BioIntelliSense

The BioButton automates the four hourly vital signs round: it records respiratory rate instead of the estimated figures that cluster at 18 or 20 on paper charts, spares patients night time checks, and lets one clinician watch hundreds of patients from a configurable dashboard. Notifications come from statistically meaningful changes in each patient's own trend, which the user terms describe as fact based and the company's study announcement calls algorithmic and rules based; the BioButton System clearance (K241101) describes notifications when data fall outside clinician specified parameters.

The privacy policy, modified August 19, 2026, says product data is used for analytics and training algorithms without naming the function those algorithms serve, and the home page makes no AI claim.

Every notification goes to a clinician, who decides what to do with it in a configurable dashboard, and large deployments add a staffed command center, as at Houston Methodist. Notifications fire on statistically meaningful change in a patient's own trend rather than at population cut offs, at fewer than one per patient per day, and the company reports clinician engagement with them at 99.71 percent.

The user terms leave every decision with the clinician: the product does not determine when a patient needs immediate intervention, and its notifications are not to be relied on alone for diagnosis or treatment. The method that turns a drift into a notification is not shown publicly, so a quiet patient and an insensitive setting look alike on the dashboard.

The measurement chain is named component by component: the BioButton on the chest, BioHub gateways or the BioMobile app, BioCloud analytics and the BioDashboard, which carries the AlertWatch software the company acquired, with a dedicated charging station for the rechargeable inpatient device. The sensor records up to 1,440 measurements per patient per day, and notifications come from statistically meaningful trends in each patient's own data. An independent 2025 study in Frontiers in Pediatrics tested the activity and position functions in children against reference measurement.

How personal baselines are set and adapted, which functions use the algorithms the privacy policy says are trained on product data, and how changes are released are not described.

BioCloud is the company's own analytics service, and the AlertWatch clinical intelligence software it acquired now runs as the BioDashboard. Data goes to contractors and service providers for IT and infrastructure support, care management and marketing services, and protected health information is never passed to them for their own purposes.

The hosting provider, the makers of the sensor, gateway and charging hardware, and any outside model behind the trained algorithms are not named.

Two peer reviewed studies in the Journal of Clinical Medicine carry the clinical case. The first, a retrospective observational study of 11,977 medical surgical patients at two US hospitals over 15 months (2024), found length of stay at the first site fell from 3.07 to 2.75 days over 12 months, 73 and 66 percent of deterioration events had a notification in the prior 24 hours, on average 14.8 hours ahead, and notifications ran at 0.97 and 0.65 per patient per day; it had no control group. The second, from Houston Methodist, followed 1,120 high risk patients transferred to intensive care and reports a 27 percent lower relative risk of death with continuous monitoring.

The company also cites a 7.7 hour shorter stay, manual vital signs time cut from 420 to 210 minutes a day, and more than 400,000 patients monitored by October 2025. An independent 2025 study in Frontiers in Pediatrics found mixed accuracy for activity and position detection in children, including 0.52 sensitivity for moderate to vigorous activity and 78.6 percent accuracy for body position.

Retention runs by data type. The privacy policy, modified August 19, 2026, keeps health and biometric data from the product for the life of the active account plus one year or as healthcare rules require, account data for a year after closure and logs for up to a year. Product data feeds statistical data sets, analytics and algorithm training, and protected health information is never passed to third parties for their own purposes. Data is encrypted on the device, in BioCloud and in transit, and health data outside HIPAA gets the same or similar safeguards.

The de identification standard for the statistical data sets and training, any opt out from training, and safety engineering for missed or false notifications go undescribed.

The privacy policy, which reaches the BioButton, gateways, apps and BioCloud, refers to the company's HIPAA business associate agreement for how long data is kept, separates protected health information from health data outside HIPAA, and leaves the Notice of Privacy Practices with the provider, stating that the company is not a medical group.

The agreement itself, the tier it applies at and its terms for subcontractors are not published. Home monitoring sends the same data over a patient's own network or a cellular gateway.

Security rests on encryption on the device, in BioCloud and in transit, reasonable safeguards and breach notice to European authorities within 72 hours where feasible, as the privacy policy states. The surface is a connected device fleet: body worn sensors, gateways placed through buildings, charging stations, a cloud layer, and rechargeable devices that pass between patients or go home with them.

No outside audit, certification or trust center appears on the site.

The BioButton System (K241101, September 2024) and the earlier BioSticker System (K191614, December 2019) are cleared under product code DRG for continuous collection of physiological data at rest in home and healthcare settings, for adults outside critical care, measuring heart rate, respiratory rate and skin temperature, with activity, position, sleep and gait as wellness measures, and for notifying professionals when data fall outside clinician specified parameters. The AlertWatch acquisition brought cleared clinical intelligence software, including AlertWatch:AC (K210160, 2021), which now runs as the BioDashboard.

The user terms of use (February 2024) keep the product inside that frame: notifications report specific conditions such as resting heart rate or skin temperature, make no diagnosis or risk recommendation, and do not decide when a patient needs immediate intervention. The marketing describes deterioration detection on each patient's own trend, a function the clearance text does not name.

Testing published so far concerns the sensor rather than a model. The BioButton System clearance (K241101) reports clinical testing in 52 subjects, with heart rate within 5 beats per minute and respiratory rate within 3 breaths per minute at rest, and states equivalent performance across race and ethnicity, gender, age and body mass index groups without per group figures. The deterioration notifications are rules based, and no fairness review, drift monitoring or governance framework covers the algorithms the privacy policy says are trained on product data.

Chest worn sensing has its own exposure. Sensor contact, adhesion and signal quality vary with chest size, tissue and body hair, and a personal baseline built on a noisier signal weakens notifications for exactly those patients.

The service level agreement makes BioCloud and the BioDashboard available 99.9 percent of each month, measured for BioCloud as the share of hourly vital sign readings delivered within 20 minutes, and returns 3, 5 or 10 percent of the monthly fee as availability falls below 99.5, 99.0 and 97.0 percent. A clinician reviews every notification before anything is done.

What the notifications say is left to the clinician. The agreement assigns all clinical questions to the customer, the user terms cap BioIntelliSense's liability to an individual user at the lesser of fees paid or $100, and no sensitivity figure for deterioration notifications is offered.

For home monitoring, insights are displayed in the BioDashboard and the customer's EMR, and the platform runs from sensor to gateway to BioCloud to dashboard, with Houston Methodist working from a central command center on it. The company presents automated vital sign collection and documentation as the manual burden it removes.

No record system, interface standard or marketplace listing is named, which leaves open whether readings post to the chart as discrete timed observations or are copied across by staff.

All personal data is stored and processed on servers in the United States, with transfers from the European Union under Standard Contractual Clauses. The data path runs from the BioButton through BioHub gateways on the hospital network, or a cellular BioHub or the BioMobile app at home, to BioCloud and the BioDashboard, and the service level agreement sets availability for both cloud services.

The hosting provider, how one health system's data is kept apart from another's, and recovery objectives are not described.

Hospitals subscribe to the BioButton as a data as a service product under a master services agreement, paying a monthly service fee that the service level agreement uses to calculate refunds for missed availability. Home monitoring is marketed as reimbursable under remote patient monitoring codes, and the company offers help to applicants for the federal Rural Health Transformation program.

No price, unit of charge or implementation cost is offered. The company makes its case from the cost of the alternative: telemetry at up to $1,400 per patient per day, 22.3 percent of patients assigned telemetry inappropriately at admission in one review, and routine vital sign checks causing 33 percent of patient sleep interruptions in a 2024 multicenter study.

The general ward patient is the center: medical surgical units, specialty care areas and emergency departments, with the same sensor following patients into hospital at home and post acute care. Intensive care is out of scope because those patients already have continuous monitors. Houston Methodist runs the BioButton across 2,653 non intensive care beds at eight hospitals, and the Journal of Clinical Medicine study covered nearly 12,000 medical surgical patients at two hospitals.

At home, patients leave wearing a rechargeable BioButton with a 10 day battery, a charger and spare adhesives, and data reaches the care team through the BioMobile app or a cellular BioHub; home monitoring is marketed as reimbursable remote patient monitoring. Home and post acute use carry no study of their own, and a 2025 Frontiers in Pediatrics study tested activity and position detection in children as research rather than a marketed use.