Harnessing the potential of data analytics

Profile photo of Josie Harding, Head of Insight and Evaluation During the pandemic, collaborating with partners to gather data for analytics was essential in looking after our population. Now it’s in place, there’s no turning back, says  Head of Insight and evaluation Josie Harding.

If you look at any list of hospital patients, many of them will be receiving support from other services too, including primary care, acute care, and local authority services such as social care or housing. But traditionally, each organisation has held its own data. This makes it hard to paint an overall picture of the challenges faced by local populations and how best to meet their needs.

When I started as Head of Insight and Evaluation at the trust in 2019, I was glad that the strategy team was taking this broader perspective, looking at how population health was affected by the wider system. I’d previously been an analyst in community safety and social care, so I’d seen the system from all perspectives. It seemed clear to me that when organisations pool information, they can get a much clearer picture of what is needed, but the process isn’t always straightforward. Then Covid hit – and suddenly a shared approach became the only way to manage the situation.

Pooling resources

By the end of March 2020, the number of Covid infections started to shoot up. We’d already started building connections with analysts in other organisations, but it was obvious we’d all have to start sharing our information immediately.

As part of the Covid response, a network was set up, bringing together data analysts from all the organisations in the system. No one had any idea what the pandemic might look like and there was a flurry of discussions about Covid modelling methods. It turned out that some network members had highly relevant experience: epidemiology, public health and even analytical techniques for monitoring pandemics.

Because we were at the hospital, our team was able to share really useful data. We had a huge number of patients, so we knew which groups were being most affected, how Covid was impacting patients, their length of hospital stay, the impact of discharge and mortality levels. These factors told our colleagues where to allocate resources such as community support after discharge.

Improving our analytical skills

We also built our analytical skills internally. We started working in Excel, but we accessed training in R – a specialist programming language that enabled us to improve our modelling. Then a network member with experience in modelling pandemics taught us how to model specifically for Covid-19, helping us predict bed numbers.

The whole process was entirely evidence led.

By Wave 2, we were also learning from the experiences of other European countries and from Australia, which was in the winter season while we were still in our summer. We started to build an understanding of the unique profile of Mid and South Essex, with our own trend lines.

One pattern we noticed was a two-week lag in any intervention. If lockdown was imposed, two weeks later we’d see fewer patients. Similarly, when lockdown ended, two weeks later we’d see an influx of patients. So, when government changes were implemented, we knew what to expect.

Planning beds around data

Around that time, the trust decided to consolidate its critical care skill sets and knowledge by creating a critical care centre at Basildon Hospital. By drawing on earlier data, we could pinpoint where we’d seen pressures in Wave 1 and estimate the need for oxygen or beds based on those peaks. It was massively motivating to see people planning the number of beds around the figures that we were providing, rather than budget or capacity. The whole process was entirely evidence led.

Daily reports

When it hit, Wave 2 was about three times the size of Wave 1. We hadn’t anticipated this, but we started doing daily analysis. Every day, as we saw the rate doubling, we adjusted our modelling so predicted rates rose much more steeply.

We began sending data reports out three times a week to four key staff, but by early 2021 about 80 staff were requesting daily reports. Because we were monitoring the situation so frequently, people had a lot of confidence in our forecasts and were using them to decide things like the number of planned surgeries that were possible while ensuring we would still have enough beds.

In the end, our figures weren't far off. That was a great advantage for our trust, but it benefited the wider system too, as it helped community providers identify pressure points and decide where to allocate capacity.

People had a lot of confidence in our forecasts.

A new way of working

The way we used data during the pandemic has been a springboard to a new way of working. Without collaborating as we did, we couldn’t have identified and supported some of the most vulnerable people, and our hospitals wouldn’t have had the resource and space to provide safe care. Now, we’re applying these techniques to working through the backlog of elective care.

It’s also paving the way for the new integrated care systems, which look at the entire journey of our patient population – from prevention through to attendance and admission – identifying areas for improvement along the way. We’re well on our way as a Trust now, and there’s no doubt the pandemic has accelerated that process.

But also, getting to know the other data analysts, with these amazing skills, was an unexpected opportunity and a real pleasure. Many of us worked in the same area – but without Covid, we’d never have met. That will be a lasting legacy of the pandemic: now this work has started, we’ll keep on working closely together across the system.