Open data: go FAIR or go home

This article is written by Dr. Allana LeBlanc, Senior Advisor to the Chief Data Officer, Health Canada. The views represented in this article are those of the author and do not necessarily represent the views of Health Canada.

• The problem: Transparency in decision-making builds trust – without understanding where data comes from, public trust can be eroded.
• Why it matters: The general public wants to know how and where data is used to make policy decisions.
• The solution: FAIR data, or data that is Findable, Accessible, Interoperable, and Re-usable can be summed up as quality data that is ready to be used (and re-used). Good data management is the foundation of FAIR data principles, and a necessary base for transparency in decision making.

While many folks may say that it was the quest for information during the initial days of Covid-19 that drove the desire for open data, this drive has actually been a few years in the making. Technological advances in science and research have enabled us to move from an external hard drive on a lab floor onto a cloud server. Servers can securely (and affordably) store terabytes of data in an accessible place where even your neighbour’s dog can make evidence-informed decisions! We have arrived!

Just kidding, while the drive for information is there, no dogs are making evidence-informed decisions, at least not yet. Yet being the operative word. The White House (along with NASA and some other powerful friends) declared 2023 the year of open science. While up here in the frozen tundra we affectionately call Canada we are not that far advanced, we like to think we have done a lot of great work so far (with even more to come!).

The FAIR data principles

Open data and open science can be nicely wrapped up as FAIR (Findable, Accessible, Interoperable, Re-useable) data principles. These principles were outlined by a set of researchers in 2016 to support humans and machines in the use and re-use of data. Summed up, they mean timely access to quality data for use. They are achieved by good data management. At an institutional level, this means putting in the supports necessary to find and use data appropriately. At the individual level, this means clearly identifying your data (including appropriate use and re-use). Collectively, this means Canadians have the information they need to make decisions by having access to scientific and research data. And, arguably more importantly, Canadians understand how, where, and why data is being used to make decisions. Transparency in data.

Wouldn’t you feel a lot better if you knew what data was being used to inform a decision about your health?

In 2020, the Government of Canada published the Roadmap for Open Science, which recommended that all Federal Departments and Agencies develop strategies and tools to implement FAIR data principles to ensure the interoperability of research data. Health Canada (along with several other Departments and Agencies) also adopted this action item in their Open Science Action Plan, which was published in 2022.

Efforts underway to support FAIR data include developing a comprehensive inventory of Health Canada data holdings, standardised machine-actionable data management plans, and promoting publication to the Open Government Portal. It also includes integrating best practices in data management and governance. Other priorities include clearly articulating who the data contact is, outlining any potential privacy or ethical concerns, and generally just being explicit about who/what/where and how data can be used for the lifespan of the data. Integrating Persistent Identifiers as much as possible can help to avoid broken links and lost information. Again we want to clearly articulate the who/what/when/and how to use data until it is disposed of.

CARE principles (Collective Benefit, Authority to Control, Responsibility, Ethics) acknowledge that FAIR principles do not fully align with Indigenous Peoples’ rights and advancing reconciliation. FAIR principles focus on the characteristics of data itself, and ignore the historical contexts and power imbalances that make that data available. CARE principles, which are complementary to FAIR, are people and purpose-oriented, are meant to give back power to Indigenous Peoples.

Open data by default

The government of Canada has an Open Government Portal where you can access datasets immediately. They also have some guides and tools to help you link datasets and filter by topic (pro tip: the priority on which these are uploaded is partially based on demand, so if you are interested in a topic, put in a request!). The goal at Health Canada is to increase the number of datasets on the portal, in hopes that people will use, and re-use the data for novel purposes.

Admittedly, initiatives like the Open Government portal, FAIR data principles, and even the year of open science are directed towards scientists, researchers, and those making policies (and decisions about policies). We cannot expect everyone to make statistical inferences on raw data, no matter how open it is. However, the cultural shift in making data open may be felt across the board. FAIR data is an important contributor to transparency of decision-making – everything from public health guidelines, to regulations on pesticides. And wouldn’t you feel a lot better if you knew what data was being used to inform a decision about your health?

So, open by default. Timely access to quality data for use. Ability to answer novel questions. Link different datasets. Data-informed decisions. Transparency in data. Perhaps dogs are smarter than we think!

👋 You can write an article like this one! Share your thoughts with a community of public servants. Learn more

(Image Credit: Unsplash)