This document discusses how big data can be used to address major challenges in prostate cancer research and clinical practice. It proposes a model to create a standardized, collaborative data platform that integrates large clinical datasets from multiple European and non-European sources. This would allow for novel analytics and computational approaches to gain new insights into prostate cancer outcomes and improve standardized care pathways. Key elements for success include involvement of experts in prostate cancer and big data as well as all stakeholders, including patients. Education is needed to encourage data sharing while protecting privacy.
3. Major Prostate Cancer challenges
90% of what is published in the
scientific literature is unreliable and
unfit to trigger a change of Clinical
Practice Guideline Recommendation
4. Major Prostate Cancer challenges
• Insufficient knowledge on patient characteristics.
• Poor definition and lack of standardisation of Prostate
Cancer-related outcomes.
• Lack of standardised “care pathways” across different
European geographies.
• Inability to incorporate real world clinical outcomes
data into the management of Prostate Cancer
(screening, diagnosis & treatment).
5. CLINICAL DATA
DATA
STEWARD
ANALYTICAL ENVIRONMENT
STORAGE OFANALYSISRESULTS
LARGE-SCALE PARALLEL COMPUTATION
MULTI-PLATFORMANALYSISPIPELINES
tranSMART
ANALYSIS USER SPACE
‘OMICSDATA
STUDYDESIGN&
DEMOGRAPHICS
HEALTH RECORDS
HARMONIZED
DATASETS
GPU
DATA ANALYST
KCL, UCL, Marsden
EORTC
Erasmus / / Nijmegen
Martini-Klinik/ NCT/
Dresden/Fraunhofer
EU partners
bringingbigdata
Epidemiology
datasets
VA Dataset
CaPSURE
Non-EU
partnerswith
data sources
HARMONISED
DATA STORE
DATA STANDARDS &
ONTOLOGY
MANAGEMENT
DATA EXPLORATION
DATA REDUCTION
DATA INTEGRATION
DATAACCESSENDPOINTS
STANDARDS
REGISTRY
STANDARDTEMPLATES
STANDARDVOCABULARIES
ETLTOOLS
i2b2
model
OMOP
model
DATA CUSTODIAN ENVIRONMENT
DATA ACCESS ANDSOURCES DATA PLATFORM
DATA ANALYTICS
UNISR/ St Raphaele
UTA
Lund University
Goeteborgs University
DATA SOURCES DATA TYPES DATA HARMONISATION DATA ANALYSIS PIONEER’S OUTCOMES
IMPLEMENTOMICS
Consensusonthemost
importantPcaoutcomes
Identificationofcritical
evidencegapsinPCa
Standardisationof
outcomedefinitionand
outcomemeasures
Newinsightson
improvedstratification
Improvedstandardized
carepathwayswith
knownbetter
predictableoutcomes
ERSPC group
Is big data the answer? Our model
7. Key elements to success of our model
• Access to large numbers of clinical data sets across the
different stages of Prostate Cancer and across different
European and non-European geographies.
• A standardised, open-access, collaborative data platform.
• Novel informatics and computational approaches.
• Involvement of key opinion leaders in Prostate Cancer
and Big Data.
• Involvement of all relevant stakeholders including
patients.
8. Encouraging social action
• Encouraging widespread adoption of data donation faces
significant hurdles due to understandable concerns
around privacy and security.
• It is possible to anonymise data but many of the most
transformative uses of big data in healthcare don’t allow
for the data to be anonymised
e.g. there’s no point in identifying the perfect clinical trial
participant if you can’t contact them.
9. Encouraging social action
• Key to overcoming the hurdle of patient consent is
education (opt-out rather than opt-in).
• The big data industry must engage with patients and
explain the benefits to them and to others of sharing
their data.
Take home message:
Our data can drive healthcare innovation and save
lives, but we must be willing to share it.
As more targeted and novel therapies are being developed there is an urgent need to identify and validate endpoints which may shorten the length of clinical trials or guide patient stratification
As more targeted and novel therapies are being developed there is an urgent need to identify and validate endpoints which may shorten the length of clinical trials or guide patient stratification
As more targeted and novel therapies are being developed there is an urgent need to identify and validate endpoints which may shorten the length of clinical trials or guide patient stratification
As more targeted and novel therapies are being developed there is an urgent need to identify and validate endpoints which may shorten the length of clinical trials or guide patient stratification
As more targeted and novel therapies are being developed there is an urgent need to identify and validate endpoints which may shorten the length of clinical trials or guide patient stratification