Special Report: Analytics
Traditionally, analytics tools have been used in the NHS to understand current situations or past trends. But with Covid-19, has the role of analytics been permanently altered? Maja Dragovic looks at the impact.
Analytics is seen as a tool which can help the NHS predict where emergency services are most likely to be needed and allow trusts to plan accordingly.
Though in a crisis such as the Covid-19 pandemic, it is important to be flexible, says Matt Hennessey, the chief intelligence and analytics officer at Greater Manchester Health and Social Care Partnership.
“In the past what we have done when facing a bit of a challenge is, we look at what we’ve got and try and make do with the existing infrastructure. If you have a great tool, you tweak it a little bit to get you up to what you want, but if you alter it for a long time or the requirement starts to grow you can start having problems with it,” he adds.
An example is that data that had to be collected from care homes early on in the epidemic but was likely to change over time.
“First we just wanted to know whether they have enough PPE,” Hennessey says.
“Strategically we knew that since this was a novel virus, it might suddenly become important to know how many residents are over 100 or what people had for breakfast.”
Hence, for Greater Manchester they needed to develop a diverse data collection tool. They decided to go for an eForm because it meant that whenever someone logged in to fill in the information, they could just add another field without having to deal with updates, which would have been the case if they used an app, which was another option.
Collecting data using one digital system proved to be useful, says Hennessey, who is also associate director at Public Health England.
“We did everything in the one system, we didn’t just use care homes (data), we looked at GP practices, pharmacists, urgent dental centres,” he explains.
“We built a dashboard based on the data we were collecting, and you could see where the challenges were. If a pharmacist were running low on gloves, you could see that just up the road there’s a GP with a two-month supply. It enabled us to respond extremely quickly.”
On top of this, Greater Manchester were also broadening the remit of people who were viewing the data.
“We started to use a single digital platform, then the other people in the system, who wouldn’t normally be looking at data, could look at it and make local decisions based on it,” Hennessey says.
“So, for example, care homes could look at the availability of GP appointments, or the hospitals were able to predict how many spare beds there were in care homes. We were able to get hospital doctors to be able to see GP records, we were able to get pharmacists to be able to see data on adult social care.”
Requests for more access to data
In terms of greater access to data, Greater Manchester is not an isolated example.
Bruce Horne, product specialist lead at Orion Health, says that during Covid-19, they saw an increase in requests amongst their customers to expand access to patient data appropriately to wider audiences within their organisations, surface that data for the purposes of population health management and explore opportunities for machine learning and AI.
“We have worked with one customer, Connecting Care, to surface a risk score within their shared care record,” Horne says.
“The use case behind this is to identify patients at “high risk” of complications from COVID-19 (as published by NHS Digital).”
For Richard Betteridge, a data scientist at Cerner, the increase in requests partially stems from the loosening of information Governance rules.
“Some of those restrictions have loosened up during the Covid crisis which meant that more clinical information is available for clinicians to support the direct care that they are providing,” he explains.
“A good practical example is that we’ve been able to bring social care and automation for north central London into the platform for the first time alongside the healthcare data.
“So, if you are a GP logging into the system, for example, to see admissions and discharges from hospital which was one of the use cases that we developed, you’d also see alongside that whether that individual was transferring to social care. So it actually provides clinicians with more rich dimensions so they can start to understand the population in which they serve, especially in the case of primary care.”
Intelligence and insight
One of the biggest impacts of the Covid-19 crisis, however, is the recognition of the importance of analytics teams within NHS organisations.
The role of analytics and data is becoming a more central component of a healthcare organisation and therefore, the analysts information teams are becoming crucial as decision-making teams need people who can interpret and understand the data on the ground, says Charlie Evans, who manages analytics practice at Cerner in Europe.
“Healthcare has become a lot more data driven,” he concurs.
In the case of the Northern Care Alliance NHS Group, it now has more than 20 analysts who support the five hospitals within it (Salford Royal, The Royal Oldham Hospital, Fairfield General Hospital, Rochdale Infirmary and North Manchester General Hospital).
At the start of the pandemic, the Alliance has set up an analytics cell to support all of the organisation, both in terms of managing the pandemic itself and more recently in restoration and recovery activity.
“Teams were able to do a lot of predictive modelling using statistical tools so we have an idea of where we’re likely to see increased infection rates and then modelling through what the impact would be in terms of our hospital resources,” explains Emma Wright, director of information and business intelligence for the Northern Care Alliance NHS Group.
“They’ve done a lot of capacity and demand modelling to understand how much activity, how much surgery we will be able to perform in this new environment.
“Overall, we’ve seen a real increase in how the organisation values its analysts and its analytical resources. They’ve certainly become part of the fabric of the organisation and they are very much now imbedded with operational teams.
“Since Covid-19, we are relying more on analytical tools. Not just in relation to Covid-19 but in other areas as well, particularly around activity restoration.”
Long term impact
Will the Covid-19 crisis permanently alter the way analytics tools are used? In New Zealand, the government is already developing a national machine learning algorithm hub to support the country’s response to the pandemic.
The hub, which is being developed together with Orion Health, will support scenario modelling, risk prediction, forecasting and planning around pandemics.
“The idea is that health and care organisations across New Zealand will be able to access information and tooling they need for effective decision making,” Horne explains.
In the UK, a national algorithm hub may not be on the cards right now, but there are other positive developments that are expected to follow.
“From a tools perspective, I can see organisations moving much more towards open source, open data that enables people to share both their data and their modelling in between organisations,” argues Wright, who also expects the on-demand platform and self-service to become more prevalent.
From an analytics perspective, Wright adds, analysts are now on the cusp of journey towards being recognised as a professional body in their own right and she believes this will encourage more people to have a career in analytics in the NHS.