Tietovarannot tieteen ja tutkimuksen Graalin malja

Size: px

Start display at page:

Download "Tietovarannot tieteen ja tutkimuksen Graalin malja"

Maximillian Collins
5 years ago
Views:

1 VTT TECHNICAL RESEARCH CENTRE OF FINLAND LTD Tietovarannot tieteen ja tutkimuksen Graalin malja Mark van Gils, Docent Niilo Saranummi, Professor Emeritus contact:

2 A holy grail? Wikipedia: Different traditions describe it as a cup, dish or stone with miraculous powers that provide happiness, eternal youth or sustenance in infinite abundance. For us? A methodology for scientific work that leads us to high-impact solutions for meaningful (health) challenges 2

3 High Impact? healthcare professionals Acute Care R&D, Scientific Community increased knowledge improved decision making Sub-acute care increased competetiveness Successful Solution more efficient care, disease mgmt Finnish & international industry reduced cost improved quality of life Patient at home Society 3

4 Is there a path to get there? 4

5 Knowledge mining from over 20 years of work in ICT for Health projects >100 R&D projects EU FP4 -> H2020; TEKES; Academy of Finland; contract research: SMEs, multi-nationals Some very successful, some 5

6 Example of a succesful attempt Decision support tools to help in early differential diagnosis of neurodegenerative diseases Research projects Deployed as commercial product in several hospitals 2017 Expanding ecosystem development continuing 6

7 In the beginning - a clear clinical need 7

Internationally recognised clinical partners from multiple centres Close involvement of top

co-operators give feedback/iterate potential solutions/protypes provide retrospective data

engineers do something sensible depression Parkinson PNFA Diversity of clinical partners is

8 Internationally recognised clinical partners from multiple centres Close involvement of top clinical partners is essential to formulate the problem attract good quality further co-operators give feedback/iterate potential solutions/protypes provide retrospective data collect propective data validate methods, and publish findings in general make sure that engineers do something sensible depression Parkinson PNFA Diversity of clinical partners is key nurses to highest senior clinicians, trainees to emeritus professors different cities, different countries 8

9 Method development The Need for Data Data is needed for first development of methods (retrospective, limited dataset) Data is needed for first validation of methods (other retrospective) Data is needed to seriously validate (prospective, at various locations) Data is needed for more advanced method development (eg from Alzheimer s early diagnostics to differential diagnostics of different dementias) (more retrospective and prospective). 9

10 Concept: understandable interpretation of a complex disease state by data-driven machine learning and visualisation methods PredictAD EU project 10

11 Iterative Method implementation into decision support software Testing in real-life situations, not only accuracy, but also usability, how it affects decision making Update of prototypes to next version and test again and again PredictND EU project 11

12 And more complex method development mechanistic predictions & biomarkers Platforms for Biomedical Research & Clinical Decision Support healthcare research citizens phenomenological inferences & associations Mechanistic Modelling Top-down Biophysical FE Models Metabolic Pathways Biosignals Modelling hypothesis parameters Phenomenological Modelling Bottom-up Statistical associations Connectivity networks personalised environment & information Human health data clinical population environmental personal & environmental data EU project 12

levels ethical approvals harmonisation of data between different sites (different equipment, ways of interviewing patients,

13 Requires even more new data collection at different centres Prospective data collection is extremely cumbersome and almost always underbudgetted and delayed technical installation issues at sites, cybersecurity, privacy, data handling, formats paperwork at many levels ethical approvals harmonisation of data between different sites (different equipment, ways of interviewing patients, treatment protocols ) suitable patients don t show up at the rate we had planned patients disappear from follow-ups missing data, poor data 13

14 Scientific credibility Publishing in high-level international journals is essential to prove credibility and continue research with more and more partners in related projects May be tricky within a project when data collection is delayed and project end date is fixed -> having a cluster of projects can be of great help Both health journals as well as technical journals are needed 14

More credibility demonstrate value of the methods in other challenges

sets of heterogeneous data (TEKES project MADDEC) Screening for

15 More credibility demonstrate value of the methods in other challenges Outcome prediction and intervention planning for Traumatic Brain Injury patients ( ) Prediction of adverse events after cardiac surgery using large sets of heterogeneous data (TEKES project MADDEC) Screening for gastric cancer using breath analysis ( ) - mass data for better risk analysis in cardiovascular diseases 15

A citizen portal, aiming to use low-cost measures, such as web-based cognitive tests, games and gait analysis, to

16 Expanding the ecosystem: citizen portal for early detection of risk of dementia Detecting individuals at high risk of developing demential as early as possible -> in home setting -> more, very different data, needed! A citizen portal, aiming to use low-cost measures, such as web-based cognitive tests, games and gait analysis, to assess the risk. Early validations (PredictND EU project) show that these measures correlate highly with traditional measures 16

When there is enough evidence taking the steps

Business development discussions (spin-off,

) TEKES Tutli project Creation of an actual

17 When there is enough evidence taking the steps towards commercialisation IPR creation, protection Business development discussions (spin-off, licensing?) TEKES Tutli project Creation of an actual start-up company Product lines (imaging, decision support) CE marking done FDA approval upcoming Combinostics 17

18 Databases PredictAD ( ) Retrospective (N>1000): ADNI, AIBL, Kuopio, Copenhagen ( ) Retrospective (N = thousands): Rotterdam, DESCRIPA, ADNI2, Sheffield, Kuopio. Prospective (N=350): Sheffield, Lido, Kuopio PredictND ( ) Retrospective (N approx 3000): Kuopio, Amsterdam, Copenhagen, ADNI, EDAR Prospective (N=800): Kuopio, Amsterdam, Copenhagen, Perugia brain injury Retrospective (N=approx 1000): Cambridge, Turku, Rotterdam Prospective (N=400): Cambridge, Turku cardiac case retrospectove (N=90.000), prospective (N=10.000) 18

19 healthcare professionals Acute Care R&D, Scientific Community increased knowledge improved decision making Sub-acute care increased competetiveness Successful Solution more efficient care, disease mgmt Finnish & international industry reduced cost improved quality of life Patient at home Society 19

20 healthcare professionals The path to reach high impact Acute Care Elaborate R&D, Scientific development increased and validation in many different settings, Sub-acute care Community knowledge requiring lots of different data collections Finnish & international industry Successful Solution Society improved decision making improved quality of life more efficient care, disease mgmt Development into an actual product or service that is brought increased to the market competetiveness Attract customers: provide reduced evidence to healthcare cost professionals of the added value Patient at home 20

Regional innovation hubs and their role to international cooperation

Regional innovation hubs and their role to international cooperation Brazilian delegation 22-26th August 2011 Ministerial Adviser, Dr. Pirjo Kutinlahti Outline Finnish cluster policy approach Centre of