Data sharing in the time of COVID-19

Four members of the COVID-19 Clinical Research Coalition participated in a high-level session on data sharing at the Geneva Health Forum on 4 May.

The speakers – Prof. Philippe Guérin, Infectious Diseases Data Observatory (IDDO); Dr Nathalie Strub-Wourgaft, Drugs for Neglected Diseases initiative (DNDi); Dr Rob Terry, TDR/WHO; and Ms Naomi Waithira, Mahidol Oxford Research Unit (MORU) – discussed the importance of sharing de-identified, anonymized individual patient data from clinical studies with other researchers so that data can be re-used to answer urgent questions.

The session, entitled Data sharing in the time of COVID-19: What works and what we need, was co-organized by the coalition in collaboration with the Drugs for Neglected Diseases initiative (DNDi) and the Infectious Diseases Data Observatory (IDDO), and moderated by Prof. François Chappuis of the University of Geneva.

“It’s arduous and expensive to collect data,” said Dr Terry, who shared a newly launched guidance document from WHO entitled Sharing and reuse of health-related data for research purposes. “We should get the maximum we can from the research that has been conducted by reusing data.”

More than 50% of research globally is funded by public money. If it’s not being reused, it’s being wasted– Dr Rob Terry, TDR/WHO

Despite widespread sharing of genomic and epidemiological data, however, the sharing of clinical study data during the COVID-19 pandemic has been relatively limited. “Data sharing is much lower than expected,” noted Ms Waithira, the Head of Data Management at MORU. “When requests are tracked, numbers are significantly lower, even when [researchers are] requesting data that was promised to be shared.”

“It is incredible the number of times the same [COVID] study has been done – so much duplication, small sample sizes and slightly different pathologies and endpoints. That is the reality,” noted Prof. Guérin, whose organization, IDDO, manages a platform for sharing COVID-19 study data. However, as long as the data are of good quality, even with disparate endpoints and differences in study design, data can be curated and pooled for analysis with greater statistical power.

“We increase detection power by putting together the data from many smaller studies and allowing secondary research. Re-use increases the value of any data collected,” agreed Dr Strub-Wourgaft, noting that for ANTICOV, an ongoing adaptive trial in 13 countries in Africa, if the principal investigators had had access to more existing study data, “…we could have identified a signal earlier, enabling us to switch to different treatments earlier. That’s what you need in the time of a pandemic – to make sense of the data quickly and react.”

The speakers also discussed the importance and the challenges of data protection laws, which are necessary to protect individuals’ data privacy but can sometimes be “over-interpreted” to the extent that they become an obstacle to data sharing.

“Data sharing has been going on as long as research has existed. Data protection is more recent,” said Ms Waithira. “There is a strong need to rationalize data protection laws, to make sure they do not work against the interests of the public they are meant to be protecting.”

“This is not a new field but for many, it looks like a new field with the technical, ethical, and legal issues,” said Prof. Chappuis, the moderator, in his conclusion. “Let’s do everything that can be done to keep these processes simple and straightforward so they can be applied and adopted widely. And so that trust, understood widely, can be secured.”


Watch the recording

Inline feedback
View all comments