One thing Britain is exceptionally good at is collecting and using health data for research, studying cohorts of people over many decades. A shudder of alarm rippled through the research world at the news this week that UK Biobank’s data had been put up for sale on China’s Alibaba site, with the science minister, Patrick Vallance, saying that more attempts to sell the data in China were expected. Some sensationalised reporting failed to make clear enough that no names, addresses, NHS numbers or other identifiers were included, nor that the Chinese government reacted fast in taking listings down and nothing was sold. But would there be a stampede of participants withdrawing from this or other research programmes?
Biobank dashed to reassure its 500,000 members, and as a longtime volunteer I received a message not only explaining what had happened but listing some of the invaluable research findings and remedies that had already sprung from our data. Remarkably, a representative for Biobank told me that only about 100 people inquired about withdrawing, and after each was spoken to, only 50 actually backed out – pretty impressive. Prof Sir Rory Collins, Biobank’s chief executive, says he will personally speak to any anxious participant.
The list of good done using Biobank data includes a blood test revealing motor neurone disease years before symptoms arise, a single gene behind almost all Alzheimer’s cases and a score to decide which overweight people have most risk factors and should be first for weight-reduction drugs.
Biobank collected exceptionally detailed data from its half-million recruits aged 40-69 in 2006: blood, urine, saliva, height, weight, hip and waist measurements, blood pressure, heart rate, grip strength, bone density, eye examinations, lung function and fitness tests. Lifestyle questions included location, education, shift work, mobile phone use, smoking, drinking, exercise, sleep, diet, mental health and cognitive function tracked over time. It took five years to sequence everyone’s genomes. They have my samples stored at -80C so future researchers can seek causes and cures long after my death. All that extraordinarily valuable data is not for sale, nor for trivial or suspect use, only for good.
This breach was apparently the work of rogue researchers from three Chinese institutions, Vallance revealed in his report to the Lords this week. It appears that someone was trying to make money from information that is rightly free and open to legitimate scientists.
The NHS, national and centralised, is the best storehouse for lifelong data. The US, with healthcare run by private companies, can’t do the same. In other countries with more locally devolved or insurance-based systems, it’s harder to collate fragmented national data. The NHS is largely to thank for making life sciences a genuinely hopeful growth sector.
Longitudinal studies have been a research jewel, allowing projects such as studying children born in the same month who are then followed throughout their lives. In the UK we have followed groups of people from 1946, 1958, 1970, 1989-90 and 2000-2002 and there is now a new study recruiting 30,000 babies this year. Longitudinal studies of various sorts have revealed to us the link between smoking and lung cancer. Others included the supreme value of early years education, the prevention of at least 100,000 cot deaths worldwide by finding the safest baby sleeping position, and the lack of social mobility after the age of five.
The organisation Use My Data, which founded by cancer patients grateful for research that saved their lives, campaigns to get people to join research projects, helping researchers devise trustworthy transparent data systems. They face down the world of paranoia, conspiracy and fakery.
The lack of panic after the Biobank theft had researchers heaving a sigh of relief. But it has resulted in some people calling for its data to stay locked up, only available to those willing to visit in person, not for researchers to use outside. Fiona Fox, of Science Media Centre, says some warn the logic of absolute data security is to lock everything so tight it becomes barely usable. Weigh the risk against the inestimable value of 22,000 researchers using Biobank data in 60 countries.
But there are problems ahead: while AI is turbocharging the speed of number crunching, it also challenges data privacy. “Even anonymised data can be de-anonymised,” writes associate professor of bioethics Simon Kolstoe in the Conversation, as “AI tools are able to find complex patterns or links in data that no human would ever be able to discover”, uncovering identities of those who contributed their information. The protection will have to come from laws on the use of data.
The really serious issue is the fall in the number of willing participants. A Biobank committee member, the epidemiologist Prof Sir Simon Wessely, tells me he has seen it drop by 10% per decade because of “survey fatigue”. Rubbish surveys after an Amazon delivery, or pressing a feedback button after using an airport toilet, have devalued important ones. “People think no one’s counting those results anyway,” he says. He points to the Office for National Statistics crisis with its employment figures: they can’t get replies. He used to get 80% replies to his surveys of mental health in the military; that’s now fallen to 40%.
Summon up your public spirit. A population-wide study recruiting now is Our Future Health, seeking 5 million volunteers, so sign up here. I’ve already done so – it’s simple, just a blood sample and a questionnaire gets you a £10 token. Everyone benefits.
Polly Toynbee is a Guardian columnist
