WHO policy on the sharing and reuse of health-related data for research purposes and guidance on the implementation

of the policy Section title

Sharing and reuse of health-related

data for research purposes: WHO
policy and implementation guidance

i
WHO policy on the sharing and reuse of health-related data for research purposes and guidance on the implementation of the policy

ii
Section title

Sharing and reuse of health-related

data for research purposes: WHO
policy and implementation guidance

i
WHO policy on the sharing and reuse of health-related data for research purposes and guidance on the implementation of the policy

Sharing and reuse of health-related data for research purposes: Sales, rights and licensing. To purchase WHO publications,
WHO policy and implementation guidance see http://apps.who.int/bookorders. To submit requests for
commercial use and queries on rights and licensing, see
ISBN 978-92-4-004496-8 (electronic version) https://www.who.int/copyright.
ISBN 978-92-4-004497-5 (print version)
Third-party materials. If you wish to reuse material from this
© World Health Organization 2022 work that is attributed to a third party, such as tables, figures
or images, it is your responsibility to determine whether
Some rights reserved. This work is available under the Creative permission is needed for that reuse and to obtain permission
Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence from the copyright holder. The risk of claims resulting from
(CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by- infringement of any third-party-owned component in the work
nc-sa/3.0/igo). rests solely with the user.

Under the terms of this licence, you may copy, redistribute and General disclaimers. The designations employed and the
adapt the work for non-commercial purposes, provided the presentation of the material in this publication do not imply
work is appropriately cited, as indicated below. In any use of the expression of any opinion whatsoever on the part of WHO
this work, there should be no suggestion that WHO endorses concerning the legal status of any country, territory, city or
any specific organization, products or services. The use of the area or of its authorities, or concerning the delimitation of
WHO logo is not permitted. If you adapt the work, then you its frontiers or boundaries. Dotted and dashed lines on maps
must license your work under the same or equivalent Creative represent approximate border lines for which there may not yet
Commons licence. If you create a translation of this work, you be full agreement.
should add the following disclaimer along with the suggested
citation: “This translation was not created by the World Health The mention of specific companies or of certain
Organization (WHO). WHO is not responsible for the content or manufacturers’ products does not imply that they are
accuracy of this translation. The original English edition shall be endorsed or recommended by WHO in preference to others of
the binding and authentic edition”. a similar nature that are not mentioned. Errors and omissions
excepted, the names of proprietary products are distinguished
Any mediation relating to disputes arising under the licence by initial capital letters.
shall be conducted in accordance with the mediation rules of the
World Intellectual Property Organization (http://www.wipo.int/ All reasonable precautions have been taken by WHO to verify
amc/en/mediation/rules/). the information contained in this publication. However, the
published material is being distributed without warranty of
Suggested citation. Sharing and reuse of health-related data for any kind, either expressed or implied. The responsibility for the
research purposes: WHO policy and implementation guidance. interpretation and use of the material lies with the reader. In no
Geneva: World Health Organization; 2022. Licence: CC BY-NC-SA event shall WHO be liable for damages arising from its use.
3.0 IGO.
Cover photo credit: WHO/Hery Razafindralambo
Cataloguing-in-Publication (CIP) data. CIP data are available at Design and layout by ACW.
http://apps.who.int/iris.
ii
Section title

Contents
Foreword iv

Acknowledgements v

World Health Organization policy on the sharing and reuse

of health-related data for research purposes 1
Vision 1
Purpose and scope 1
Objectives for data-sharing for research purposes 1

Guidance for implementation of the WHO policy on the

sharing and reuse of health-related data for research purposes 3
Who is this guidance for? 3
Aim of the guidance document 3
What data are covered? 3
How to use this guide 5
Where can I get help and advice? 5
Making a data management and sharing plan 5
Why? 5
The type of data will influence the data-sharing plan required 5
What information does the data set describe? 6
Which data can be shared openly? 7
Which licence can be used to share data? 7
Preparing the data for sharing 7
Choosing a data repository 8
What if it is not possible to share the data set openly? 8
Consent and data sharing for health-related purposes. 8
Anonymization and de-identification 10
Choosing an appropriate repository 10
Clinical trials 13
Sharing individual participant data (IPD) in research 13
Equity in data sharing and capacity-building 13
Which Creative Commons licence should be used for data
that are made openly available? 13
Describing the data set with metadata 14
Writing a Data Availability Statement 14

References 15
Annex 1. DRAFT template for creating a data management and sharing plan 16
Annex 2. Visual guide to practical data de-identification 17

iii
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

Foreword
When data related to research activities are shared ethically, equitably and
efficiently, there are major gains for science and public health. However, the
extent of sharing of health research data sets is lower than is needed to unlock
these gains. This is why we are prioritizing the development of practical guidance
to assist the many World Health Organization (WHO) colleagues who manage
health research to enable greater sharing of research data. WHO staff members
may be responsible for research projects funded or sponsored by WHO, or they
may be involved in research in other ways – such as by coordinating networks of
researchers or providing technical support to ministry of health staff who manage
or fund research programmes.

As I write this Foreword we have seen the problems caused by the lack of data-
sharing on COVID-19. Many thousands of COVID-19 clinical trials have taken
place, but most have been too small or inadequately designed to provide
useful evidence for policy-makers. A more mature ecosystem for curating
and aggregating data sets would have been extremely helpful in collating the
evidence base for decision-making. It is therefore very timely for WHO to publish
this guidance which provides practical assistance to our staff in how to ensure
that that final data from health research with which WHO is associated are
further shared for reuse. WHO believes that the global scientific community
should embrace the norm of sharing of data sets at the time of publication and,
in some cases (such as pathogen genomic data), even prior to publication.

In parallel, the legitimate equity and career development needs of researchers

around the world need to be addressed through initiatives that ensure access to
medical products that are developed and to capacity development, including in
data management and analysis. WHO’s view is that the sharing of health data is
a global public good. Countries that share data should be applauded, and never
penalized, for contributing to that public good.

Dr Soumya Swaminathan
WHO Chief Scientist

iv
Section title

Acknowledgements
This guide for WHO staff was developed by a WHO Science Division working
group chaired by Vasee Moorthy, Senior Advisor, Research for Health.

Robert Terry was the lead writer. Main technical contributors were: Lisa Askie,
Draurio Barreira Cravo Neto, Craig Burgess, Sarah Charnaud, Ian Coltart, Janet Diaz,
Nathan Ford, Lisa Haintz-Carbonin, Ghassan Karam, Katherine Littler, Fuad Mirzayev,
John Reeder, Andreas Reis, David Schellenberg, Anthony Solomon, Soe Soe Thwin
and Sachiyo Yoshida.

WHO thank the following external expert reviewers: Philippe Guerin,

Georgina Humphreys, Steve Kern, Rebecca Li, Rebecca Lawrence, Laura Merson,
Sally Rumsey, and Tim Smith.

v
WHO policy on the sharing and reuse of health-related data for research purposes and guidance on the implementation of the policy

vi
World Health Organization
policy on the sharing and
reuse of health-related data
for research purposes

Vision
Great advances in science and public health can be achieved through the appro-
priate sharing and reuse of health data, permitting analyses that: 1) allow for the
fullest possible understanding of health challenges; 2) help develop new solutions;
and 3) ensure that decisions are based on the best available evidence. Data, and
the knowledge derived from the use of that data, should be recognized as a global
public good, and data-sharing and data reuse should be maximized in ways that
are effective, ethical and equitable in order to improve public health.

Purpose and scope

The purpose of this document is to clarify the policy and practice on the reuse and
onward sharing for research purposes of health data collected under the auspices
of WHO technical programmes. This covers use in both emergency and non-emer-
gency situations and complements the following: the Policy on use and sharing of
data collected in Member States by the World Health Organization (WHO) outside
the context of public health emergencies; the Policy statement on data sharing by
the World Health Organization in the context of public health emergencies; and
the Joint statement on public disclosure of results from clinical trials. This policy
covers the reuse of health data for research purposes. Its scope includes research
data generated by research undertaken directly by WHO, or funded by WHO, as
well as the use of other health data for research purposes.

This document sets out the objectives of this WHO policy and the obligations of
WHO staff and researchers funded by WHO. The following section of this docu-
ment entitled Guidance on the implementation of the WHO policy on the sharing
and reuse of health-related data for research purposes provides further referenc-
es and resources to assist in the development of a data management and shar-
ing plan that is in alignment with the vision of this policy.

Objectives for data-sharing for research purposes

The sharing of data collected under the auspices of WHO technical programmes
must be undertaken in ways that are:
• Equitable – any approach to the sharing of data should recognize and balance
the needs of: participants and researchers who generate and use data; other
analysts who might want to reuse those data; and those communities who ex-
pect health benefits to arise from research.
• Ethical – all data sharing should balance and protect the privacy of individu-
als and the dignity of communities while acknowledging the imperative to im-
prove public health through the most productive use of data.
• Efficient – any approach to data sharing should be aimed at enhancing/opti-
mizing the quality and value of the use of those data and enabling their con-
tribution to improving public health. Data sharing should be done as promptly
and in as open a manner as possible, building on existing norms, policies and
practices and reducing unnecessary duplication and competition.
• FAIR – WHO encourages data management and sharing to follow the FAIR principles
(Box 1).
1
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

World Health Organization In summary, following these principles should ensure that the data are:
policy on the sharing and • Findable – manual or machine searchers can easily locate data using the
Internet and a unique and persistent identifier.
reuse of health-related
• Accessible – once users find the required data, they need to know how those
data for research purposes data can be accessed, possibly including authentication and authorization.
• Interoperable – which means that, where possible, the data are stored in the
simplest nonproprietary software format (e.g. spreadsheet data stored as a
CSV file rather than in a commercial software such as Excel).
• Reusable – in addition to the above, all other information or software required
to access and use the data are provided, as are rich metadata (summary)
describing clearly what the data contain and how they are organized under a
clear and accessible data usage licence.

In order to address the above principles, WHO staff and researchers supported
by WHO should develop a data management and sharing plan for each data set
for which the Organization has responsibility in order to ensure the following:
• Data sets that contain no participant information, or which have been an-
onymized, should be deposited in an appropriate data repository with a per-
sistent identifier, such as a Digital Object Identifier (DOI), and made available
under the Creative Commons Attribution 4.0 International (CC BY 4.0) licence,
which clearly describes the terms of reuse.
• These data sets, described by accurate metadata, must include a description
with links to the underlying data, individual participant data or extended data,
and to any relevant materials or software necessary to understand, assess and
replicate the research.
• The same conditions apply in the case of research articles that are funded in
whole or in part by WHO. These must include a Data Availability Statement
with links to underlying data or extended data and any relevant materials or
software necessary to understand, assess and replicate the research.
• In cases where data cannot be made publicly available for ethical, legal and/or
confidentiality reasons, a metadata record should be created in an appropriate
data repository with a persistent identifier, such as a DOI. The Data Availability
Statement should indicate the restrictions, the process for applying for access
to the data, and the conditions that will apply for reuse.
• The chosen repository, hosted either by WHO or by a third party, should be
able to demonstrate in a transparent manner how its governance arrange-
ments meet any technical, ethical and legal requirements that may be applica-
ble to it, including any national requirements as appropriate.
• Chosen repositories should maintain appropriate user agreements that gov-
ern the sharing and contribution of data stored by them. Data use agreements
should address needs for sharing and reuse in line with this policy.

A data management and sharing plan must be developed as part of any techni-
cal programme or research project in order to guide the data collection, cura-
tion, storage, access, analysis, archiving and, in rare cases, disposal throughout
the project or research cycle.

2
Guidance for implementation
Section title

of the WHO policy on the sharing

and reuse of health-related data
for research purposes

Who is this guidance for?

The guidance is intended for WHO staff and other researchers who are funded in
whole or in part by WHO.

Aim of the guidance document

The aim of this document is to provide guidance on how to develop a data man-
agement and sharing plan in order to share a WHO data set in a digital format in
a way that is in line with WHO’s policy on the sharing and reuse of health-related
data for research purposes (the “Policy” or the “WHO Policy”).

What data are covered?

The Policy covers both emergency and non-emergency situations and comple-
ments the following from the perspective of reuse: the Policy statement on data
sharing by the World Health Organization in the context of public health emergen-
cies; (1) the Policy on use and sharing of data collected in Member States by the
World Health Organization (WHO) outside the context of public health emergen-
cies; (2) and the Joint statement on public disclosure of results from clinical trials.
(3) The scope of the Policy includes research data generated by research under-
taken directly by WHO or funded by WHO, as well as the use of other health data
for research purposes.

This guide is a summary of advice, providing references and links to more de-
tailed resources. In many areas, there are no best practice recommendations
or accepted standards. In such cases, the practices that WHO considers to be
good practice are set out in the Policy.1 Consequently, users will need to exercise
judgement, taking into account all relevant considerations, in order to identify
the most appropriate mechanism for storing and sharing specific data set/s. For
instance, sharing pathogen genomic data requires different considerations than
sharing individual patient data (IPD) that underlies a clinical trial.2

It is recommended that users of this guide develop a data management and

sharing plan (see Annex 1) in line with the principles and objectives of the Poli-
cy. The plan should meet the WHO principles of being effective, ethical and equi-
table – as articulated in the Policy. The plan should also follow the technical FAIR
principles as far as possible and should aim to make the data as open for reuse
as possible while balancing the necessary ethical and legal factors that protect
the interests of the stakeholders related to the data set.

1 This guide may contain links to third party resources or external websites. WHO is not responsi-
ble for the accuracy or content of any third party resources or external links. Reference to any third
party resource or external link does not imply that the resource or link, or their author or entity, is
endorsed or recommended by WHO. These references are provided for convenience only.

2 in this document we use IPD to mean individual participant data to cover all instances of WHO
research that involve people and their personal data. Please note in clinical research IPD is more
typically defined as individual patient data.
3
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

BOX 1. The FAIR Guiding Principles

To be findable:
F1 (meta)data are assigned a globally unique and persistent identifier
F2 data are described with rich metadata (defined by R1 below)
F3 metadata clearly and explicitly include the identifier to the data it describes
F4 (meta)data are registered or indexed in a searchable resource.

To be accessible:
A1.0 (meta)data are retrievable by their identifier using a standardized communications protocol
A1.1 the protocol is open, free and universally implementable
A1.2 the protocol allows for an authentication and authorization procedure, where necessary
A2.0 metadata are accessible, even when the data are no longer available.

To be interoperable:
I1 (meta)data use a formal, accessible, shared, and broadly applicable language for knowledge
representation
I2 (meta)data use vocabularies that follow FAIR principles
I3 (meta)data include qualified references to other (meta)data.

To be reusable:
R1.0 (meta)data are richly described with a plurality of accurate and relevant attributes
R1.1 (meta)data are released with a clear and accessible data usage license
R1.2 (meta)data are associated with detailed provenance
R1.3 (meta)data meet domain-relevant community standards.

This approach usually means choosing an appropriate third party repository

and curating the data so that they are: 1) well described using a clear summary
(metadata); 2) structured so they can be reused and aggregated with similar data
sets; and 3) where necessary, anonymized to reduce the possibility of individu-
al participants being identified. The chosen repository must be assessed to en-
sure that it meets the principles outlined in this document – particularly where
the data are sensitive (e.g. in the case of IPD). There are some instances where
WHO may directly manage a repository within its own resources; WHO-funded
researchers should investigate whether their university or host institute has es-
tablished its own data repository or set up an institutional subscription with a
data-sharing organization.

4
How to use this guide
The recommendation is that a data management and sharing plan should be de-
veloped before any data collection has begun. Nevertheless, there are still many
practical solutions to sharing data in an ethical and effective way, even after a pro-
gramme of activity has completed. WHO encourages users to read the various sec-
tions of this guide and follow up on those resources that are most relevant to their
situation. It is recommended that the template in Annex 1 be used to draft a data
management and sharing plan. WHO-funded researchers may choose to discuss
this with the WHO technical programme that supported their work. WHO will un-
dertake periodic reviews of these plans in order to further improve this guidance,
particularly where standards and best practices are developed.

Where can I get help and advice?

Help and advice are available through the Science Division. Please contact Rob
Terry for technical advice, Ian Coltart for licensing/publication requests, and Va-
see Moorthy for questions on genomic data sharing. Where it is unclear who to
contact, these colleagues can triage requests to the correct staff member.

Making a data management and sharing plan

Why?
Any proposal to share data requires careful consideration of the relevant techni-
cal, legal and ethical issues. Users need to understand the data that are availa-
ble and how they can be used. To this end, a structured description of the data
set, known as the metadata, as well as a Data Availability Statement, are needed.
The benefits of developing a data-sharing plan include the following:

• A data-sharing plan demonstrates good research practice and marks the com-
pletion of the research cycle.
• Data sets can be used and cited by others, providing mechanisms of credit for
the work involved in creating the data set.3
• Data sets can be aggregated into other studies, can be reproduced and can ena-
ble the creation of new knowledge for the benefit of science and public health.
• The plan ensures appropriate protection and privacy of participants where relevant.
• The plan ensures preservation and long-term archiving of the data.
• A data-sharing plan satisfies the requirements of many journals and funders,
including WHO.

The type of data will influence the data-sharing plan required

Understanding the data set will help decide how data sharing can be achieved.
Many universities provide advice through their library services, and a number
of journals and data organizations have published guidance. The simplest entry
guides on understanding the sharing of research data have recently been pub-
lished by the F1000Research publisher (2021). This describes what open data
is, how and where to choose a repository, how to structure data (e.g. in spread-
sheets), applying the FAIR principles, and how to address legal and ethical issues
such as achieving anonymization. (4)

3 A 2019 study estimated an increase of 25% in citations where data were made open. See: Fane B,
Ayris P, Hahnel M, Hrynaszkiewicz I, Baynes G, Farrell E. The state of open data report 2019. Digital
Science. 2019. https://doi.org/10.6084/m9.figshare.9980783.v2. 5
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

A comprehensive resource for managing clinical health data for research purposes
has recently been published by the Global Health Network that created a knowl-
edge hub for the European and Developing Countries Clinical Trials Partnership
(EDCTP) which was launched in 2021. (5) This resource sets out the knowledge
management steps throughout the research cycle, explains how to develop a data
management and sharing plan, and provides a repository finder tool and other re-
sources. In addition, the Global Health Network offers free online training courses
in data management and sharing, as well as training on general research skills
– with certification upon completion. (6,7) These courses have been designed to
suit clinical research and researchers in low- and middle-income settings. The
knowledge hub also has links to many other resources. (8)

What information does the data set describe?

The main concerns when planning to share health data relate to the history and
provenance of the data. They include:
• Do the data contain any information that could directly identify an individual
(e.g. name, telephone number, email) or indirectly identify a person by linking
a detailed geographical location with a disease?
• Has participant or patient consent been given for reuse and what specifically
does the consent allow in terms of reuse?

Where data are collected under a research protocol (e.g. clinical trial) the con-
sent obtained should allow for publication of the analysis and deposition in a re-
pository of the underlying data reported in a study. This type of consent will not
cover the sharing of individual participant data unless it has been adequately
anonymized (see below). One prerequisite for further sharing of data is that the
data-sharing agreement includes appropriate legal provisions governing receipt
of data by WHO or the WHO-funded researcher.

For WHO, any processing of personal data4 should be done in compliance with
WHO’s Personal Data Protection Policy. The “Research Data” section of the Per-
sonal Data Protection Policy applies in cases where the direct purpose of the
processing of personal data is scientific research.

Sharing identifiable participant (or patient) data (IPD) requires more planning
but can be safely achieved to answer extremely important health questions to
which answers may not be possible from single studies. An example would be to
identify groups that are under-represented in a number of research studies in or-
der to understand, for example, contraindications in medicines with respect to
children or pregnant women. This may also include analysing electronic patient
records using data collected as part of service delivery but that do not have a
standing consent, provided that the data provider has secured all necessary per-
missions according to the applicable national laws of the country where the data
were collected. Several solutions are available to enable sharing through man-
aged access repositories. Further information is provided below.
Users also need to understand what formats their data are in – for instance,
4 Personal data: Any information relating to an individual who is or can be identified from that in-
formation. Personal data include: biographical data (biodata) such as name, sex, civil status, date
and place of birth, country of origin, country of residence, individual registration number, occu-
pation, religion and ethnicity; biometric data such as a photograph, fingerprint, facial or iris im-
age; as well as any expression of opinion about the individual, such as assessments of status and/
or specific needs. Processing of personal data: Any operation, or set of operations, automated or
not, which is performed on personal data, including but not limited to the data collection, record-
ing, organization, structuring, storage, adaption or alteration, retrieval, consultation, use, trans-
fer (whether in computerized, oral or written form), dissemination or otherwise making available,
correction or destruction.
6
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

images (e.g. MRI scans), audiovisual formats (e.g. interviews), Office software
(e.g. Excel), computer software programmes, a database (e.g. REDCap), DNA se-
quence information etc. Many different repositories exist to handle these types
of data and it is likely that, with some research, a suitable resource will be found.

Which data can be shared openly?

In most cases, data that contain no participant information (e.g. a survey of in-
sect vectors) or that have been anonymized can be shared by deposition in an
appropriate repository that complies with the WHO Policy. This guide will help
you decide which repository is best suited to your data.

Which licence can be used to share data?

The WHO Policy requires that WHO’s data are shared under a Creative Commons
Attribution 4.0 International (CC BY 4.0) licence. This open licence means users
must acknowledge the use of the data in any subsequent publication. The data
set must be cited – e.g. using the DOI assigned to the data set when deposited.
However, this type of licence means that those responsible for sharing the data
do not need to be contacted or involved in the reuse. This is analogous to some-
one citing a research paper authored by someone else. This makes the data
sharing more efficient and increases its reusability. The option of applying this li-
cence is usually available when depositing the data.

Preparing the data for sharing

The data-sharing toolkit of the EDCTP Knowledge Hub describes the major steps
in preparing data for sharing. (9) These steps have been adapted and are sum-
marized below. They should form the basis of the Data Management and Sharing
Plan. The EDCTP toolkit provides further detailed advice for each step.

1. Choose a suitable repository and set up an account.

2. Understand the terms and conditions of the Data Deposit Licence. In cas-
es where WHO wishes to conclude an agreement with a third party reposi-
tory, the agreement must be reviewed and cleared by the WHO Legal Office.
WHO-funded researchers are responsible for the data they collect and for
concluding their own agreement with the repository.
3. Clarify ownership, rights and permissions. Ensure that rights have been ob-
tained to share the data – and, if appropriate, that existing consent covers
the data sharing and reuse intended – and define any conditions that apply.
Ensure that the Data Deposit Licence does not conflict with WHO Policy by
placing additional restrictions on further reuse.
4. Structure and organize the data in files that are logical and clearly named.
5. Check that the data are accurate and consistently labelled. Some disciplines,
particularly in chemistry and the omics, have a recommended structure for
reporting and depositing in a repository. Consider also supplying non-propri-
etary files – e.g. CSV files as well as structured data sheets such as Excel. This
maximizes reuse (i.e. users do not require MS Office to use the files) and long-
term preservation (i.e. when proprietary file formats become obsolete). (10)
6. Prepare supporting documents, user guides and any tools (e.g. software)
needed to understand and use the data and deposit them along with the
data. This may require that a data dictionary is developed to define each of
the variables in the data set.
7. Deposit the data and obtain a DOI that can be used to cite the data set in any
reports or journal articles as part of the Data Availability Statement.

7
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

Some of these steps are expanded below with specific reference to the require-
ments of the WHO Policy.

Please note: This is only a short and introductory summary. Data depositors should
ensure that they have access to the required expertise in making a final decision.

Choosing a data repository

Table 1 provides a simple decision tree as a guide to choosing a suitable data
repository. The links give access to online search tools. However, the list is not
exhaustive, so it is advised to consult the resources available in the list of refer-
ences at the end of this document in order to make an informed choice. Where
specific platforms are referenced, these are provided for information and should
not be considered an endorsement or recommendation by WHO. Researchers
funded by WHO should check if their university or host institution has its own
data repository or an account with a general data repository that is compliant
with WHO’s policy. Most repositories will require registration. Some repositories
offer a free service up to a specified data limit and then a subscription service
beyond that; others charge for all deposits. Where a charge to deposit applies, it
is worth investigating if there are waivers or discounts for researchers from low-
and middle-income countries (LMICs).

What if it is not possible to share the data set openly?

In certain cases it may not be possible to deposit the data set in a repository. This
might be for legal reasons (e.g. where some of the data were obtained from a pro-
prietary source) or for ethical reasons when consent or appropriate permission
cannot be obtained for sharing data (e.g. electronic records) from a Member State.

In these cases, two options are available, although neither is ideal and they
should be considered only as a last resort. One option is to retain the data within
WHO or at the institution of the WHO-funded researcher’s institute. If a publica-
tion is made with these data, a Data Availability Statement should be included
with the publication.

A second option is possible where it is not planned to use the data in a publica-
tion. In this case, the data depositor should prepare a description of the data
set – known as the metadata – and deposit that description (but not any relat-
ed data files) in a repository. This will also generate a specific web address (DOI)
that can be used to cite the summary of the data easily. Within the metadata the
depositor can also include instructions on how others can make contact and dis-
cuss if there is a possibility to access or use those data. This is analogous to the
Data Availability Statement in a publication. It will be desirable to have the nec-
essary procedures in place to handle a request to access and use these sensitive
data sets, including a sound rationale for refusal if necessary. There remains an
obligation to ensure that the data – e.g. clinical trial data – are stored and ar-
chived properly for future reference. While not ideal, the creation of a summary
at least makes available an inventory of WHO data. Please seek advice if you are
unsure how to proceed.

Consent and data sharing for health-related purposes.

There is no single approach to obtaining consent for research purposes. WHO
has worked with the Council for International Organizations of Medical Sciences
(CIOMS) to produce ethical guidelines for health-related research involving hu-
mans. (11) These guidelines are the basis for creating sound consent processes,
and a number of templates are used by WHO to match different contexts. With
8
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

respect to data sharing, CIOMS Guideline 12 on Collection, storage and use of

data in health-related research states that:

“Researchers, sponsors and research ethics committees must share data for fur-
ther research where possible.”

Guideline 12 also sets out the conditions where a research ethics committee can
grant a waiver for informed consent on the sharing of stored data where no indi-
vidual consent exists:

“When researchers seek to use stored data collected for past research, clinical or
other purposes without having obtained informed consent for their future use
for research, the research ethics committee may consider to waive the require-
ment of individual informed consent if:
1. the research would not be feasible or practicable to carry out without the
waiver; and
2. the research has important social value; and
3. the research poses no more than minimal risks to participants or to the
group to which the participant belongs.” (12)

Further conditions for data sharing are spelled out in Guideline 24 on Public ac-
countability for health-related research. Guideline 24 states that:

“Researchers should prospectively register their studies, publish the results and
share the data on which these results are based in a timely manner. Negative
and inconclusive as well as positive results of all studies should be published or
otherwise be made publicly available.”

There is a following commentary on this guideline which states:

“There are compelling reasons to share the data of health-related research. Re-
sponsible sharing of clinical trial data serves the public interest by strengthening
the science that is the foundation of safe and effective clinical care and public
health practice. Sharing also fosters sound regulatory decisions, generates new
research hypotheses, and increases the scientific knowledge gained from the
contributions of clinical trial participants, the efforts of clinical trial researchers,
and the resources of clinical trial funders …”

“… The risks of data sharing may be mitigated by controlling with whom the
data are shared and under what conditions, without compromising the scientif-
ic usefulness of the shared data. Organizations that share data should employ
data use agreements, observe additional privacy protections beyond de-identifi-
cation and data security, as appropriate, and appoint an independent panel that
includes members of the public to review data requests. These safeguards must
not unduly impede access to data.” (13)

In summary, these guidelines support and encourage data sharing when the ap-
propriate governance processes are in place to protect individual privacy. The CI-
OMS guidelines underpin many Good Clinical Practice documents, including the
guidance provided by WHO. (14)

In drafting a consent form that adequately covers the secondary use of the data, (15)
the United Kingdom’s Data Service has produced useful guidance, particularly
with respect to language to be avoided. It advises that:
9
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

“Consent forms should not preclude sharing of research data. So, promises to
destroy any data or that data will only be seen or accessed by the research team
should be avoided.”

“Terms such as ‘fully anonymous’ or ‘strictly confidential’ should be avoided,

as they are often impossible to define. Better is to indicate how data will be an-
onymized (e.g. by removing all personal information that could directly identify
an individual) and that whilst data will be made available to other researchers,
confidentiality will be protected.” (16)

Anonymization and de-identification

Data variables are categorized as direct and indirect identifiers. Direct identifi-
ers link directly to a person’s identity without further information (e.g. name,
address, medical record number, photograph, etc). Indirect identifiers require
some deduction to identify the person and include, for instance, a higher level of
address, disease condition, gender, generic job title, date of birth. By combining
these variables, there is a reasonable probability that an individual can be iden-
tified, (e.g. a 50-year-old man with diabetes on a certain street in Geneva).

“Anonymization” is the process by which those variables that can identify an individu-
al are irreversibly removed from a data set. “De-identification” is the process by which
these identifiers are masked in some way or replaced by a code that has a key. Ap-
plying the key can reinsert the identifiers to enable re-identification. As such, control
of the key is limited and its use allowed only in certain circumstances. De-identifica-
tion through an encrypted code and key is a common approach used in clinical trials.

There are no global standards for anonymization or de-identification. Many

countries have adopted data protection laws describing what constitutes anon-
ymous data. It is important to be aware of any applicable legal frameworks, in-
cluding in the place of collection and storage of the data. (17)

Many data managers within the research community use the standards recom-
mended in the United States of America or in Europe, depending on their respec-
tive places of business and applicable legal frameworks. (18, 19)

In addition, the Personal Data Protection Commission of Singapore and the Austral-
ian National Data Service (ANDS) have both produced useful guides. (20, 21) An info-
graphic from ANDS is reproduced in Annex 2. Calculation of the risk of identification
is itself an imprecise measure and there are alternative ways to measure it. (22)

WHO-funded researchers should seek advice from within their university or

host institution and should contact WHO if they require further assistance.

Choosing an appropriate repository

Repositories can be organized in a number of ways. The simplest repositories
are often free to use up to a certain data limit with deposition and access via a
website without any mediation or review by committee. Consequently, they can
be described as open repositories. There is a structure for describing the data
with a summary – the metadata – but there is no requirement to format the data
to a certain standard. Open repositories are sometimes called data lakes, as the
data are all in one place but in variety of formats with different standards. Reuse
conditions are set by the depositor from “open”, under a Creative Commons li-
cence, to “closed” whereby only the metadata are available. Examples of open
10 repositories include Figshare, Mendeley and Zenodo.
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

Other repositories may include more structure for deposition, with certain con-
ditions for access – such as a requirement to structure the data according to a
certain standard or a restriction on the republishing of those data elsewhere. For
some repositories, particularly those dealing with more complex and sensitive
data such as IPD, the steps for depositing and accessing the data are moderated.
These managed resources may include reviews by data access committees, data
transfer agreements and some continued involvement of the depositor in the type
of reuse. There may be a requirement to deposit data in a certain standard struc-
ture or, alternatively, curation to a standard structure may be provided by the re-
pository. There are also different ways in which the data may be stored, either to-
gether in a centralized location, or kept in different places but accessed through a
specialized website – sometimes known as a federated system. There are numer-
ous variations of these repositories, each with its own benefits and drawbacks.

If the data being shared are relatively simple and there are no concerns about pri-
vacy the open repositories offer a straightforward solution. The data are securely
archived, free to access and the depositor has no need to be involved in further re-
search. The route to access and reuse of the data is quick and simple. However, as
there are no requirements for specific standards, persons accessing the data will
have to undertake their own curation when combining data from different sources.

Managed platforms require more resources and the process of accessing the
data may be lengthy, requiring review from a data access committee and the
resolution of questions regarding the protocol. However, if numerous data from
many sources (e.g. electronic patient records from many different hospitals in
different countries) are combined, the fact that the data are in a single standard
(e.g. C-DISC for participant/patient data) means that the analysis can proceed
more quickly and the utility of the whole data set is greatly enhanced.

The simple decision tree in Table 1 can guide users into making the first level
of decision about where to deposit their data. Again, if you require assistance
in making this choice, please contact: [email protected], [email protected] or
[email protected] .

11
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

TABLE 1. Decision table – how to choose the right repository

Does the data contain

identifiable patient Suggested action
information?

Can the data be effectively

NO YES anonymized by the data depositor?
Can it be shared without further
managed access?

Choose a structured
discipline-specific repository
NO YES
(e.g. EDCTP Knowledge Hub,
Repository Finder, re3data.org
Fairsharing.org/databases Vivli)
OR
Choose a general data repository
(e.g. Mendeley Data, Zenodo, Figshare)
OR
Choose an institutional data repository
if compliant with WHO Policy
(e.g. it issues a DOI).

Identify a suitable managed

NO
Managed access repository. Search EDCTP
access Knowledge Hub, Repository Finder
required (e.g. IDDO, TB PACTS).

NO 1. Create a metadata record

Data cannot (summary) to obtain a DOI in
be shared for any suitable repository.
ethical, legal 2. If the data are part of a publication,
or technical explain the rationale in a Data
reasons Availability Statement.

Choose a discipline-specific
repository. Search EDCTP
Knowledge Hub, Repository Finder,
re3data.org Fairsharing.org/databases
OR
Choose a general data repository
(e.g. Mendeley Data, Zenodo, Figshare)
OR
Choose an institutional data
repository if compliant with WHO
Policy (e.g. it issues a DOI).

12
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

Clinical trials
A number of established good practice arrangements exist for clinical trials and
their reporting of the underlying data. WHO has issued a Joint statement on
public disclosure of results from clinical trials.3 The statement covers the regis-
tration of a trial prior to commencement, the use of the trial ID in all subsequent
communications, the timely reporting of results within 12 months from prima-
ry study completion (the last visit of the last subject for collection of data on
the primary outcome) in a trial registry and no longer than 24 months after trial
completion in a journal publication.

Sharing individual participant data (IPD) in research

IPD can be shared in an appropriate public repository, as listed in the reposito-
ry finder tools highlighted in Table 1, provided that the data are adequately an-
onymized. These data sets can be used in systematic reviews to aggregate re-
ported results into a finding with a greater degree of confidence. As noted above,
any processing of personal data by WHO should be done in compliance with the
WHO Personal Data Protection Policy.

Equity in data sharing and capacity-building

A fundamental objective of the WHO Policy is that data are shared in ways that are
ethical, efficient and equitable. For WHO, equitable data sharing means balancing
the needs of all stakeholders – the data depositors, those generating the data (in-
cluding researchers in the case of research data sets), the secondary users of data
sets, and the communities from where the data have originated. WHO considers it
to be crucial to address the needs of data depositors and explores every avenue to
advance equity and capacity-building, including within the area of data sharing.

With regard to research data sets, many researchers in LMICs consider that data
sharing policies – particularly those of international funders of research – may
disadvantage them if the WHO Policy requires data to be shared within short
timescales or in ways that do not involve the researchers in the analysis. WHO
will promote data agreements – for instance, through advocacy with funders of
international research –that include involvement of researchers from the coun-
tries where data are collected. All subsequent publications should reflect those
contributions in the authorship.

Capacity-building is closely linked to achieving equitable approaches to data

sharing. WHO seeks to build research capacity through its technical programmes
that support research and training courses in LMICs, as well as other support.
This includes efforts to provide training in data management and curation as
well as the skills required to manage and analyse health-related data.

Which Creative Commons licence should be used for data

that are made openly available?
The data should be shared under a Creative Commons Attribution 4.0 Interna-
tional (CC BY 4.0) licence. This means the data are free to use without seeking
further permission from the depositor but the users must provide a proper attri-
bution or citation as to where they obtained the data from. This is why obtaining
a DOI is useful as it provides a precise and permanent location for the data set so
that the depositor can receive citation credit.

13
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

Describing the data set with metadata

Metadata is “data about the data” which describe the properties of the data set.
Metadata should be structured so that it can be machine-read (e.g. by an Inter-
net search engine) to maximize findability. (23) The elements covered by a meta-
data statement are often prescribed by the repository used and commonly in-
clude the following:

• Title: This is often the title of the research project or, if the data are from a techni-
cal programme, a description of that programme and the type of data included.
• Owner: A contact email and an address should be provided for the data crea-
tor analogous to the corresponding author on a research paper. If the data be-
long to WHO, then WHO might be added as the publisher of the data.
• Contributors: Analogous to co-authors in research manuscripts, the term
“contributor” describes who else should be credited with creating the data set.
• Subject: These are keywords and some repositories will provide drop-down
menus (e.g. MeSH). The subject may include when, where and how the data
were collected and the type of data (e.g. trial data, household survey).
• Description: As in an abstract on a research paper, the is a brief description of
the data. This may include information describing the structure and titles of
the data files.
• Licence: The WHO Policy requires Creative Commons Attribution 4.0 Interna-
tional (CC BY 4.0).
• Date: The date when the metadata were created, plus any modifications. This
is often automatically added by the repository.
• Format: The standard by which a file is encoded. A free standard is pre-
ferred. However, if a proprietary standard is used, copies should be supplied
in non-proprietary files where possible (e.g. alongside common files such as
REDCap or Excel).
• Related publication: A url or DOI of any related publications and journal arti-
cles should be included.
• Language: The language(s) in which the data are presented.

Writing a Data Availability Statement

This is increasingly required as part of the submission process for publishing in a
journal and is recommended by the International Committee of Medical Journal
Editors (ICMJE). If the data set is deposited in a repository, the following head-
ings should be included:
• repository name
• title of the data set
• DOI
• list of all data items (including the full file name, and a description of its contents)
• data licence
• any restrictions on access to the data.

N.B. Where data are NOT shared through a repository, a statement to this effect is
required and should include details of who may be contacted – and how – by in-
terested parties wishing to discuss access to the data.

14
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

References
1 Policy statement on data sharing by the World Health Organization in the context of public health emergencies. Geneva: World
Health Organization; 2016 (https://www.who.int/ihr/procedures/SPG_data_sharing.pdf, accessed 29 January 2022).

2 Policy on use and sharing of data collected in Member States by the World Health Organization (WHO) outside the context of public
health emergencies. Geneva: World Health Organization; 2017 (https://www.who.int/publishing/datapolicy/Policy_data_sharing_
non_emergency_final.pdf, accessed 29 January 2022).

3 Joint statement on public disclosure of results from clinical trials. Geneva: World Health Organization; 2017 (https://www.who.int/
news/item/18-05-2017-joint-statement-on-registration, accessed 29 January 2022).

4 Understanding open data. London: F1000 Research Ltd; 2021 (https://think.f1000research.com/open-data/, accessed 28 January 2022).
5 EDCTP Knowledge Hub. European and Developing Countries Clinical Trials Partnership Data management portal (https://edctp-
knowledgehub.tghn.org/Dat-man-por/, accessed 28 January 2022).

6 EDCTP e-Learning courses. European and Developing Countries Clinical Trials Partnership Data management portal (https://edctp-
knowledgehub.tghn.org/training-courses/e-learning-courses/, accessed 28 January 2022).

7 Global Health Training Centre (https://globalhealthtrainingcentre.tghn.org/, accessed 28 January 2022).

8 Free tools and resources. European and Developing Countries Clinical Trials Partnership Data management portal (https://edctp-
knowledgehub.tghn.org/data-sharing-toolkit/collated-external-resources/, accessed 28 January 2022).

9 EDCTP Data sharing steps. European and Developing Countries Clinical Trials Partnership Data management portal (https://edctp-
knowledgehub.tghn.org/data-sharing-toolkit/data-sharing/, accessed 29 January 2022).

10 EDCTP Formats list. European and Developing Countries Clinical Trials Partnership Data management portal (https://edctpknowl-
edgehub.tghn.org/data-sharing-toolkit/data-sharing/formats-list/, accessed 29 January 2022).

11 International Ethical Guidelines for Health-related Research Involving Humans, fourth edition. Geneva: Council for International Or-
ganizations of Medical Sciences (CIOMS); 2016 (https://cioms.ch/wp-content/uploads/2017/01/WEB-CIOMS-EthicalGuidelines.pdf ,
accessed 29 January 2022).

12 International Ethical Guidelines for Health-related Research Involving Humans, fourth edition, Guideline 12. Geneva: Council for In-
ternational Organizations of Medical Sciences (CIOMS); 2016:47–52 (https://cioms.ch/wp-content/uploads/2017/01/WEB-CIOMS-Eth-
icalGuidelines.pdf, accessed 29 January 2022).

13 International Ethical Guidelines for Health-related Research Involving Humans, fourth edition, Guideline 24. Geneva: Council for In-
ternational Organizations of Medical Sciences (CIOMS); 2016:91–93 (https://cioms.ch/wp-content/uploads/2017/01/WEB-CIOMS-Eth-
icalGuidelines.pdf, accessed 29 January 2022).

14 Handbook for good clinical research practice (GCP): guidance for implementation. Geneva: World Health Organization; 2005 (https://
apps.who.int/iris/handle/10665/43392, accessed 29 January 2022).

15 Research data management. Colchester: United Kingdom Data Service, University of Essex (https://www.ukdataservice.ac.uk/man-
age-data/legal-ethical/consent-data-sharing.aspx, accessed 29 January 2022).

16 Documenting consent. Colchester: United Kingdom Data Service, University of Essex (https://ukdataservice.ac.uk/app/uploads/doc-
umenting-consent.pdf, accessed 29 January 2022).

17 Greenleaf G. Global tables of data privacy laws and bills, fifth edition. 145 Privacy Laws & Business International Report. 2017;14–26
(https://ssrn.com/abstract=2992986, accessed 29 January 2022).

18 Guidance regarding methods for de-identification of protected health information in accordance with the Health Insurance Portabil-
ity and Accountability Act (HIPAA) Privacy Rule. Washington (DC): US Department of Health & Human Services (https://www.hhs.gov/
hipaa/for-professionals/privacy/special-topics/de-identification/index.html#standard, accessed 29 January 2022).

19 The European Medicines Agency (EMA) approach to anonymization draws most of its reference from the US Institute of Medicine re-
port: Sharing clinical trial data: maximizing benefits, minimizing risk. Washington (DC): National Academies Press; 2015 (https://pu-
bmed.ncbi.nlm.nih.gov/25590113/, accessed 29 January 2022).

20 Guide to basic data anonymisation techniques. Singapore: Personal Data Protection Commission; 2018 (https://iapp.org/media/pdf/
resource_center/Guide_to_Anonymisation.pdf , accessed 29 January 2022).

21 De-identifying your data. Canberra: Australian National Data Service; 2018 (http://www.ands.org.au/working-with-data/sensitive-da-
ta/de-identifying-data, accessed 29 January 2022).

22 El Emam K, Abdallah K. De-identifying clinical trials data. Cranbury (NJ): Applied Clinical Trials; 2015 (https://www.appliedclinicaltri-
alsonline.com/view/de-identifying-clinical-trials-data, accessed 29 January 2022).

23 Describe: metadata and documentation. European and Developing Countries Clinical Trials Partnership Data management portal
(https://edctpknowledgehub.tghn.org/data-sharing-toolkit/data-management-sub/Describe-Metadata/, accessed 31 January 2022).

15
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

ANNEX 1. DRAFT template for creating a data management

and sharing plan
In preparing a data management and data sharing plan, WHO-funded researchers should seek advice
from within their university or host institution. Use the template below as a checklist to develop your
data management and sharing plan.

Topic Action Notes

What type of data?

Who are the stakeholders?

What rights and permissions

– do you have?
– do you need to obtain?

Choosing a repository Third party or WHO?

Will you need to sign a data deposit For WHO, review and clearance by WHO
agreement? Legal office

Choosing NOT to share Create a metadata record Do you have a process in place to deal
with enquiries for access?

Where appropriate, what consent exists?

Do you need to apply for a waiver of
requirement for consent?

If necessary, what level of anonymization or

de-identification do you need to apply?

Clinical trials Register Decide when and how you will report

Sharing IPD:
– open repository?
– managed repository?
– WHO repository?

Equity How can equity in data sharing, access

and/or capacity-building be ensured?

Timelines for sharing

Choosing the right licence WHO Policy requires Creative Commons

Attribution 4.0 International (CC BY 4.0)
licence

Create your metadata

Write a Data Availability Statement

Source: Wilkinson MD, Dumontier M, Aalbersberg IJ, Appleton G, Axton M, Baak A et al. Comment: The FAIR Guiding Principles for scien-
tific data management and stewardship. Nature: Sci Data. 2016;3(160018) (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4792175/
pdf/sdata201618.pdf, accessed 31 January 2022).

16
Guidance for implementation of the WHO policy on the sharing and reuse of health-related data for research purposes

ANNEX 2. Visual guide to practical data de-identification (24)

What do scientists, regulators and lawyers mean when they talk about de-identification? How does anonymous
data differ from pseudonymous or de-identified information? Data identifiability is not binary. Data lies on a
spectrum with multiple shades of identifiability. This is a primer on how to distinguish different categories of data.

De-identified Anonymous
Degrees of Pseudonymous data data
identifiability data Direct and known
Direct and indirect
identifiers have
indirect identifiers
Information containing direct Information from which direct identifiers been removed or
have been removed or
and indirect identifiers have been eliminated or transformed, manipulated together
manipulated to break
but indirect identifiers remain intact. with mathematical and
the linkage to real
technical guarantees to
world identities.
prevent re-identification.

Explicitly Potentially Not readily Key Pseudony- Protected De- Protected Anonymous Aggregated
personal identifiable identifiable coded mous pseudony- Identified de-Identi- anonymous
mous fied

Direct identifiers
Data that identifies
a person without
additional information
or by linking to
information in the Intact Partially Partially Eliminated or Eliminated or Eliminated or Eliminated or Eliminated or Eliminated or Eliminated or
public domain masked masked transformed transformed transformed transformed transformed transformed transformed
(e.g., name, SSN)

Indirect identifiers
Data that identifies an
individual indirectly.
Helps connect pieces
of information until
an individual can be Intact Intact Intact Intact Intact Intact Eliminated or Eliminated or Eliminated or Eliminated or
singled out (e.g., transformed transformed transformed transformed
DOB, gender)

Safeguards
and controls
Technical,
organizational
and legal controls
preventing employees, Not relevant Limited Controls Controls Limited Controls Limited Controls Not relevant Not relevant
researchers or other Due to nature or none in place in place or none in place or none in place Due to nature due to high
of data in place in place in place of data degree of data
third parties from re- aggregation
identifying individuals

Selected Name, ad-

dress, phone
Unique device
ID, license plate,
Same as Po-
tentially Iden-
Clinical or
research
Unique, arti-
ficial pseudo-
Same as Pseu-
donymous,
Data are
suppressed,
Same as
De-Identified,
For example,
noise is cal-
Very highly
aggregated
examples number, SSN, medical record tifiable except datasets nyms replace except data are generalized, except data are ibrated to a data (e.g., sta-
government-is- number, cookie, data are also where only direct identifiers also protected perturbed, also protected data set to hide tistical data,
sued ID (e.g., IP address (e.g., protected by curator retains (e.g., HIPAA Lim- by safeguards swapped, etc. by safeguards whether an census data, or
Jane Smith, MAC address safeguards and key (e.g., ited Datasets, and controls (e.g., GPA: 3.2 and controls individual is population data
123 Main 68:A8:6D: 35:65 controls (e.g., Jane Smith, John Doe = 5L7T = 3.0-3.5, gen- present or not that 52.6% of
Street, 555- :03) hashed MAC ad- diabetes, HgB LX619Z) (unique der: female = (differential Washington,
555-5555) dresses and le- 15.1 g/dl = sequence not gender: male) privacy) DC residents
gal representa- Csrk123) used anywhere are women)
tions) else)

24 This infographic is adapted from the Australian National Data Service document a Visual-Guide-to-Practical-Data-DeID.pdf (2018) (CC-BY-ND 4.0)
http://www.ands.org.au/working-with-data/sensitive-data/de-identifying-data. 17
Sharing and reuse of health-related data for research purposes: WHO policy and implementation guidance

Notes

18
ycilop eht fo noitatnemelpmi eht no ecnadiug dna sesoprup hcraeser rof atad detaler-htlaeh fo esuer dna gnirahs eht no ycilop OHW
WHO policy on the sharing and reuse of health-related data for research purposes and guidance on the implementation of the policy

World Health Organization

20, Avenue Appia
1211 Geneva 27
Switzerland

www.who.int

20