Harrison
Thanks for sharing your concerns
I am also a systems engineer doubling as anything that is needed,
now in clinical emergency
I am temporarily bypassing restrictions publishing protocols and hopefully
soon some results here
https://protocols.io/share/DfVgOaP
Here is one of my protocols for EM treatment of covid
https://www.protocols.io/view/emergency-medicine-em-protocol-for-the-all…
I am looking for trial partners with Lab facilities for a clinical study
Let me know if you or others are interested in a collaboration
Dr Paola Di Maio
On Thu, Apr 2, 2020 at 9:01 AM harrison via RDA-COVID19 wrote:

Hello everyone,
Regarding the objectives of the WG, I think we should aim to create the
bases of an open-access health/social data repository.
Something like this is spontaneously happening with all the dataset that
are being uploaded in Github, my personal concern is that GitHub is owned
by Microsoft. In that sense, I think that we should aim to something more
like “arXiv” (htttps://arxiv.org/ ) (“DataRxiv” ??
.org domain is currently owned by amazon), which is hosted/owned by Cornell
University (private institution, but educational).
As it is pointed out in the case statement data uploaded to this
repository should, at least, adhere to FAIR
principles. Besides that, here
are some complementary ideas that came into my mind until now:
1.- Data should be anonymised before uploaded.
2.- A chain of trust to data security responsible (individuals and
institutions) should be indicated. (As pointed-out in Christopher
Harrison’s e-mail)
3.- Extremely sensitive data must be flagged and access to it must be
granted only with permission from the owners. A protocol to speed-up the
procedure permission should be available in emergency situations like
COVID19.
4.- Every time a dataset is uploaded this must be curated by at least two
experts (like peer-review system). The curators should also be indicated in
the dataset information.
5.- Potential bias in the dataset (gender, race, etc.) should be
mentioned in the dataset information. We don’t want to repeat Amazon kinds
of issues about hiring only men…
Regarding point 3, as this kind of data is usually very sensitive, maybe
this repository could host dataset description and information, but
not-necessarily the data itself. This will at least to help know who owns
the data, to whom we need to talk to get it and/or work with, and to
elucidate if this data could help our current research or not. This could
even be helpful to encourage collaborations between
researchers/labs/policy-makers/etc.
Regarding Christopher concerns about IRBs ( “*…An IRB’s goal is to
protect patients privacy and rights and providing a method to review
ethical use a study…*” ), we may even consider ways to make that the
dataset uploaded or pointed in this repository include among their
description/information if they already passed through and IRB (or
something similar) or if they are about to do it, prepared to do it, etc.
I know all this may sound obvious so several of you, I just wanted to
contribute a bit and I think it is worthy to outline it.
I whish you all a enjoyable confinement, if you are confined as me.
Else, keep enjoying life and take care.
Jorge Andrés Clarke De la Cerda
PhD in Applied Sciences
On Thu, Apr 2, 2020 at 7:57 AM paoladimaio via RDA-COVID19 wrote:

Hello everyone,
I would personally disagree with the group starting an actual repository for data, though agree that perhaps a metadata repository would be useful for the research community. Ideally though this metadata should be produced to an agreed standard so that it easier for the research community to find.
Simply I believe that setting up an effective and safe repository for health data would be difficult and presents a great number of challenges. In particular I would disagree with the notion of expecting data depositors to anonymise data as, having worked in a data archive, I have seen that very often data owners do not necessarily know how to do this well. If we wanted to make this a requirement for data to be made available we would need to provide guidance as to a minimum standard of anonymisation, tools and tests that should be done to justify that anonymisation, and probably have well-trained staff to check the data before it is goes public. Allowing the self-deposit of ‘anonymous’ data would be to gamble public confidence in the research community on the capabilities of the least-able data owner that makes their data available. Once one dataset in the repository was shown to be nowhere near anonymous enough, the public reaction would probably lead to the end of the project.
I don’t really foresee how we could put together a GDPR-compliant data resource for sensitive data (remembering that health data is a special category and so would require additional controls) without investment and considerably more time. If we consider a widely-used security model such as the ‘5 Safes’, I think it would be very difficult for us to develop a process in the time we have. 
As we have limited time it may be better for the group to focus on helping data owners produce safe data, guidance for researchers on how to use the data, and bringing metadata together to help researchers locate data that they could potentially request to use.
Kind regards,
Simon

Simon –
I agree, a metadata ‘catalog’ of available assets (datasets, offered data services etc.) is a practical and achievable objective
Best
Keith
——————————————————————————–
Keith G Jeffery Consultants
Prof Keith G Jeffery
E: ***@***.***
T: +44 7768 446088
S: keithgjeffery
———————————————————————————————————————————-
The contents of this email are sent in confidence for the use of the
intended recipient only. If you are not one of the intended
recipients do not take action on it or show it to anyone else, but
return this email to the sender and delete your copy of it.
———————————————————————————————————————————-
– Show quoted text -From: ***@***.***-groups.org
Sent: 02 April 2020 13:43
To: RDA-COVID19
Subject: Re: [rda-covid19] health data and irb’s
Hello everyone,
I would personally disagree with the group starting an actual repository for data, though agree that perhaps a metadata repository would be useful for the research community. Ideally though this metadata should be produced to an agreed standard so that it easier for the research community to find.
Simply I believe that setting up an effective and safe repository for health data would be difficult and presents a great number of challenges. In particular I would disagree with the notion of expecting data depositors to anonymise data as, having worked in a data archive, I have seen that very often data owners do not necessarily know how to do this well. If we wanted to make this a requirement for data to be made available we would need to provide guidance as to a minimum standard of anonymisation, tools and tests that should be done to justify that anonymisation, and probably have well-trained staff to check the data before it is goes public. Allowing the self-deposit of ‘anonymous’ data would be to gamble public confidence in the research community on the capabilities of the least-able data owner that makes their data available. Once one dataset in the repository was shown to be nowhere near anonymous enough, the public reaction would probably lead to the end of the project.
I don’t really foresee how we could put together a GDPR-compliant data resource for sensitive data (remembering that health data is a special category and so would require additional controls) without investment and considerably more time. If we consider a widely-used security model such as the ‘5 Safes’, I think it would be very difficult for us to develop a process in the time we have.
As we have limited time it may be better for the group to focus on helping data owners produce safe data, guidance for researchers on how to use the data, and bringing metadata together to help researchers locate data that they could potentially request to use.
Kind regards,
Simon
—
Full post: https://www.rd-alliance.org/group/rda-covid19/post/health-data-and-irbs
Manage my subscriptions: https://www.rd-alliance.org/mailinglist
Stop emails for this post: https://www.rd-alliance.org/mailinglist/unsubscribe/68906

I fully agree with Simon. A data repo will be hard to realize. A metadata repo might be do-able, sufficiently challenging already, but of great potential, provided that it is well-maintained and where possible and relevant links to the distributed repositories.
One thing needed is to be clear about what we mean with metadata.
E.g., references to repositories? Data about repositories? Aggregate data? Anything else?
Best, Ronald

Regarding a metadata repository, please check out https://www.wikidata.org/wiki/Wikidata:WikiProject_COVID-19 it might already fit the purpose and
will greatly benefit from contributions by this group.
Kind regards
Dieter

Hi Simon
The group will focus on generating guidelines so researchers in each field are aware of the appropriate standards and repositories. We don’t intend to establish a data repository.
There are some initiatives the EC or its projects are supporting. OpenAIRE for example is creating a dashboard to increase discovery of COVID19 research and EMBL/EBI are rapidly deploying a COVID19 research data platform that should be integrated with EOSC to facilitate data sharing.
All best
Sarah
– Show quoted text -From: ***@***.***-groups.org on behalf of SimonParker via RDA-COVID19
Sent: 02 April 2020 13:42
To: RDA-COVID19
Subject: Re: [rda-covid19] health data and irb’s
Hello everyone,
I would personally disagree with the group starting an actual repository for data, though agree that perhaps a metadata repository would be useful for the research community. Ideally though this metadata should be produced to an agreed standard so that it easier for the research community to find.
Simply I believe that setting up an effective and safe repository for health data would be difficult and presents a great number of challenges. In particular I would disagree with the notion of expecting data depositors to anonymise data as, having worked in a data archive, I have seen that very often data owners do not necessarily know how to do this well. If we wanted to make this a requirement for data to be made available we would need to provide guidance as to a minimum standard of anonymisation, tools and tests that should be done to justify that anonymisation, and probably have well-trained staff to check the data before it is goes public. Allowing the self-deposit of ‘anonymous’ data would be to gamble public confidence in the research community on the capabilities of the least-able data owner that makes their data available. Once one dataset in the repository was shown to be nowhere near anonymous enough, the public reaction would probably lead to the end of the project.
I don’t really foresee how we could put together a GDPR-compliant data resource for sensitive data (remembering that health data is a special category and so would require additional controls) without investment and considerably more time. If we consider a widely-used security model such as the ‘5 Safes’, I think it would be very difficult for us to develop a process in the time we have.
As we have limited time it may be better for the group to focus on helping data owners produce safe data, guidance for researchers on how to use the data, and bringing metadata together to help researchers locate data that they could potentially request to use.
Kind regards,
Simon
—
Full post: https://www.rd-alliance.org/group/rda-covid19/post/health-data-and-irbs
Manage my subscriptions: https://www.rd-alliance.org/mailinglist
Stop emails for this post: https://www.rd-alliance.org/mailinglist/unsubscribe/68906

I fully agree with Ronald and Simon.
And because we must be ambitious, shouldn’t the EOSC in Europe and the NIH Data Commons in US, as well as other major global initiatives, collaborate on this? Everything else is “in theory”…
It’s no longer time for theories…
All the Best.
Carlos.
Carlos Luis Parra Calderón
Consortium Coordinator
FAIR4Health Project
https://www.fair4health.eu/
Head of Technological Innovation
Virgen del Rocío University Hospital
Andalusian Health Service
Head of Research and Innovation in Biomedical Informatics, Biomedical Engineering & Health Economy,
Institute of Biomedicine of Seville
Honorary assistant
School Computer Science
University of Seville
_________________________________________________________________________________________________

Hi Sarah, all,
(apologies for missing the group email off my first message)
Thanks to you and the co-chairs for initiating this activity and reiterating the objectives in your last email. Like other publishers I’m sure, PLOS is receiving COVID related submissions and fast tracking them through peer review and publication (https://collections.plos.org/covid-19). As the goal of this group is to produce guidelines, I am particularly interested in two key questions and am considering how PLOS could help in this regard:
1. Are any group members active researchers who would use the guidelines, and/or do we need to engage candidate researchers as part of our efforts to test them?
2. What are the high value (for reuse) data formats and data types we should be seeking to be prepared/shared according to the guidelines?
The practical guidance and recommendations may be more impactful when developed and applied in the context of worked exemplars/case studies/early-adopters. This approach may also help test/validate our solution (guidelines) against the problem (need for best practice to maximise efficiency).
If anyone has answers or suggestions for these two questions, feel free to email me directly.
Looking forward to contributing to this effort.
Iain
Iain Hrynaszkiewicz
Publisher, Open Research
***@***.***
Main: +44 (0) 1223-442810 I Direct: +44 (0) 1223-446976
Twitter
PLOS
Empowering researchers to transform science
Carlyle House, Carlyle Road, Cambridge CB4 3DN, United Kingdom
Twitter
California (U.S.) corporation #C2354500, based in San Francisco
– Show quoted text -From: ***@***.***-groups.org
Sent: 02 April 2020 13:57
To: RDA-COVID19 ; ***@***.***
Subject: Re: [rda-covid19] health data and irb’s
Hi Simon
The group will focus on generating guidelines so researchers in each field are aware of the appropriate standards and repositories. We don’t intend to establish a data repository.
There are some initiatives the EC or its projects are supporting. OpenAIRE for example is creating a dashboard to increase discovery of COVID19 research and EMBL/EBI are rapidly deploying a COVID19 research data platform that should be integrated with EOSC to facilitate data sharing.
All best
Sarah
________________________________
From: ***@***.***-groups.org on behalf of SimonParker via RDA-COVID19
Sent: 02 April 2020 13:42
To: RDA-COVID19
Subject: Re: [rda-covid19] health data and irb’s
Hello everyone,
I would personally disagree with the group starting an actual repository for data, though agree that perhaps a metadata repository would be useful for the research community. Ideally though this metadata should be produced to an agreed standard so that it easier for the research community to find.
Simply I believe that setting up an effective and safe repository for health data would be difficult and presents a great number of challenges. In particular I would disagree with the notion of expecting data depositors to anonymise data as, having worked in a data archive, I have seen that very often data owners do not necessarily know how to do this well. If we wanted to make this a requirement for data to be made available we would need to provide guidance as to a minimum standard of anonymisation, tools and tests that should be done to justify that anonymisation, and probably have well-trained staff to check the data before it is goes public. Allowing the self-deposit of ‘anonymous’ data would be to gamble public confidence in the research community on the capabilities of the least-able data owner that makes their data available. Once one dataset in the repository was shown to be nowhere near anonymous enough, the public reaction would probably lead to the end of the project.
I don’t really foresee how we could put together a GDPR-compliant data resource for sensitive data (remembering that health data is a special category and so would require additional controls) without investment and considerably more time. If we consider a widely-used security model such as the ‘5 Safes’, I think it would be very difficult for us to develop a process in the time we have.
As we have limited time it may be better for the group to focus on helping data owners produce safe data, guidance for researchers on how to use the data, and bringing metadata together to help researchers locate data that they could potentially request to use.
Kind regards,
Simon
—
Full post: https://www.rd-alliance.org/group/rda-covid19/post/health-data-and-irbs
Manage my subscriptions: https://www.rd-alliance.org/mailinglist
Stop emails for this post: https://www.rd-alliance.org/mailinglist/unsubscribe/68906
Hi Sarah, all,
(apologies for missing the group email off my first message)
Thanks to you and the co-chairs for initiating this activity and reiterating the objectives in your last email. Like other publishers I’m sure, PLOS is receiving COVID related submissions and fast tracking them through peer review and publication (https://collections.plos.org/covid-19). As the goal of this group is to produce guidelines, I am particularly interested in two key questions and am considering how PLOS could help in this regard:
1. Are any group members active researchers who would use the guidelines, and/or do we need to engage candidate researchers as part of our efforts to test them?
2. What are the high value (for reuse) data formats and data types we should be seeking to be prepared/shared according to the guidelines?
The practical guidance and recommendations may be more impactful when developed and applied in the context of worked exemplars/case studies/early-adopters. This approach may also help test/validate our solution (guidelines) against the problem (need for best practice to maximise efficiency).
If anyone has answers or suggestions for these two questions, feel free to email me directly.
Looking forward to contributing to this effort.
Iain
Iain Hrynaszkiewicz
Publisher, Open Research
***@***.***
Main: +44 (0) 1223-442810 I Direct: +44 (0) 1223-446976
Twitter
PLOS
Empowering researchers to transform science
Carlyle House, Carlyle Road, Cambridge CB4 3DN, United Kingdom
Twitter
California (U.S.) corporation #C2354500, based in San Francisco
From: ***@***.***-groups.org
Sent: 02 April 2020 13:57
To: RDA-COVID19 ; ***@***.***
Subject: Re: [rda-covid19] health data and irb’s
Hi Simon
The group will focus on generating guidelines so researchers in each field are aware of the appropriate standards and repositories. We don’t intend to establish a data repository.
There are some initiatives the EC or its projects are supporting. OpenAIRE for example is creating a dashboard to increase discovery of COVID19 research and EMBL/EBI are rapidly deploying a COVID19 research data platform that should be integrated with EOSC to facilitate data sharing.
All best
Sarah
– Show quoted text -From: ***@***.***-groups.org on behalf of SimonParker via RDA-COVID19
Sent: 02 April 2020 13:42
To: RDA-COVID19
Subject: Re: [rda-covid19] health data and irb’s
Hello everyone,
I would personally disagree with the group starting an actual repository for data, though agree that perhaps a metadata repository would be useful for the research community. Ideally though this metadata should be produced to an agreed standard so that it easier for the research community to find.
Simply I believe that setting up an effective and safe repository for health data would be difficult and presents a great number of challenges. In particular I would disagree with the notion of expecting data depositors to anonymise data as, having worked in a data archive, I have seen that very often data owners do not necessarily know how to do this well. If we wanted to make this a requirement for data to be made available we would need to provide guidance as to a minimum standard of anonymisation, tools and tests that should be done to justify that anonymisation, and probably have well-trained staff to check the data before it is goes public. Allowing the self-deposit of ‘anonymous’ data would be to gamble public confidence in the research community on the capabilities of the least-able data owner that makes their data available. Once one dataset in the repository was shown to be nowhere near anonymous enough, the public reaction would probably lead to the end of the project.
I don’t really foresee how we could put together a GDPR-compliant data resource for sensitive data (remembering that health data is a special category and so would require additional controls) without investment and considerably more time. If we consider a widely-used security model such as the ‘5 Safes’, I think it would be very difficult for us to develop a process in the time we have.
As we have limited time it may be better for the group to focus on helping data owners produce safe data, guidance for researchers on how to use the data, and bringing metadata together to help researchers locate data that they could potentially request to use.
Kind regards,
Simon
—
Full post: https://www.rd-alliance.org/group/rda-covid19/post/health-data-and-irbs
Manage my subscriptions: https://www.rd-alliance.org/mailinglist
Stop emails for this post: https://www.rd-alliance.org/mailinglist/unsubscribe/68906

Hello everyone,
Thanks Simon for the precisions, I totally agree with you.
Creating a metadata repository could be “feasible” and without doubts
very useful.
Regarding Ronald concerns, we should agree to a minimum quantity of
information contained in the metadata in order to strongly simplify the
data-search process, ideally, get to know contact coodinates to the data
owner.
Regarding all the currently existent repositories, I think a
repository-map is another feasible idea.
As Carlos Luis said, “this is no longer time for theories…” Here
is a short and interesting
reading, the response from Harry Crane to John Ioannidis’s article “*A
fiasco in the making …” .*I point out this because I think it will be
good to also think in measures to prevent delaying in critical response and
desensitising the public to the real risk we face.
Best
Jorge Andrés Clarke De la Cerda
PhD in Applied Sciences
On Thu, Apr 2, 2020 at 4:16 PM iainh_z via RDA-COVID19 wrote:

All:
Agree that establishing and maintaining catalog of available datasets is
essential. It is important for such resources to be both user(readable)
and machine friendly (as a web service). Registry technology and
standard such as DCAT (https://www.w3.org/TR/vocab-dcat-2/) should be
considered for this. Issues such as access, anonymization, or privacy
should be discussed/addressed separately. A registry or catalog is there
to help discover data and tell you where it is. But access is handled by
the custodian.
best
*P

All –
Presumably we could harvest from metadata catalogs of relevant institutions/universities, convert to – say – DCAT and make available.
Best
Keith
——————————————————————————–
Keith G Jeffery Consultants
Prof Keith G Jeffery
E: ***@***.***
T: +44 7768 446088
S: keithgjeffery
———————————————————————————————————————————-
The contents of this email are sent in confidence for the use of the
intended recipient only. If you are not one of the intended
recipients do not take action on it or show it to anyone else, but
return this email to the sender and delete your copy of it.
———————————————————————————————————————————-
—–Original Message—–
From: ***@***.***-groups.org
Sent: 02 April 2020 16:14
To: ***@***.***-groups.org
Subject: Re: [rda-covid19] health data and irb’s
All:
Agree that establishing and maintaining catalog of available datasets is essential. It is important for such resources to be both user(readable) and machine friendly (as a web service). Registry technology and standard such as DCAT (https://www.w3.org/TR/vocab-dcat-2/) should be considered for this. Issues such as access, anonymization, or privacy should be discussed/addressed separately. A registry or catalog is there to help discover data and tell you where it is. But access is handled by the custodian.
best
*P

I strongly agree with the sentiment of Simon’s comment, and Sarah’s reminder that the aim of the WG is to establish guidelines, not infrastructure.
To those in the WG frustrated by the delays that buracracy cause in sharing and accessing data – I understand this as I dealt with it for over a decade, so I support making the buracratic process as efficient as possible. However this should not be at the stake of the moral and legal principles that underpin the IRB/ethics-review process and researchers should not take it onto themselves to temporarily suspend the legal and ethical requirement to respect individuals’ right to privacy as they see fit. This would not only be damaging for health research everywhere, but would also put those researchers on dangerous footing with their employers and the law.

Hi all,
Adding my voice and thoughts to the mix here.
I agree with the overall sentiment here that this should not be about
setting up a repository. I’m also not convinced that the role of RDA and
this WG is to establish a metadata repository… though I think this would
be hugely valuable, I’m not sure RDA is set up in a way that can create,
and maintain such a capability. I think it maybe better to look for
existing metadata repository providers and looking to them to
expand/re-purpose their capability to provide such a service.
In regards to the purpose of this group and how we can add value here, I’d
like to draw attention to what is different in the current situation to
what RDA does normally. (and forgive me if I’m stating something that is
obvious or have already been recoginsed).
– Normal work we do in RDA is around trying to codify the
complicate data landscape so that we have convergence of thoughts, ideas,
and practices leading to best-practices etc.
– Right now, we are in chaos! The world has been thrown into chaos and
what this group is doing is to help everyone make sense of what is going
on, and provide some framework for this to occur in, at the fastest
possible pace.
I think the Cynefin framework
provides some good
guidance on how to approach the current situation, and our task at hand.
Normally RDA work is primarily around taking what is in the Complicated
domain, and making it Simple. Currently we are very much squarely in the
Chaotic state when it comes to how to respond to COVID-19. Right now, our
role here is to help the world move from the Chaotic domain to the Complex
domain as quickly as possible.
[image: image.png]
Why I think this is important here is because it can change HOW we go about
doing the work here. While we typically deliberate long and hard at RDA on
picking the Best practice/framework from all that is available, what is
required now is to pick/develop some framework quickly, and put it out
there for guidance and use. We should then look at how that has helped, and
continually iterate over this looking at emergent uses.
Also, as RDA members, we have an extensive network for
disseminating information. Looking at the discussions and comments that
are flowing through in the group mail, there are many valuable ideas,
initiatives going on. One thing we could do is find a ways of looking at
is going on across the world that are worthy of adoption/adaption by
others, and disseminating that through our networks for others to build
upon.
Another area that I would like to draw attention to is the work by Dave
Jones at ESIP on Operational Readiness Levels
for data(Slide deck
,
webinar
recording ). This provides a
framework for rapidly evaluating the ‘usable’ value of data in a fast
changing environment. Establishing something akin to this would allow
classification of the growing volume of data, into there fitness for
use/reuse.
[image: image.png]
Kind regards
Kheeran
On Fri, Apr 3, 2020 at 10:03 AM kristan.kang via RDA-COVID19 wrote: