European healthcare professionals say their field is lagging behind when it comes to digitising and sharing patients’ health information. This could be putting lives at risk.
“21st-century healthcare is providing 20th-century care,” said Nigel Hughes, scientific director of Real World Evidence at Janssen Clinical Innovation, who was among the experts who met under the framework of the European Health Parliament (EHP) initiative to discuss the future of data in health.
More than 60 young healthcare professionals from across the continent met in Brussels in early December to discuss the EU health system, the challenges it faces, and the creative thinking that could help tackle them.
Electronic records could save lives
Hughes’ 26-year-old daughter Rose has Addison’s disease, a long term disorder resulting from insufficiency of adrenal hormones. People with this condition suffer life-threatening adrenal crises, for which they have to be hospitalised.
Hughes said: “Last week Rose had an Addisonian crisis, which is potentially fatal. Without fail, every single time, it’s a disaster, because it’s considered a rare disease and nobody knows how to manage it. Except for Rose, who’s had it for 26 years, and she has probably the most experience in the hospital.”
Hughes says this situation is worsened for patients due to a lack of digital record-keeping. “There is no such thing as a memory or record in the system to check what happened last time.”
Medical history isn’t digitally shared between European healthcare providers. For Rose, this means that every time she has a medical emergency, she finds herself explaining to healthcare professionals how her crisis was dealt with in the past.
Hughes is an advocate for electronic health records, which he says could be used for both care and research.
“Healthcare is not a data-centric industry”
Hughes refers to the music app Spotify as an example of how to successfully track users' histories.
“It’s fantastic,” he says, “it gives you your top five tracks, top five artists, and you can even go back every year. It gives you all sorts of statistics, a new playlist to listen to, the whole year of top tracks and so forth.”
“So why can Spotify do that, but healthcare professionals have no idea what happened to Rose last month, last year, or five years ago?” Hughes asked his audience in Brussels. “Because healthcare is not a data-centric industry,” he concludes.
But managing data accurately is anything but easy
For all the advantages, turning healthcare into a data-centric industry is not without its perils – mostly because of the difficulties providing a clear legal and regulatory framework.
One of the disadvantages of tools such as algorithms, supercomputing and artificial intelligence is that there is “a poor amount of representative diverse data, which represents the populations,” says Hughes. Data-driven technologies are only able to learn from the data that they have access to, which means people who don’t see a doctor regularly, or don’t have access to one will not be part of the data that is gathered.
“We are concerned about the bias that some of these algorithms and protocols can have,” says Sarah Collen, senior policy manager at the NHS European Office, “and that’s why the NHS [the UK’s National Health Service] is working on developing a code of conduct on data-driven technologies.”
Collen says developers need to be aware of where their data starts and finishes. “It is a tool that may have a bias against certain groups, whether that's women, children, people with a rare condition, or people from a different ethnic background,” she says.
But even if data governance is hard, Collen says, “it is a fundamental right and it is our job to work out how to use that data carefully, and in a secure way, to provide quality effective care.”
“All the data we collect is biased”
According to Bob Ghosn, head of the Emergency Unit of the International Department at the Red Cross in Belgium, “all the data we collect is biased. But that does not mean that we should not collect it, and certainly, it does not mean that we should not be data-driven organisations. It just means that we need to be aware of this biases and try to mitigate them.”
Ghosn says that we tend to think that digital data is more biased than other data, but that nothing proves this to be the case.
Mismanaging data has unintended consequences
The 2013 Ambien scandal serves as an example of the disastrous effects of data mismanagement. The US Food and Drug Administration (FDA) had to reduce the dose of the sleeping pill Ambien for women, after realising their bodies metabolised it slower than men’s. The discovery came after several car accidents that occurred when women who had taken Ambien fell asleep at the wheel, because they had not fully metabolised the drug after the eight hours they were told this would require.
Outside of health, it was recently brought to light that Apple credit cards were giving less credit to women than their male spouses – thanks to an AI algorithm. Apple co-founder Steve Wozniak spoke out against it, saying he could borrow ten times as much as his wife on their Apple-branded Cards, even though she had the better credit score.
EHP is a partnership initiative comprising six entities: EU40, Johnson&Johnson, the European Patients’ Forum (EPF), the College of Europe, Porter Novelli, and the European Affairs Recruitment Specialists (EARS).