Collaboration and analysis platform for inpatient and outpatient sleep medicine by HTW and Charité

The challenges

Sleep medicine is a growing field in healthcare due to the high prevalence of sleep disorders. The International Classification of Sleep Disorders 3rd edition (AASM 2014) contains 67 distinct sleep disorders. Beside these primary diagnoses, there are also secondary sleep disorders which may be related to medical or mental problems, including substance or drug use or misuse.

The disorders are diagnosed via medical interviews and overnight recordings of multidimensional biosignals, either at a sleep laboratory using so-called polysomnography (PSG) or at home using polygraphy (PG). PSG monitors various body functions including brain waves, eye movements, muscle activity, heart activity (ECG – electrocardiography, blood pressure, pulse wave), respiration (airflow, respiratory movement), blood gases (oxygen saturation, carbon dioxide tension), snoring, and movement of the sleeping person. A PG uses a reduced set of sensors focused on respiratory and cardiovascular events Once a sleep recording has been performed, data are transmitted to a specially trained sleep expert who manually scores the recordings on a computer screen to determine the sleep profile and relevant patterns and events.
Data collected during a home sleep testing is transferred from a patient’s home to a sleep laboratory.  A sleep laboratory shares data with the attending general practitioner and with sleep experts for second opinion. There is also a need for data sharing between healthcare institutions when a patient moves to another place. Additionally, new data transfer requirements are constantly emerging such as data transfer from wearables to analysis facilities. Currently, printed material is still common for sleep reports, while CD ROMs are used for PSG data sharing.
Cloud-based solutions are being slowly introduced in sleep medicine, but currently only in a unidirectional manner, from therapy devices such as breathing support to the device manufacturer. There is no solution yet for secure multi-actor access and analysis of sleep-related health data. Such a scenario does not only require a fine-grained access control and protection of the identifying data but also encryption of the medical data itself, as not only video recordings, but also long term multidimensional biosignal recordings and reports are typically implicitly re-identifiable. This is in particular the case for rare sleep disorders such as narcolepsy – a disorder where patients suddenly fall asleep.

The demonstrator

Figure 1. Secure system for sleep medicine

The proposed system shown in the figure below builds upon an existing central data management system capable of storing sleep data. The system processes metadata, which are structured data, including textual descriptions and sleep parameters within reports. Metadata are stored as key-value pairs. The system also processes unstructured bulk data such as biosignal recordings. Security controls have been applied to enable secure cloud-based sleep medicine, while supporting the aforementioned data sharing requirements.

Metadata are encrypted on the client side using the symmetric searchable encryption (SSE) before storage on the SSE-enabled cloud storage within the Data Management. SSE keys used for the encryption are also themselves locally encrypted using attribute-based encryption (ABE) before storage in the Key Storage. In addition, access to data and services are protected by attributed-based access control (ABAC). Biosignal recordings are encrypted before upload to a bucket storage in the cloud. Search for unstructured data is supported based on structured metadata of the file headers.
The current security system supports the main sleep medicine diagnosis process, from biosignal recording by the patient, to access by the sleep lab professionals who score the recordings and create a report that can be shared with the general practitioner through the platform.

Impact

The reports compiled based on interpretations of the multidimensional signals from the PSG or PG recordings and the medical interviews, are the basis for treatment of patients with sleeping disorders. Usually, a patient returns to the sleep physician at annual intervals because most sleep disorders are chronic conditions.
The proposed ASCLEPIOS solution will facilitate secure exchange of data based on fine-grained access control and cryptographic schemes, promoting the digitalisation of sleep medicine for enhanced collaboration among practitioners, and the preservation of complete sleep-related records – an important aspect taking into account the chronic dimension of sleep disorders requiring the periodic observation of the patients.