IoT Analytics for smart Health and Care IoT Week 2015 Professor Dr. Ch. Thuemmler Technische Universität München, Edinburgh Napier University christoph.thuemmler@tum.de
IoT relevant Statistics Gartner 2013: IoT (excluding PCs, tablets and smart phones) will grow to 26 billion Units installed by 2020. This is a 30 fold increase from 0.9 Billion in 2009. Cisco: By the end of 2014 the number of smart phone subscriptions exceeded the world population with a tendency to grow further Cisco: VNI mobile Forecast: Globally, mobile data traffic will reach 24.3 Exabytes per month by 2019 (the equivalent of 6,079 million DVDs each month), up from 2.5 Exabytes per month in 2014.
Inevitably there has to be interdependency between architecture and topology and data analytics strategies
6 Billion Subscriptions in 2014 Value 18 Billion Euro in 2017 More than 100.000 m-health Apps on the market 5G has the potential to make m-health a Disruptive Technology
Theory Practice Google Health Microsoft Health Vault Trust Legal-ethical Healthcare = Critical Infrastructure
Software to Data Paradigm
Politics EHRs allow health professionals to communicate more quickly and accurately by identifying relevant information more easily and to better plan complex treatment Industry Don t come to us telling us you can upload [data] into our electronic medical record. We don t necessarily want it there our physicians don t want it all there. They really don t need to know how much exercise each of their patients is getting on a daily basis; they just don t have time to process all of that. Christine Folck
Politics Delivering care in a way which is integrated around the individual patient is essential to a new way of working which truly puts the patient at the heart of what we do. Industry If you think about how healthcare is delivered, it s on an ad hoc basis. Someone comes into a hospital, someone comes into a pharmacy, someone comes into a doctor. But beyond those touchpoints, the patients are on their own. There s no real continuity of care.
Data analysis relevant trends in health and care IoT BYOD Bring your own device Neuroempiricism Storing and managing information following neuro-biological principles Subsidiaric - Solve local problems locally FOG / Edgecloud
Distributed Small - and Big Data Deep IoT Analytics Patient to own, control and manage / pre-process / process their own data on their personal devices Health care providers, health insurers, etc. to manage relevant information and legally required documentation Many devices will be gateways for their personal CPS and sensors Distributed Small and Big Data Analysis
Public Health Data analytics of the social web Epidemic monitoring through Twitter data analysis as effective as data analysis of CDC Especially in emerging economies Deep IoT analysis must include mobile phones / smart phones
IERC Book 2015 Pre-processing and data aggregation at device level is a remedy for the congestion problem that often occurs in centralised and hierarchical architectures and will lead to a more scalable design Distributed approach to help to reduce latency and to link more devices Deep IoT analytics processes to be robust and resilient to compensate for missing data as individuals will have the right to withhold information analogue to the right of the silence of the chip in RFID.
Availability of Raw Data only through new Services Interoperability complex as data frequently wrapped in strings with unknown protocols Interoperability complex due to a huge variety of grown legacy systems Huge legal, administrative and technological efforts are required to allow for relatively simplistic transmission of data sets across national borders in Europe (epsos) Health and Care community not yet ready for commercial FI based XaaS solutions (FI-STAR)
Conclusions Deep IoT analytics in Health and Care will be based on a distributed architecture Networks will be more intelligent (FOG, Edge-Cloud, Smart Grid) Information will be processed not only centrally but on different levels Governance of Distributed Analytics approach for Health and Care will be critical People will have control over their personal information and strategies providing Deep IoT Analytics have to honor this Deep IoT analytics requirements need to be flagged up to the 5G program as early as possible and there should be early interaction especially with regards to large scale demonstrators
References Cisco (2015), VNI report 2015, http://www.cisco.com/c/en/us/solutions/service-provider/visualnetworking-index-vni/index.html Ferrer-Roca O, Tous R, Milito R (2014), Big and Small Data: The Fog, IEEE Conference Publications Niehaus E, Herselman M, Babu AN (2009), Principles of Neuroempiricism and generalization of network topology for health service delivery service, Indian Journal of Medical Informatics, Vol 4 (1), pp 1-16 Nunna S, Kousaridas A, Ibrahim M, Dillinger M, Thuemmler C, Feussner H, Schneider A (2015), Enabling Real-Time, Context-Aware Collaboration through 5G and Mobile Edge Computing, IEEE Conference Publication epsos, www.epsos.eu FI-STAR, http://fistarblog.com Achrekar H, Gandhe A, Lazarus R, Yu SH, Liu B (2011) Predicting Flu Trends using Twitter Data, IEEE Conference Publications Lampos V, Cristianini N (2010) Tracking the flu pandemic by monitoring the Social Web, IEEE Conference Publication delivery