Estonian approach - from biobanking to precision medicine, Andres Metspalu, Director, Estonian Genome Center, (Estonia), GetPersonalized! summit in Helsinki on 25 May 2015

1. Estonian approach - from
biobanking to precision medicine
Andres Metspalu
Professor & Director
Estonian Genome Center (Estonia)

2. Estonian approach:
from biobanking to precision
medicine
Andres Metspalu, M.D., Ph.D
Estonian Genome Center
University of Tartu
Helsinki
May 25th, 2015
GetPersonalized! Summit in Helsinki

3. 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.

4. For personalized medicine we’ll need:
E-infrastructure (e-health, digital prescriptions, digital signature,
electronic health records etc.)
Public support
Primary care providers support
Population based biobank
DNA analysis technology
Databases on genotype phenotype associations

5. 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)

6. § 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.

7. -0.05
-0.03
-0.01
0.01
0.03
0.05
0.07
-0.05 -0.03 -0.01 0.01 0.03 0.05
P21(4.68%)
PC1 (8.65%)
Austria Bulgaria Czech Republic
Estonia Finland (Helsinki) Finland (Kuusamo)
France Northern Germany Southern Germany
Hungary Northern Italy Southern Italy
Latvia Lithuania Poland
Russia Spain Sweden
Switzerland
-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
PC223.8%
PC1 36.6%
Europe
CHB
JPT
YRI
CEU

8. Uuringubaas: TÜ Eesti Geenivaramu geenidoonorite kohort
ca 52000 isikut

9. Public opinion and awareness of the EGCUT 2001-
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%
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
In favor of
the idea of
EGCUT
Never
heard of
EGCUT
Against it
Annely AllikKellele:
Kuupäev:
Tingimused:
Tähtaeg:
Kliendi nr.:
7 päeva
29.06.2010
013
SUMMAARVE SISU
24 900,00Uuring: Eesti elanike arvamus Geenivaramu projektist
Leping nr. C-201-10, 24.05.2010.a.
Tartu Ülikooli Eesti Geenivaramu
Tiigi 61b
Tartu 50410
Maksja:
22.06.2010
2902552Arve nr.
Kaidi KandlaEsindaja:
24 900,001
Hind
tk.
Ühik Kogus
EEK
SUMMA
EUR
1 591,40

10. 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”

11. e-Health

12. 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

13. PHARMACIS AND
FAMILY DOCTORS
2009
X-Roads, ID-card, State IS Service Register
HEALTHCAREBOARD
-Healthcareproviders
-Healthprofessionals
-Dispensingchemists
STATEAGENCYOFMEDICINES
-CodingCentre
-Handlersofmedicines
POPULATIONREGISTER
PHARMACIS
2010january
BUSINESSREGISTER
HOSPITALS
2009
FAMILYDOCTORS
2009
SCHOOLNURSES
2010september
EMERGENCYMEDICALSERVICE
2013
NATIONAL
HEALTH
INFORMATION
SYSTEM
2008 december
PRESCRIPTION
CENTRE
2010 january
PATIENT PORTAL
2009
XROADS GATEWAY
SERVICE
2009
Architectural “Big picture”
From Artur Novek (eHealth)

14. Documents total – 10.8 mio
1.1 mio persons medical data (growth 20% during 2012)
National Patient Portal

15. 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 DB
1 mio person have documents (82% of
population)

16. Do we have enough information to
start?

17. 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

18. Example of an individual disease trajectory
18
Birth
B01 Varicella [chickenpox]
B05 Measles
B26 Mumps
L23.9 Allergic contact
dermatitis, unspecified cause
K02 Dental caries
A37 Whooping cough
J18.0 Bronchopneumonia,
unspecified
B00 Herpesviral [herpes
simplex] infections
J30 Vasomotor and
allergic rhinitis
B35 Dermatophytosis
J01 Acute sinusitis
J31 Chronic rhinitis,
nasopharyngitis and
pharyngitis
F32.1 Moderate depressive
episode
G44.2 Tension-type
headacheI10 Essential (primary)
hypertension
J20 Acute
bronchitis
.
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
Birth
J01
M54
G50.0
S51.0
J35.0
J38.4
H65.1
I21.9
K29.9
R49.0
F32.01
F43.22
D23.6
G44.0
J31.0
.
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
Recruitment
Birth
F43.22
K29.9
.
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
Male, born 1970, joined 2007
Smoker, BMI=27.5, low physical activity
Health
Insurance Fund
NE Med
Center

19. 19
.
J01 Acute
sinusitis
M54 Dorsalgia
G50.0 Trigeminal
neuralgia
S51.0 Open wound
of elbow
J35.0 Chronic tonsillitis
J38.4 Oedema of larynx
H65.1 Otitis media, acute
and subacute allergic
I21.9 Acute myocardial
infarction, unspecified
F43.22 Mixed anxiety and
depressive reaction
K29.9 Gastroduodenitis,
unspecified
R49.0 Dysphonia
F32.01 Depressive episode
with somatic syndrome
F43.22 Mixed anxiety and
depressive reaction
K29.9 Gastroduodenitis,
unspecified
D23.6 Benign
neoplasms of skin
G44.0 Cluster headache
syndrome
J31.0 Chronic
rhinitis
.
2007 2008 2009 2010 2011 2012 2013

20. 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“)

21. Polygenic risk scores
Calculated 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
Frequency
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
050010001500

22. 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 risk
Above-average (50-90%) genetic risk
Highest 10% of the genetic risk
24 27 5
80
134
52
119
198
84
116
154
47
(n=1810) (n=2020) (n=1337) (n=685)

23. 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 8.3 10
8
81 cases
p 1.9 10
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)

24. ROC curves (BMI=25..35)

25. Model-based predictions T2D

26. Coronary Artery Disease
Males Females
CAD prevalence in individuals aged 40-75
0%
5%
10%
15%
20%
25%
Below-average genetic risk
Above-average (50-90%) gen. risk
Highest decile of the gen. risk
146
144
52
81
97
34

27. Incident Myocardial Infarction
A weaker,
but still
significant
effect seen
in women:
p=0.005

28. 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)

29. Virtuous Cycle of
Clinical Decision Support
Measure
Guideline
CDS
Practice
Registry
http://www2.eerp.usp.br/Nepien/DisponibilizarArquivos/tomada_de_decis%C3%A3o.pdf

30. 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 hospital
Commercialization
- IPR
- Business agreements
The Estonian Program for
Personal Medicine I

31. 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.

32. *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

34. 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.

35. 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

36. 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.

37. Thank you!
www.biobank.ee

38. 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
NDHRD
2/2014
Pop
reg
3/2014
HIF
6/2014
Cancer
Reg
12/2010
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
Reg
1/2012
Death
Reg
12/2012
Death
Reg
8/2013
Death
Reg
6/2014

39. 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

40. GD in national registries
• Population Registry 51800 GD 99.8%
• Health Insurance Fund 51607 GD 99.5%
• Estonian eHealth 39880 GD 76.9%
• Tartu University Hospital 22492 GD 43.4%
• North Estonia Medical Center 21202 GD 40.9%
• Cancer Registry 2644 GD 5.1%
• Causes of Death Registry 2349 GD 4.5%
• Myocardial Infarction Registry 945 GD 1.8%
• Tuberculosis Registry 260 GD 0.5%
40