, no indemnity and no remediation position.

Substitution is what raises the stakes. The commercial argument is that continuous wireless monitoring frees nurses from manual spot checks and allows safe coverage of more patients, which means wards reduce an existing safety practice in reliance on the system. Availability then becomes a clinical property rather than a service quality one, and no uptime commitment or historical availability figure is published.

The populations sharpen it further than for a general ward product. An extremely premature neonate and a labouring mother are the two settings where undetected deterioration escalates fastest, and both are separately served here. Continuous fetal heart rate monitoring in particular is an area where monitoring failure carries severe and well litigated consequences, and nothing states what is warranted about signal continuity or alarm delivery.

No detection performance is published for any alerting function, so the residual risk a unit accepts when changing its observation protocol cannot be quantified.

The partner channel adds an allocation question, since where a monitoring manufacturer resells or co develops, the contracting party a hospital faces may not be Sibel.

One pre emptive note: further clearances or deployments cannot move this grade. Only contractual terms or published availability and alarm delivery performance will.

Ask for the availability commitment, what is warranted on alarm delivery, and who contracts in partner led deployments.

Interoperability certified by regulators against an open international standard, which is the strongest position this axis has recorded.

The sensors are cleared operating under an open medical device communication standard designed to replace proprietary monitoring networks, and the European certification was announced specifically as the first wireless wearable patient monitoring platform certified to key interoperability standards. That converts interoperability from a marketing claim into a regulated technical property that a third party can verify, which is a categorically different thing from a vendor stating it integrates well.

The commercial corroboration is strong. A major established patient monitoring manufacturer, which is both an investor and a co development partner, publicly cites that open standard as the reason for the collaboration, and the pair were selected by a national capital region to deploy across its hospitals. A manufacturer with its own proprietary monitoring estate endorsing an open standard is a meaningful signal.

Device level openness extends further. The display application is cleared to show data from third party cleared monitors alongside the company's own sensors, so the platform ingests as well as emits, and a development kit lets partners build applications on the cleared sensors, reportedly within a week.

What is missing is the record system. No electronic health record is named, no clinical messaging standard such as HL7 or FHIR is described, and nothing states whether continuous vital signs post to the chart as discrete observations, which is the documentation burden the product is partly sold to relieve.

Ask which record systems are integrated and through what standard.

The on premise topology is clear and the hosted layer is not described.

What can be established is concrete. Sensors sit on the patient, a central hub provides bedside and multi patient display within the facility, and data syncs to a cloud platform. That places display and alarming close to the patient rather than depending on a round trip, which is the correct architecture for a monitoring product where alarm latency matters.

Everything about the hosted side is absent. No cloud or hosting provider is named, no region is stated, no residency commitment is made, no tenancy model is described and no continuity position was located.

The geography makes this consequential rather than academic. Operation spans more than 20 countries including European deployment under a regime with data location and transfer requirements, alongside deployments in India, Pakistan, Nigeria and Rwanda, several of which impose their own localisation rules. Nothing describes whether processing is regionalised or consolidated.

Continuity deserves attention specific to this product. A ward that has replaced manual observation rounds with continuous wearable monitoring has reorganised its nursing workflow around the system, so an outage removes an observation regime rather than degrading one, and no availability commitment or fallback behaviour is described.

One further question follows from low resource deployment. Sites with intermittent connectivity need local buffering, and whether the hub operates independently of cloud availability is unstated.

Ask where data is hosted by region, whether the hub functions during connectivity loss, and the availability position.

Cost is absent from every published surface. Two dedicated passes located no pricing page, no unit of charge, no range, no tiering, no implementation fee position and no minimum commitment.

The unit question is genuinely complex here and entirely unaddressed. The platform combines consumable elements, since sensors use skin adhesives that must be replaced, with durable elements, since the chest sensor is rechargeable with a stated five hour charge cycle, alongside central hub hardware and cloud software. A hospital therefore faces a per patient consumable cost, a device fleet cost and a platform cost, and nothing indicates how any of the three is charged or how they relate.

The product spread compounds it. Neonatal, adolescent and adult, maternal, remote home monitoring and clinical trial use are different buyers with different volumes, and a research platform sold to pharmaceutical sponsors will not price like a ward deployment.

The channel adds a further unknown. A major patient monitoring manufacturer is both investor and co development partner, and where that relationship involves resale or bundling into existing monitoring contracts, the price a hospital sees may be set by the partner rather than by Sibel.

No return proxy is published either, despite a clear argument that continuous monitoring frees nursing time from manual spot checks, which is directly calculable and is not calculated anywhere.

Ask for the consumable cost per patient day, device and hub capital cost, software licensing basis, and whether partner channel pricing differs.

The widest evidenced population coverage in this index, established through separate clearances rather than a single stretched claim.

The span runs from birth to old age. Neonates including those born extremely premature are covered by a miniaturised sensor designed for fragile skin, adolescents from 12 and adults are covered by the main platform, and mothers and fetuses are covered simultaneously through labour, delivery and the postpartum period. Each was cleared separately, which matters because a neonate, an adult and a fetus present entirely different physiology, and a single validation would not carry across them.

Care settings extend equally far: general medical and surgical wards, step down units, hospital at home programmes, home remote monitoring, sleep diagnostics and clinical trials.

The geographic and economic range is what makes this exceptional. The same platform is deployed in Copenhagen hospitals and in India, Pakistan, Nigeria and Rwanda, and the company describes a dual mandate of meeting rigorous regulatory standards while remaining deployable in the world's most resource limited settings. Most vendors in this index address one health system's economics; this one is designed against two, and the maternal work targets a setting where most of the roughly 700 daily maternal deaths worldwide occur without any vital signs monitoring at all.

What is not evidenced is depth of installed base by setting, since no site count or population split is published.

Ask for the installed base by setting and country, and which products carry real deployment volume.