Slide template for use in presentation - PowerPoint PPT Presentation

About This Presentation
Title:

Slide template for use in presentation

Description:

Catalogue Biolink NL & Catalogue 2.0 – PowerPoint PPT presentation

Number of Views:38
Avg rating:3.0/5.0
Slides: 12
Provided by: AlexClu5
Category:

less

Transcript and Presenter's Notes

Title: Slide template for use in presentation


1
Catalogue Biolink NL Catalogue 2.0
2
Deliverables of Biolink NL
  • Improve the access to biomaterials and data by
    setting up
  • A national infrastructure for linkage of biobanks
    to registries (including technical, governance,
    and ethical/legal aspects).
  • A web based Portal including procedures for
    linkage and a central catalogue of linked
    biobanks/registries that can be searched for
    specific patient populations. The catalogue
    should contain minimal harmonized dataset and
    also informed consent information.

3
What level of catalogue do you mean?
Biolink NL
  • Level 1 List of Biobank Cohorts
  • available biobank sample collections with a
    minimal set of properties such as size, data
    topics and material types.
  • Level 2 Data item dictionary
  • information on data items available in each
    cohort such as questionnaire questions, lab
    measurements, diagnoses or GWAS
  • Level 3 Group level information (chunked)
  • aggregate information such as number of
    available samples for high blood pressure cases.
  • Level 4 Individual level data per Item
  • complete set of (anonymised) data in a biobank
    for which most biobanks will have in-house
    database.
  • Level 5 Identifiable data per Individual
  • services to link to data from other sources and
    registries such as the national cancer registry,
    through pseudonimization service.

Biolink NL
Biolink NL
4
Which catalogues and why
5
Catalogue level 5
feasibilities
Privacy
Patient/data selection
Flexible/ usability
.....
Legal demands
Valid results
Preconditions
Requirements
.....
Data linkage
Demands Bio- databanks!
Catalogue Subject level
Semantic integration
Up-to-data
6
What will the catalogues look like?
Level Catalogue Content External access Privacy demands Gover-nance Frequent updates Remarks
2 Contents/linkage possibilities Y 0 ?
3 Key numbers Y Fed by catalogue V
5 Linked subject level data N Few variables per subject 2) Version 1 Mondriaan infrastructure
7
Getting started with RP Catalogue 2.0
  • Who does what
  • Programmer (Bart, 1.0 fte)
  • Coordinator and biobank liaison (David, 0.5 fte)
  • Standardization/bouwstenen, workflow, power user
    (Linda, 0.3 fte)
  • RP5 / Mondriaan integration and policies (Willem,
    0.2 fte)
  • Integration PSI, CTMM, ....
  • What data to use
  • What biobanks are willing to share data in public
    domain
  • Ask biobanks that are in BBMRI-NL catalogue
    (David/Morris)
  • Ask studies in CTMM (Linda/David)
  • Ask PSI board (Linda, ...)
  • Other sources? (e.g. BBMRI-SE)
  • What data items to share

8
Eerste concrete acties
  • Aansluit team (David Willem)
  • Willem neemt contact op met biobanks
  • Assistance in selectie en mapping van data items
  • Procedure voor level 2, 3, 4 en 5 data overdracht
  • Technisch faciliteren van de import (techneut
    naar techneut)
  • User interactie ontwerp team (Morris Chantal?)
  • Alternatieve ontwerpen (wizard vs filtering)
  • Toegangsbeperkingen en anonimiteit
  • Harmonisatie team (Linda)
  • Afspraken maken over data items en hun codering /
    formats
  • Mapping van variabelen (protocollen en features)
    tussen bronnen
  • Governance sustainability team (Willem)
  • Business case voor catalogus
  • Wat voor informatie mag wel/niet gedeeld worden
  • Afspraken over uitwisseling / samenwerking
  • Open source? Project place (NFU sponsor?)

9
Interesse
  • Pilot projecten
  • Inladen van Mondriaan level 3
  • Inladen van CTMM level 3

10
Deelname aan level 3 en 5
Biobank
Periodieke uitspoel van gepseudonimiseerde
microdata
Level 5 catalogus
aggregatie
Level 3 catalogus
11
Deelname aan alleen level 3
Biobank
aggregatie
DB
Level 3 catalogus
12
Uitdagingen
  • Wie beheert de catalogus
  • Samenwerking en inzet middelen Biolink NL in
    Catalogus 2.0
  • Inrichten efficiënte aansluit procedure
  • Kennis opbouw data
  • Impact minimalisatie voor biobanken
  • Langere termijn
  • Sustainability
  • Samenwerking inbedding NFU/BBMRI

13
Privacy level 5
Compliant with e.g. WBP and case law CBP
14
Governance result of 2 years of Mondriaan
experience catalogue 3 and 5
  • Sources set the limits
  • Scientific use only
  • Independent, non-profit governance
  • Different sources have different limits
  • Sometimes level V is less scary than level III
  • Control ownership
  • Whats in it for me?
  • Catalogue design consequences
  • Level 3 catalogue level vs source level
  • Level 5
  • No public access,
  • No release of data without approval by a source
  • Release of answers.
  • Industry

15
Mondriaan catalogue (level 5)
  • In production September 2010
  • Access approx 10 pers. (UMCU/UMCG/RUG)
  • Semantic integration
  • Compliant with privacy laws
  • Established procedures on data-handling and
    ownership
  • Data available in catalogue
  • GP data 0,5 M subjects (Utrecht Almere)
  • EPIC- NL (40K subject)
  • IADB (0,5 M subjects Groningen)
  • Psychiatric data provence of Utrecht

16
Mondriaan catalogue (level 5)
  • Data expected 2012-13
  • SFK public pharmacy data of 14 M subjects
  • Achmea (5 M subjects)
  • UPOD
  • Smart
  • Vumc GP network
  • LifeLines
  • LINH

17
Mondriaan Meta-catalogue
Researcher
Mondriaan
City/ region
Catalogue
Data exchange HDR-Mondriaan
Gronin- gen
Data delivery
Sneek
Zwolle
Almere
VUmc
Twente
Leiden
EPIC
Nijm- egen
SFK
Zeeland
Ehoven
Ehoven
S- Limburg
Researcher
18
Proces data pseudonimisatie/ anonimisatie
data-upload naar Mondriaan
De TTP verzorgt de pseudonimisatie en koppeling
van data
Bron- pseudoniem
2) De TTP stuurt bronpseudoniemen naar de
client. Koppelingen zijn nog niet mogelijk met
deze pseudoniemen. Dit i.v.m. privacy.
Versleutelde NAW
1) NAW-gegevens worden versleuteld bij de
bron en naar TTP gestuurd. De medische data
worden niet naar de TTP gestuurd. .
DB
DB
Gepseudonimiseerde (medische) data van de bron
3) De client stuurt alleen de bronpseudoniemen
Medische data naar Mondriaan
19
Catalogue on patient level be careful!
20
Koppelen
  • Custodix
  • Kanskoppeling
  • 3 stappen
  • normalisatie inputgegevens hashing
  • Blocking
  • Matching
  • Matching combinatie van fellegi-sunter
    algoritme(s) en scriptable matchers
  • Matches non-matches
  • Calibratie
  • ZorgTTP
  • Wel normalisatie hashing inputgegevens
  • Geen koppeling, maar uniform versleutelde
    identifiers of combinaties daarvan
Write a Comment
User Comments (0)
About PowerShow.com