New health gateway solves one of healthcare’s biggest digital problems

The system focuses on local data processing, which offers several advantages over cloud-dependent IoT architectures. By executing analytics and running machine-learning inference directly at the fog layer, the B-Health IoT Box ensures low-latency responsiveness and reduces bandwidth needs. These capabilities support continuous monitoring applications that require rapid evaluation of biometric data.

A new fog-computing medical data gateway capable of harmonizing large volumes of health and well-being information from heterogeneous devices has been validated in real-world deployments involving thousands of users, marking a significant advance in the push for interoperable digital health ecosystems.

The system, known as the B-Health IoT Box, is detailed in the study “The B-Health Box: A Standards-Based Fog IoT Gateway for Interoperable Health and Wellbeing Data Collection,” published in Sensors.

Researchers behind the project developed the gateway to address long-standing weaknesses in digital health infrastructure, where fragmented data streams, incompatible communication protocols, and limited support for contextual information hinder the development of complete Personal Health Records and slow the adoption of next-generation connected care technologies.

The study reports that the B-Health IoT Box collected and transmitted more than 1.5 million datasets across multiple clinical, occupational, and community settings while supporting over 4500 users, demonstrating its maturity and scalability.

Interoperability failures driving system redesign

Most patient records hold only clinical data that mirrors hospital-based Electronic Health Records, omitting the diverse behavioral, environmental, and physiological indicators that everyday devices can capture. The authors highlight that although IoT and Internet of Medical Things systems have spread across the market, their benefits remain limited because health data often arrives in inconsistent formats that cannot be merged without extensive customization. These gaps prevent clinicians and researchers from using sensor-based data for monitoring, early diagnosis, or personalized care.

The B-Health IoT Box was developed as a direct response to these interoperability failures. The system uses a layered architecture aligned with Internet of Things reference models and integrates established healthcare standards such as HL7 FHIR, openEHR, the ISO/IEEE 11073 family, and the Continua Design Guidelines. By embedding semantic harmonization within the fog layer itself, the gateway processes, normalizes, and structures data before it is transmitted to cloud systems or electronic health platforms. This design reduces reliance on remote processing, lowers latency, and supports near real-time decision-making at the point of care.

A fog-computing architecture built for health systems

At the physical level, the B-Health IoT Box connects to a wide array of devices including wearables, smart textiles, environmental sensors, and medical-grade instruments. A communication layer manages the flow of information through wired and wireless protocols such as USB, Bluetooth Low Energy, Wi-Fi, Zigbee, and Ethernet. Above this, the information layer interprets incoming data, applies transformations, and stores results temporarily to support fast local processing.

Its function layer carries out the core intelligence. A service manager orchestrates workloads and ensures system continuity. A fog-computing node handles real-time filtering, event detection, lightweight analytics, and local AI inference, enabling the system to recognize abnormal physiological or environmental patterns without depending on cloud connectivity. This structure allows the gateway to deliver timely feedback for applications such as fall detection, posture monitoring, physical therapy guidance, or acute-care alerts.

The application layer provides interfaces for configuration, visualization, and third-party integration, while the interoperability layer connects the device with external health information systems. The top cloud layer supports links to European eHealth platforms developed under Horizon 2020 projects, ensuring that data moves smoothly into regional and cross-border health infrastructures.

Support for health standards and modular device integration

A defining feature of the B-Health IoT Box is its strong compliance with healthcare communication standards. By integrating the Continua Design Guidelines and the IEEE 11073 framework, the gateway aligns data models from both medical-certified devices and consumer sensors. The study emphasizes that the platform can ingest data from smart t-shirts, insoles, fitness wearables, posture sensors, and traditional medical devices, treating each as a reliable source of health and contextual information.

Custom adapters allow non-compliant technologies to become interoperable by transforming raw device output into semantically aligned formats. Once harmonized, the B-Health IoT Box exports structured health information in FHIR, enabling seamless exchange with Electronic Health Records, Personal Health Records, and digital health portals. This approach reduces fragmentation and supports the creation of richer patient datasets.

The gateway also uses general-purpose IoT protocols such as MQTT, HTTP(S), RESTful APIs, JSON, and I²C, offering flexibility across deployment environments. This combination of standards ensures compatibility with existing digital health infrastructures while supporting growth into new use cases.

Local intelligence, reduced latency, and privacy-centric design

The system focuses on local data processing, which offers several advantages over cloud-dependent IoT architectures. By executing analytics and running machine-learning inference directly at the fog layer, the B-Health IoT Box ensures low-latency responsiveness and reduces bandwidth needs. These capabilities support continuous monitoring applications that require rapid evaluation of biometric data.

Security and privacy were key priorities in the design. The platform uses encrypted communication, encrypted storage, SSH-based authentication, role-controlled access, configurable firewalls, and comprehensive logging. These features make the gateway suitable for clinical environments and research deployments that must comply with strict regulations, including GDPR. Its architecture is also prepared for privacy-preserving federated-learning extensions that would allow distributed model training without exposing sensitive data.

Comparison with commercial IoT platforms

The study includes a comparison with commercial IoT ecosystems such as AWS Greengrass, Azure IoT Edge, Losant IoT, and Particle Photon. Unlike these general-purpose systems, the B-Health IoT Box integrates healthcare standards natively, removing the need for additional adaptation layers. Its modular hardware, built on open components like Raspberry Pi with 3D-printed enclosures, enables rapid prototyping and on-site customization, which is less feasible with closed industrial platforms.

The authors note that embedding semantic harmonization in the fog layer gives the system a strong advantage over cloud-centric competitors, which often require extensive preprocessing pipelines at the server level. By bringing semantic alignment closer to the data source, the gateway improves consistency and enables intelligent decision-making at the edge.

Application scenarios across Europe

The B-Health IoT Box has been tested across several European Union projects, each demonstrating its versatility and adaptability.

One major deployment involved back-pain prevention and treatment training within the Smart4Health and SmartBear Horizon 2020 initiatives. The system collected training metrics from a lumbar extension physiotherapy machine, integrating data such as angle measurements, force testing, repetitions, and movement patterns. These values were harmonized and structured as FHIR DiagnosticReport resources, allowing clinicians and patients to use standardized, machine-readable performance summaries for monitoring progress and adjusting therapy programs.

Additional deployments took place in occupational and industrial environments. Projects such as DIH4CPS, ICU4COVID, and AGILEHAND used the gateway to monitor posture, biomechanical load, and workplace environmental data for workers and healthcare staff. These large-scale tests produced more than a million datasets, validating the system’s ability to handle high data volumes and diverse IoT configurations.

Across all settings, deployments followed ethical guidelines and GDPR requirements, ensuring secure processing, participant consent, and strong privacy safeguards.

Addressing gaps in digital health infrastructure

Existing IoT health systems struggle with a series of persistent challenges: data management complexity, scalability constraints, inconsistent standards, human-machine interaction barriers, and cybersecurity vulnerabilities. The B-Health IoT Box is positioned as a comprehensive response to these issues.

By offering a unified architecture that can ingest, process, harmonize, and distribute data from heterogeneous sources, the gateway reduces technical burdens on health organizations and developers. Its layered architecture supports emerging concepts such as digital twins, advanced time-series analytics, and edge-based disease monitoring. These capabilities position the system as a foundation for next-generation personalized and preventive care models.