A team of healthcare data specialists in Bradford has been working on the collation of datasets from a range of sources in an effort to track the spread of Covid-19 in its region.
The Bradford Institute for Health Research (BIHR) and Bradford Teaching Hospitals NHS Foundation Trust have also been creating a series of dashboards for researchers and clinicians to investigate factors that could have an influence on where the pandemic is most intense.
They have been using local authority adult social care data, investigating a range of other sources and aiming to make connections through the unique property reference numbers (UPRNs) for people’s addresses.
Kuldeep Sohal, programme manager at BIHR, said it already has pseudonymised healthcare data from the GP practices across Bradford and Airedale and the three hospital trusts, and some local authority data on adult social care. It also had agreements for data on children’s social care and had been exploring links with the housing department at the local authority and private housing providers across Bradford.
“The discussions have accelerated over the past two weeks with the local authority and housing providers, who are willing to share the data to provide the linkage with our existing healthcare data,” she said.
“We’ll be looking at exploring other data with local authorities, in particular credit data. There are a host of datasets we’ve asked for to support our Covid-19 response, and we’ll be looking to use the UPRN for the linkages.”
The UPRN is a 12-digit code specific to an address that makes it possible to identify a property in datasets from a range of sources.
Sohal said that, while the work has been going on for some time, the pandemic has intensified the interest in sharing data among other NHS organisations, local authorities and West Yorkshire Police.
Dan Mason, programme manager for research analytics, said the team currently has patient data from the hospital and two other nearby trusts – Bradford District Care NHS Foundation Trust and Airedale NHS Foundation Trust – and is able to place them on a map. But it sees the “holy grail” in being able to link patients’ NHS numbers with property reference numbers.
“That would give us a way in to look at correlations with all sorts of other datasets and find indicators of what could be happening at a whole system level,” he said, adding that at the moment the data is limited because the data is only available for patients who have entered a hospital – and so far the number of cases has been lower in West Yorkshire than in the south of England.
Dr Tom Lawton, head of clinical artificial intelligence at the trust, said the effort would also contribute to scientific research, providing data on a localised cluster in line with good academic practice.
The team is looking at other surrogate markers that could be linked with the data, such as from the local NHS 111 service run by Yorkshire Ambulance Service, and from tracker apps that could help to provide clues on which symptoms are most significant in the spread of the virus.
Dashboards for response
It has built dashboards that have been feeding into the response at the trust and across Bradford, and is looking to increase their capabilities with more NHS data.
“In terms of analytics for the district response, we’re looking at whether we can link into data on where people are discharged to and patient pathway analyses in downstream care settings, but we’re still putting that together,” Mason said.
There is also the potential to follow up using local authority and education data to establish how patients and their families manage after discharge.
“This is having a huge societal impact and we must be able to capture that and work with the partners in our region to mitigate the problems,” said Dr Michael McCooe, clinical director at the trust.
“A lot of this started as research but we’re keen this data is accessed and used by our partners to improve things for our citizens.”
Image by Muenocchio, public domain