Collecting and preserving health data

Description

Privacy, anonymization, distribution, etc.

Activity M4 held online on November 17 2021, 12:30 to 14:00.

Video

To be released.

Notes

Presentations

• Adama Faye is an engineer interested by data collection
• Pierre Lonchampt is interested in open-source compliance with data-related regulations and standards (e.g. GDPR, HIPAA)
• Vanessa Sekpon is epidemiologist in Benin, collaborating with Echopen, interested in how to preserve data
• Emmanuel Kellner is working on a fine dust sensor for citizen collecting data and does a PhD in Geneva
• Fabio, work on games for health
• Gaston Kayenga, Congo, intéressé par la collecte de données en temps de guerre

Challenges and best practices

• Fabio - Breathing Games we thought to register the data of the games locally and to enable the player to share them via a transfer button to reduce the challenges regarding privacy Pierre - Checked with regulators? Not the resources for the moment, we had a discussion within our group including with clinicians - in the future, when the community is bigger, we would like the users to be involved in what to do with the data
• Emmanuel - we are in a prototyping phase and explore different approaches regarding needs, fears and solutions in data such as geolocalization - for the moment we inform the users that data are shared - as the air pollution sensor collects data of the whole traject, for example when it is put on a bike - what is more challenging is the possibility to trace where the person lives or works - people could add dots where people could hide certain zones of one kilometer - we have open data collection and serveurs, which eases the evolution of the techniques - the user also can change the setting files to change names, etc - we have to define how we will set a standard Pierre - does it involve an ethics committee ? No, we are an association, discussing with people knowledgeable about such challenges, for example active in MyData etc., we started from the academia and are working on disclaimers, open ways - an ethics committee could just ask to inform people, we want to be more proactive and get people involved in what they want to do with data
• Fabio - writing a report on open science in Canada - the public research funds will implement requirements for subsidized projects to share their data - thinking from the beginning of a funding request what data will be collected, how it will be used, etc - there are platforms where medical/health data can be shared at different levels - aggregated data to the public, specific data to a certain group of people - some platforms offer an independent review council where the providers of data set the rules for accessing the data (being affiliated to a university, having a research proposal, etc) and the person interested has to provide information to the council
• Emmanuel - the Swiss data center offers a public repository with data curators, and people can propose algorithms that can be applied to different data - so the collect and analysis of data is separated - there is a possibility to redo studies
• Pierre - similar to a software you want to develop without sharing the code, where different groups work on different part of the softwares without having access to the whole
• Vanessa - question on changing id of people - can they then be traced? Fabio - the user will decide when the data is shared to whom, by email Emmanuel - in LogAir, we do not care about who is collecting data, but can trace the sensor id, we set random id and the user can change it - we have no interest to trace individual data, but the person can trace their own activities using the id - maybe later we will need to identify which devices have to be fixed, etc.
• Pierre - I discover the european regulation GDPR, worked with the HIPA confidentiality standard on health data that is unified in the USA on a device that gathers data to be analysed on a server, before sending them back to a clinician - the ethics committee did not want the date of data gathering to be collected because they said people could watch the hospital videocameras to find the patient -  this will probably become more strict with sometimes a lack of bon sens
• Fabio - there is much to do for small organizations when multinationals have the means to hire many lawyers Pierre - yes the challenge is about resources, but regulators may do more follow up with big companies
• Brice - caregivers will often ask patients, users to collect data, for example can ask medicine students to collect data, patients locally are open to contribute, to share data as the physician is a sort of god, the person who does research and has students can more easily collect data - most data collection is done on - certain students who have a smartphone can use an android app to collect data but it has some costs - and there is a risk of lack of battery, data loss, etc
• Fabio - there is also open-source / gratis software for collecting the data - completed by the group
  • https://www.gnuhealth.org/
  • LibreOffice Calc, tableur
  • Répertoires en accès libre
  • https://www.revues.scienceafrique.org/
  • MySQL, base de données
  • Jupyter, Anaconda, Gephi, outils d’analyse et de modélisation de données
  • KoboToolBox, collecte sur smartphone.   [La demoiselle/dame qui presentait hier a C7 avait aussi une autre source?]
  • https://openmrs.org/
  • https://www.open-emr.org/
  • Plateforme de partage de données en santé avec comité de gestion des accès indépendant : Clinical Study Data Request , Vivli
  • https://www.openhumans.org/
  • https://physionet.org/
• Vanessa - Kia medical is a medical file that is being set in Bénin to collect data from the patient
• Brice - Covid, there is not a good system to gather information and funnel it to the authorities - what are options as software? Fabio - I’ll ask and put them here
• Pierre - science fiction is a topic that is a bit similar, it is difficult to follow-up with all the questions - SF can give us interesting inputs about where we could be in hundreds of years  - for example David Brin who wrote the Postman (not his best book) and wrote Earth and Existence, the Transparent society - a hypothesis is that capitalism could work if everyone had access to information - another thing is there could be a market of research engines, everyone could have a hacker developing a research engine to better find information accessible online
• Fabio - we have three levels in open science, 1) open access, enabling readers to freely access publications, and also publishing at low cost to avoid a North-South gap in the ability to publish, 2) open data, ensuring research can be reproduced and verified, especially a challenge in the pharma where many researches are not published, errors not shared, and methodologies are flawed, 3) open hardware including infrastructure, research equipment and devices, not only sharing access to these but also encouraging people to mutualize the ressources, ideally in a distributed way - meaning there is no central authority deciding, but communities contributing to a common project, with a collaborative way to make decisions on the contributions that will be integrated in the core - many examples can be found in IT (for example the social networks Diaspora or Mastodon)