Estonian approach - from biobanking to precision medicine

Andres MetspaluProfessor & DirectorEstonian Genome Center (Estonia)

Estonian approach: from biobanking to precision

medicine

Andres Metspalu, M.D., Ph.DEstonian Genome Center

University of Tartu

HelsinkiMay 25th, 2015

GetPersonalized! Summit in Helsinki

Per Med Definition in EU documents

• Personalised medicine refers to a medical model using molecular profiling for tailoring the right therapeutic strategy for the right person at the right time, and/or to determine the predisposition to disease and/or to deliver timely and targeted prevention

This is the EU H2020 definition of the Per Med.

For personalized medicine we’ll need:

E-infrastructure (e-health, digital prescriptions, digital signature, electronic health records etc.)Public supportPrimary care providers supportPopulation based biobankDNA analysis technologyDatabases on genotype phenotype associations

Estonian Biobank• Estonian Genome Center,

University of Tartu• A prospective, longitudinal,

population-based database with health records and biological materials

• 52,000 participants - 5% of the adult population of Estonia

• Individuals are recruited by GPs, physicians in the hospitals and medical personnel in the EGCUT recruitment offices

• Estonian Human Genes Research Act (HGRA)

§ 3. Chief processor of Gene Bank

• (1) The chief processor of the Gene Bank is the University of Tartu, whose objectives as the chief processor are to:

• 1) promote the development of genetic research;• 2) collect information on the health of the Estonian population and genetic

information concerning the Estonian population; • 3) use the results of genetic research to improve public health.

-0.05

-0.03

-0.01

0.01

0.03

0.05

0.07

-0.05 -0.03 -0.01 0.01 0.03 0.05P2

1 (4

.68%

)

PC1 (8.65%)

Austria Bulgaria Czech RepublicEstonia Finland (Helsinki) Finland (Kuusamo)France Northern Germany Southern GermanyHungary Northern Italy Southern ItalyLatvia Lithuania PolandRussia Spain SwedenSwitzerland

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

-0.12 -0.07 -0.02 0.03

PC2

23.8

%

PC1 36.6%

Europe

CHB

JPT

YRI

CEU

Uuringubaas: TÜ Eesti Geenivaramu geenidoonorite kohort

ca 52000 isikut

Public opinion and awareness of the EGCUT 2001-2014

June, 2001

Sept, 2002

Feb, 2002

March, 2003

March, 2004

Sept, 2004

Dec, 2005

May, 2007

Apr, 2008

July, 2009

June, 2010

April, 2011

July, 2013

Sept, 2014

18%16%

13%

19% 18%16%

22%

33%

28%

36%

59%

55%

67%70%

38% 39% 40%

36%

43%

39%

44%

33% 35%36%

30%33%

27% 27%

4% 4% 3% 4% 4%2% 2% 2% 3% 2% 2% 2% 2% 1%

In favor of the idea of EGCUT

Never heard of EGCUT

Against it

Estonian Government

• 16.12. 2014. the Cabinet approved the Pilot Program of Personalized Medicine”

• Current government has “Personalized Medicine” as one of its activities for the next period

• From 1.03.15. the Ministry of Social Affairs has a new position “Deputy Secretary General for E-services and Innovation”

e-Health

The Estonian ID card

• The ID card is a mandatory ID document for all Estonian residents from the age of 15

• Enables secure digital authentication and signing• A digital signature has the same legal consequences as a hand-written signature• Does not have any additional information

– No bank account, no health information etc.• Active cards: 1 214 428 (31.12.2013)

– Estonian Population 1 286 540 (01.01.2013)– Estonia has been issuing electronic ID cards from January 1st 2002– Also Mobile-ID

PHARMACIS AND FAMILY DOCTORS2009

X-Roads, ID-card, State IS Service Register

HEA

LTH

CAR

E BO

ARD

- Hea

lth c

are

prov

ider

s- H

ealth

pro

fess

iona

ls- D

ispe

nsin

g ch

emis

ts

STAT

E AG

ENCY

OF

MED

ICIN

ES- C

odin

g Ce

ntre

- Han

dler

s of

med

icin

es

POPU

LATI

ON

REG

ISTE

R

PHAR

MAC

IS20

10 ja

nuar

y

BUSI

NES

S RE

GIS

TER

HO

SPIT

ALS

2009

FAM

ILY D

OCT

ORS

2009

SCH

OO

L N

URS

ES20

10 s

epte

mbe

r

EMER

GEN

CY M

EDIC

AL S

ERVI

CE20

13

NATIONAL HEALTH INFORMATION SYSTEM2008 december

PRESCRIPTION CENTRE2010 january

PATIENT PORTAL2009

XROADS GATEWAY SERVICE2009

Architectural “Big picture”

From Artur Novek (eHealth)

Documents total – 10.8 mio

1.1 mio persons medical data (growth 20% during 2012)

National Patient Portal

15

Acceptance• ePrescription covers 94% of issued

prescriptions.• Over 90% of Hospital discharge letters –

digital• Over 97% of stationary case summaries

have sent to the central e-Health DB1 mio person have documents (82% of population)

Do we have enough information to start?

Health Insurance Fund

Estonian Myocardial Infarction Registry

Tuberculosis Registry

Estonian Cancer Registry Estonian Causes of Death Registry

Population Registry

Tartu University Hospital

North Estonia Medical Center National Digital Health Record Database

The broad informed consent.Phenotype data

Phenotype database

Coding center High security area.

Data collector:baseline phenotype data (demographics, health, behavior,...)

From questionnaires to the national registries

17

Example of an individual disease trajectory

18

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

1970

1977

1980

1981

1981

1983

1984

1985

1988

1998

1999

1999

1999

2001

2003

2003

2006

2014

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

1970

2008

2008

2008

2008

2008

2009

2010

2010

2010

2010

2010

20102011

2012

2013

2014

Recruitment

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

1970

2010

2010

2014

Male, born 1970, joined 2007

Smoker, BMI=27.5, low physical activity

Health Insurance Fund

NE Med Center

19

2007 2008 2009 2010 2011 2012 20132007

2008

2008

2008

2008

2008

2009

2010

2010

2010

2010

2010

2010

2010

2010

2011

2012

2013

2013

Polygenic risk scores: questions to address

• Which markers to include?– We use 7500 independent markers for prediction of

genetic risk of Type 2 diabetes and 10200 markers for Coronary Artery Disease

• Optimal weights?- We show that optimal risk scores are obtained with double weights: regression coefficients from a large-scale meta-analysis are additionally weighted, to give more weight to more significant markers (correction for „winners curse“)

Polygenic risk scoresCalculated as S = β1X1 + β2X2 + … + βkXk,

X2,…, Xk - allele dosages for k independent markers (SNP-s),

β1, β2, … , βk – weights

Individuals at high genetic risk

Polygenic risk score for type II diabetes: histogram of the score in 7462 individuals (Estonian Biobank)

Score

Fre

qu

en

cy

3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

05

00

10

00

15

00

Practical usefulness of the polygenic risk score?

BMI <25 BMI 25..30 BMI 30..35 BMI >35

T2D prevalence in individuals aged 45-80

0%

5%

10%

15%

20%

25%Below-average (<50%) genetic riskAbove-average (50-90%) genetic riskHighest 10% of the genetic risk

24 27 580

134

52119

198

84

116

154 47

(n=1810) (n=2020) (n=1337) (n=685)

Incident T2D: analysis of 264 incident cases in overweight individuals free of T2D at recruitment

0 1 2 3 4 5 6

0.00

0.02

0.04

0.06

0.08

0.10

0.12

Follow-up time (years)

Cumulative risk of T2D in 3421 individuals BMI at recruitment >24 kg/m2, age 35-74

Highest genetic risk score quintile (>80%)Average risk score quintiles (20-80%)Lowest risk score quintile (<20%)

p 8.310 8

81 cases

p 1.910 3

153 cases

30 cases

No significant sex difference, genetic risk score is the strongest predictor after BMI

(fasting glycose and insulin measurements are not available for this cohort)

ROC curves (BMI=25..35)

Model-based predictions T2D

Coronary Artery Disease

Males Females

CAD prevalence in individuals aged 40-75

0%

5%

10%

15%

20%

25%Below-average genetic riskAbove-average (50-90%) gen. riskHighest decile of the gen. risk

146

144

52

81

97

34

Incident Myocardial Infarction

A weaker, but still significant effect seen in women: p=0.005

What is Clinical Decision Support Software (CDSS)?

• CDSS provides clinicians with knowledge presented at appropriate times

• It encompasses a variety of tools such as computerized alerts, clinical guidelines, and order sets

• CDSS has the potential provide the necessary level of personalized guidance to providers at the point of care that will be necessary in the era of genomic medicine

• This is a tool to advise PCP (like radiology report)

Virtuous Cycle of Clinical Decision Support

Measure

Guideline

CDS

Practice

Registry

http://www2.eerp.usp.br/Nepien/DisponibilizarArquivos/tomada_de_decis%C3%A3o.pdf

Approved at the Estonian Government Research and Development Council on 17.12.2013.- Health care - Educating health care professionals - Educating the patients - Further development of the eHealth incl. decision support

systems- Research and Development - Sequencing individuals from the Estonian Biobank (ongoing), decision support analysis software, microarray analyze of 50 000 people from the biobank and return the important information to all who want by the PCP or in hospitalCommercialization - IPR - Business agreements

The Estonian Program for Personal Medicine I

Currently the pre-PILOT PROJECT is underway until the end of June 2015. Then the pilot will last up to 2018.

The Estonian Program for Personal Medicine II

The precision medicine method should be seenbasic as a additional “instrument” for physicians in diagnosing, treating and preventing disease, but also for research and clinical investigations, drug trials etc.

*Committee on a Framework for Developing a New Taxonomy of Disease; ‘Towards Precision Medicine’, National Research Council November 2011

Future

Clinical Care

Research Cohort

From Prof. Böttinger

Evidence Generating Medicine

• The next step beyond evidence-based medicine

• The systematic incorporation of research and quality improvement considerations into the organization and practice of healthcare

• To advance biomedical science and thereby improve the health of individuals and populations.

Challenges and issues

• Awareness executives, doctors and patients• New technologies and data empower patient with

more possibilities to manage own health• Ethical issues

– Right to know and right not to know– Treatable and non-treatable conditions– Big data

• Not enough knowledge about associations between DNA variants and diseases

• Large work-load to keep database of known risk markers updated

Conclusions• Estonia has great potential to plan and implement

personalized medicine solutions for the whole country, starting with the pilot project for 50 000 gene donors

– Genetic research with 5% of population genetic and continuously updated phenotype information

– Nation wide Health Information Exchange platform– 10 years of experience of national level e-services (PKI, X-Road,

ID-card, security framework)– High level public trust and acceptance

Estonian government:25. We will establish a course towards personalised medicine based on modern gene technology.

Thank you!www.biobank.ee

Timeline: linking to registries

38

2007 2008 2009 2010 2011 2012 2013 2014

Pop reg

03/2010

Pop Reg

11/2010

Pop reg

10/2011

Pop reg

4/2012

Pop reg

1/2013

NDHRD2/2014

Pop reg

3/2014

HIF6/2014

Cancer Reg12/2010

Cancer Reg

6/2012

Tub Reg

12/2012

University Hospital

8/2013

MI Reg

2/2014

North Estonia Medical Center

5/2014

01/2002 - 31/2010

Baseline

1/2011 - 6/2014

2nd visit

1/2003 - ...

Electronic follow-up

2002

Death Reg

9/2010

Death Reg

5/2011

Death Reg1/2012

Death Reg12/2012

Death Reg

8/2013

Death Reg6/2014

Prevented deaths of CHD in Finland

By treating 2144 persons who have >20% risk when the genetic information is known 135 deaths would be prevented in 10 years!

Tikkanen, E. 2013. Genetic risk profiles for coronary heart disease

GD in national registries

• Population Registry 51800 GD99.8%

• Health Insurance Fund 51607 GD99.5%

• Estonian eHealth 39880 GD76.9%

• Tartu University Hospital 22492 GD43.4%

• North Estonia Medical Center 21202 GD40.9%

• Cancer Registry 2644 GD 5.1%

• Causes of Death Registry 2349 GD 4.5%

• Myocardial Infarction Registry 945 GD1.8%

• Tuberculosis Registry 260 GD0.5%

40