The design and development of the first mICF proof of concept started in February 2016. This frontend prototype, developed in Finland, includes the co-creative design of a mobile interface enabling real mICF users (n=60) to capture ICF-related structured information and to provide a functioning profile through linking with the FunctionMapper. The objectives of this phase of iterative, agile development of the mobile user interface are to learn more about user needs and acceptance based on tests and feedback; to create solutions that support findings; to test the application in various settings; and collect and analyse feedback for continuous learning. The rapid prototyping of this mICF frontend application was initiated by co-creative service design workshops that is essentially a philosophy where a product or service is constantly being improved: learning, refining, experimenting, modifying and then learning again. The methodologies of lean UX are used in the agile, iterative mock-up prototyping for creating the proof of concept through the active participation of persons with disabilities.
Health Databank and interoperability layer
The Health Databank comprises of all data aspects as related to the system. This includes not only the management of data, but also the analytics. It is envisioned that some of the analytical processes will be automated using machine learning techniques. Such techniques will not only allow the end user to extract informed decisions, but also make provision for outcome predictions based on the data.
Distributed computing architectures are now the providing platforms for Big Data analytics, very suitable for the types of data being produced in the mICF context. These architectures allow for real-time data interaction (including probing the data in a reiterative manner). Being distributed and typically hosted in the cloud, allows very short response times, much less than dealing with data in a centralised warehouse. One of the biggest advantages of modern Big Data architectures is the fact that they can handle complex content instead of just structured content. This is very important in the mICF context, since we will be processing structured and unstructured data. Structured data will be presented as ICF codes, but some interfaces will generate additional (unstructured) data which will inform the analytic (machine learning) models.
From a cost perspective the modern analytic architectures are also more affordable, since they are typically built on hosted services using open source software components. However, we will need to be cautious about ethical implications of person-owned and person-shared data and of widening population-level inequalities.
Interoperability (the back-and-forth exchange of data among different points-of-service/systems) is a very important aspect of the system’s design. Modern international health care enterprise standards such as Health Level Seven International (HL7) FHIR (Fast Healthcare Interoperability Resources) will be used to ensure that the informed decisions made by the data analytics layer are modelled and communicated to external points-of-service (for example social services). Internally, the interoperability between the FunctionMapper, mobile interfaces and the data components of the system will follow industry standard communication protocols (i.e. RESTful API) to allow extendable and flexible integration.
The necessary measures according to national privacy and personal data security laws will be put in place to ensure the protection and privacy of person-owned as well as person-shared data. Standard measures that incorporate not only compliance but also interoperability are to be implemented so as not to widen existing population-level inequalities.