Four lessons the pandemic has taught us about health data


Data has played a vital role during the pandemic. It helps to assess the impact of new treatments, monitor trends in the quality of care and identify the needs of the most vulnerable. The government’s release of data on the number of cases, hospitalizations and deaths has also been an important way for the media and the public to track the impact of COVID-19.

As the UK slowly emerges from the pandemic, data can play an equally important role. This includes understanding long-standing issues (like rising levels of mental health problems and long wait times for care), more immediate priorities (like rolling out integrated care systems), or even how how the NHS in England compares to other healthcare systems internationally.

At the Health Foundation, we have been working since 2014 to use data to improve health and care. However, the COVID-19 pandemic has been a steep learning curve for everyone. We want to make sure we learn from the successes and failures in how data has been used during the pandemic and take advantage of the unique opportunities ahead to improve health and care for all.

Here are my four lessons from the pandemic when it comes to health data:

1. Don’t let missing data make certain experiences invisible

Despite best efforts to ensure datasets are representative of the entire population, the experiences of many groups remain missing from the data we collect about people’s health. This affects the ability of the NHS and the wider public health system to meet everyone’s needs.

To give an example, we know that people with learning disabilities are generally at increased risk of infectious diseases. But four months into the pandemic, there was still very little data on how they had been affected by COVID-19, as Chris Hatton explains. The situation has gradually improved, but these information gaps have made it much more difficult for this group to mobilize action. Similar issues with missing data have affected many other groups in our society during the pandemic.

This poses a daunting challenge for anyone interested in using health data. Missing data is not just a technical problem. We are increasingly dependent on data collected through digital systems, and people’s ability to use these systems is shaped by deep practical, economic, political and social considerations. We need to understand these impacts so we can design systems that work for everyone.

2. Include public and patient engagement as a standard in health data projects

There are many reasons to involve the public and patients in health data projects. This gives the audience confidence that their information is being used in accordance with their wishes, and it can also improve the end product.

Examples abound. The COVID-19 contact tracing app then made a valuable contribution to managing the pandemic. However, by any measure, he had a rocky start. This was partly due to public concern about how the data would be used and whether the app was the right solution. These problems could have been avoided with more public participation.

It’s an approach that doesn’t always come naturally to health data scientists. But we’ve been working to develop our own practice in this area at the Health Foundation, and we’re already reaping the rewards.

3. Make open analytics the norm

A recurring theme throughout the pandemic is that policy makers and the public have not had access to the data they need. Often, data is collected from different parts of the system, but not shared. For example, local public health teams have not always had access to data held nationally. But where data has been reliably shared, tremendous progress has been made.

The Open Data Institute has done a fantastic job helping organizations make their data open. It can be difficult, with technical and cultural issues to resolve, but the benefits are a more efficient and transparent ecosystem.

Creating linked data systems is something that is close to our hearts at the Networked Data Lab, where we are also the first to share code as well as data. We recently published our first Networked Data Lab analysis of the impact of the pandemic on the clinically extremely vulnerable population. My colleague Kathryn Marszalek talks more here about how open analytics could help us solve key problems in population health and care services.

4. Address the unequal benefits of data and technology

There is still a lack of good mixed-methods data on the impact of technology on health inequalities. A handful of studies in the US have indicated that people from specific ethnic minorities may be less likely to receive care thanks to new technologies, but these studies have not been replicated in the UK.

There are other reasons to worry. For example, our survey during the pandemic found that people in lower-paying jobs were less likely to download the contact tracing app, despite arguably being at higher risk of contracting the disease. This apparent mismatch between needs and benefits suggests that technologies could be worsening health inequalities.

Over the past year, we have announced two new partnerships in this area. We are working with the Ada Lovelace Institute to examine the impact on health inequalities of some of the technologies introduced during the pandemic. And we’re partnering with the NHSX AI Lab to support research aimed at improving the impact of the next generation of artificial intelligence technologies on black and minority ethnic communities in particular.

What can be done?

Everyone can play a role in improving the impact that health data and new data-driven technologies can have. It’s important to be constantly curious about how data shapes the world around us. During the pandemic, food deliveries were prioritized using data partly taken from non-confidential health records. It is a remarkable fact. Fast forward twenty years, what impact could health data have on society?

Data teams should ensure that data collections are representative and subsequent analysis is inclusive. They should also involve the public and patients in their work, and share code using GitHub wherever possible. Managers should make resources available for this and set aside time for their own development to familiarize themselves with modern approaches to reproducible and ethical testing.

With the right action, we can ensure that advances in the use of health data benefit everyone.

Adam Steventon (@AdamMSteventon) is director of data analysis at the Health Foundation.

This content was originally featured in our email newsletter, which explores perspectives and expert opinion on a different health or healthcare topic each month.


Comments are closed.