Data through our lifetimes: Harnessing the power of longitudinal studies for health
3 March 2021
The 2020 UK Health Data Research Alliance Symposium tackled some of the biggest topics in health data research. Panellists Neena Modi, Mike Inouye, Nic Timpson, Angus Patrick and Diana Kuh discussed how to unlock the potential of the rich health data within the UK’s impressive collection of longitudinal cohort studies.
Our health changes throughout our lifetimes from childhood through to old age. However, many research studies only provide a snapshot in time and don’t capture the longer term changes that underpin health and disease.
Longitudinal cohort studies solve this problem by following people over a long period of time – years or even decades – tracking their genetics, mental and physical health, lifestyles and more. Here in the UK we are lucky to have many of the world’s oldest longitudinal studies – the research equivalent of a vault stuffed with gold bullion, providing rich insights into the nation’s health.
One example is the University of Bristol’s Avon Longitudinal Study of Parents and Children (often known as the ‘Children of the 90s’) – a UK Health Data Research Alliance member – which has followed 14,000 Bristol women since they were pregnant in the early ‘90s. Over 30 years they’ve patiently answered questions on everything from their health and lifestyle habits to their sleep patterns, and given urine, DNA and placenta samples and more.
Many of the original participants’ children and grandchildren are now also taking part in the research, and today it boasts data on 30,000 people across three generations.
Meanwhile the National Survey of Health and Development (NSHD) is the oldest birth cohort study in the UK, tracking 5,500 people born during one week in 1946 across their entire lives and making them some of the most studied humans on the planet.
It’s certainly interesting to hear what a post-war toddler ate for breakfast or how a woman’s sleep patterns compare with her children’s – but why is this health data so prized by researchers?
“Lifetime data from these studies has provided evidence about the impact of poor lifetime socio-economic conditions, and psycho-social stress on adult physical and mental health.”
Explained Professor Diana Kuh, who led the NSHD for several decades.
The Children of the 90s study, for example, provided evidence that cot death was more common in babies put down to sleep on their tummies. In turn, this led to lifesaving government health campaigns advising parents to settle infants to sleep on their backs. And the NSHD showed that lighter newborns are more likely to have high blood pressure as adults.
Recently, longitudinal studies have come to the fore in gathering data on COVID-19. As an example, the Bristol team acted quickly to gather data from participants to help assess how people’s immune systems are responding to COVID-19 infection by tracking cases of disease through to recovery. This gives clues on how they’re likely to respond to the new vaccines.
“We’re able to react quickly with policy-relevant material and be dynamic in the way that we collect data.”
Said lead investigator Professor Nic Timpson.
“In terms of COVID-19, these established cohort studies are really critical for assessing the long-term consequences of exposure to infection and lockdowns on the health and wellbeing of the population.”
Genetic and genomic data from large-scale population studies like longitudinal cohorts can also be used to identify groups at higher or lower risk of diseases, whether that’s COVID-19 or anything else.
“Risk prediction may affect screening strategies – when or who might be brought in for testing. It may impact who we might give a particular therapy to, how we might calculate things like benefit versus harm.”
Said Mike Inouye, director of the Cambridge Baker Systems Genomics Initiative at the University of Cambridge.
Longitudinal population studies could be even more useful if they were to be linked up with routinely collected health data, such as information held in NHS health records. This approach holds lots of promise but is still in its early days, especially for older records that have not been digitised.
As Mike explained, the major reason to link health records to research studies is because bigger is always better when it comes to data analysis:
“We have incredible data assets available at our disposal with very large sample sizes and independent multiple external data sets where the designs are very similar, so you can then do much more robust statistical analysis or machine learning on those data sets.”
Another potential source of large amounts of real-time data is health apps and wearable devices such as fitness monitors, which are becoming more commonplace. INSIGHT hub public representative Angus Patrick was excited about the potential of using the data collected from these devices to further enrich longitudinal studies.
“People are wearing a Fitbit or an Apple watch for 24 hours a day – that’s been capturing an awful lot of data for an awfully long time.”
Said Angus Patrick.
Although adding data sources like wearables and genomics to longitudinal studies has huge potential for generating rich health insights, it’s important that participants are informed and happy with the data that’s being collected about them.
As a longitudinal study goes along, the consent of participants is constantly updated, making such studies well placed to collect these new types of data.
“These new technologies are really exciting. We’ve been using activity sensors, and there are instances in our studies now of real-time video recordings of child and mother interactions in the home. That’s only possible through new technologies that offer remote data capture in real time.”
Our health changes throughout our lifetimes, so we need health data research that spans whole lifetimes too. The UK’s stash of ‘gold bullion’ in longitudinal studies is only set to grow in the future as technology and tools improve, bringing exciting opportunities for understanding and improving health for life.
Content from the day: All of us together – UK Health Data Research Alliance Symposium
On this page you will find outputs from the second UK Health Data Research Alliance on 1 December 2020, including slides, video recordings and the post-event report.
The promise of AI for health data research
2 February 2021
The 2020 UK Health Data Research Alliance Symposium tackled some of the biggest topics in health data research. Panellists Jim Hendler, Cian Hughes and Laura Robinson explored the promise of using...