Abstract
Lack of standardization is highly visible while we use historical data sets or compare our model with others that use IoMT devices from different vendors. The problem also concerns the trust in highly decentralized and anonymous environments where sensitive data are transferred through the Internet and then are analyzed by third-party companies. In our research we propose a standard that has been implemented in the form of framework that allows describing requirements for methods and platforms that collect, manage, share, and perform data analysis form the Internet of Medical Things in order to increase trust. Further, we can distinguish two types of IoMT devices: passive and active. Passive devices measure some parameters of the body and save them in databases. Active devices have the functionality of passive devices and moreover, they can act in a defined way, eg.: inject directly into the patient’s body some elements such as a medicament, electric signals to the nervous system, stimulus pacemaker, etc. Nevertheless how to create a safe and transparent environment for using data active sensors, developing safe ML models, performing medical decisions based on the created models and finally deploy this decision to the specified device. While the IoMT devices are used in real-life, professional healthcare the control system should offer tools for backtracking decisions, allowing e.g. to find who made a mistake, or which event caused a particular decision. Our framework provides backtracking in the IoMT environment in which for each medical decision supported by ML models we can prove which sensor sends the data, which data was used to create prediction/recommendation, what prediction was produced, who and when use it, what medical decision was made by who. We propose a vendor transparency framework for each IoMT devices and ML models that will process the medical data in order to increase patient’s privacy and prevent for eventual data leaking.
Highlights
The rapid development of a new type of IoMT devices dedicated to professional and non-professional customers conduct a problem with vendor lock-down, ease integration, validation different data sources, and securely management of the created data sets which permissions who can use the data
We focus on a framework that is dedicated to collect data from many IoMT devices in real-time and to perform AI/ML analysis in privacy-be-design isolated containers
We propose in our framework track each medical decision and describe it by devices, sensors, algorithms, results which were used by the doctor while he prepares a certain decision
Summary
The rapid development of a new type of IoMT devices dedicated to professional and non-professional customers conduct a problem with vendor lock-down, ease integration, validation different data sources, and securely management of the created data sets which permissions who can use the data. Vendor-agnostic – each IoMT device should offer public API for accessing to data registered by sensors in some universal format such as shown in Listing 1 where userId is a unique user identifier, deviceId is a gateway id, sensorId is a unique id of sensor which produce data, typeOfSensor is a proposed classification that unified types of sensors and make easier the process of finding appropriate data, timestamp is a Unix-like description of date and time when data was registered by the gateway, sessionId is a unique value for the user which describe that measured value comes from one training/measurement session, objectAsBase is any value obtained by the sensor sensorId encoded using base, hashValue is a result of a hash function (e.g., SHA-256 or SHA-512) from a concatenation of all mentioned attributes (string values). HL7 FHIR does not specify the aspect of using the IoMT or ML models
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
