NHS England enhances BPM with data analytics

Data analytics and business process management (BPM) are increasingly converging as modern businesses seek to unlock new possibilities that support enhanced operational excellence (OPEX). By adopting data analytics in BPM, organizations can harness data-driven insight, optimize processes, find hidden efficiencies and deliver successful strategies.

At National Health Service (NHS) England – the publicly funded healthcare system in England and the second largest single-payer healthcare system in the world – data analytics helps transform multiple aspects of BPM. That’s according to James Friend, director of digital strategy at NHS England – London Region.

“We know that BPM works best when it’s set in the context of data,” he says, and that is particularly true in healthcare. “Our error rate risk appetite is zero because we don’t want to harm any patients. We spend a lot of time through a national program called ‘Getting it right first time’ to drive better outcomes. There are a number of different examples, but it’s ultimately about getting decision-makers to think about what the data is saying to start off with and having the right effector capacity to deliver the change.”

Start with the problem

The approach to BPM is to start with asking what the “problem” is, says Friend. “Be really, really clear what the root cause of the problem is, because so many times we see people try to apply process change management to parts of a path of which aren’t the issue.” Data is a really good source of knowledge in this regard, he adds. “We’re investing a lot of money bringing more data into theater management. Productivity and smooth running are key for patient and clinician experience.”

The key piece of knowledge is understanding the drivers of performance opportunity. “Once you’ve got that bit there and a clear definition, can you articulate that in some form of metric? What is the baseline data that you are trying to change?” When you know enough baseline data points to understand your current, upper and lower confidence levels, you can better gauge change and expectations around it, Friend says.

“We often see business improvement initiatives that might give you a run of more data points that the mean ‘up’ is good, but it’s not changing the overall mean or doing it in a sustainable way. You’ll get the odd one that falls way below and you’re back where you started. We’re looking for a sustainable process transformation.”

James Friend, NHS England, discusses defining the problem in process management

The OneLondon Health Data Strategy Program

A couple of years ago, NHS organizations across the UK’s Capital City launched the OneLondon Health Data Strategy Program to create a more proactive system using data at scale, Friend says.

“The principle behind the program is to use data insight for better individual and population healthcare, to support clinical academic research or healthcare product research and development. We wanted to test whether our data infrastructure – the data itself and the systems that support it – was fit for purpose for these aims.”

The program commissioned several projects, one being around cancer diagnosis pathways. “The goal is for everybody to have their diagnosis within 14 days, because we know the faster you work, the better your clinical outcome.” The project team looked at all the data points across the NHS Trusts in London that provide prostate, sarcoma and neck cancer care to find the biggest correlating reasons for patients meeting or not meeting that faster diagnostic standard, Friend says.

“First, they realized that there were about 500 different data fields available to them, so they did quite a big workshop process to narrow down which of the fields were actually going to helpful. I think they ended up building a model with 90 different data fields in it.”

The factor with the greatest correlation to patients not achieving the faster diagnostic standard was that the hospital or healthcare provider was choosing primarily to communicate with the patient by the Royal Mail. This meant that patients were less likely to achieve optimal diagnostic standard if the hospital relied on post to communicate with them, Friend says. The NHS has since invested in the NHS App to communicate quicker and more effectively with patients.

3 aims for all process redesigns

NHS England – London Region aims for every process or digital transformation to achieve one or more of three things, Friend says.

The first is whether it helps NHS England align its clinical capacity with pathway demand, addressing process bottlenecks and making them more efficient. “Money isn’t necessarily the constraining factor in the NHS, it’s actually people, skills and clinical capacity. The more clinical time we use by getting nurses to type things into an electronic patient record when they could be delivering care [for example], the less efficient we are. If we can streamline that process, it makes a big difference.”

The second thing relates to “making the right thing to do for the patient the easiest thing to do for the clinician,” Friend says. Businesses – particularly those in highly regulated sectors – have traditionally responded to incidents by making it harder to do the wrong thing. “While that [approach] has a role, the key is making it easier to do the right thing – incentivizing people to do the right thing in the easy way.” Pulse surveys play a big role here, asking frontline clinicians about the things that make it difficult to do their job and then addressing them, with patient-led designed part of the puzzle too, Friend adds.

The third thing is about getting the patient to the most appropriate environment for their assessment, for their treatment and for their care, some of which may be digital. “If you think about a sales model, that’s very much about understanding and delivering the customer’s needs – making sure you meet the needs afterwards. In healthcare, treatment can be thought about as the product and post treatment care as the ongoing support.”

He cites the ‘click and collect’ retail model example. “You’ve got a virtual needs assessment bit, which confirms the best purchase, you’ve got the physical part of collecting the product and then you’re probably back to virtual again for the aftercare.” For the NHS, this type of offering is more like a digital outpatient appointment, a physical theatre appointment and physical aftercare.

“What’s really interesting from a process design perspective is that every hospital is different, and therefore every hospital has a different dividend. If you happen to have your X-ray department built next door to your A&E [accident and emergency] department, it may well be the right process to take the patient to the X-ray and bring them back again.” However, if your A&E department is at the other end of a hospital site from the radiology department, you might want to have a process that has mobile X-ray machines in your A&E department. “You’ve got to be able to define that environment based on everything else you’ve got around it, getting the patients the most appropriate environment for their assessment, for their treatment and for their care.”

If one or more of these factors is being addressed, and there’s clear data that’s helping to address the cause of a defined problem, you have the building blocks for effective process redesign, Friend adds. “Then you can monitor and build in metrics-backed patient and staff feedback.”

Ekaterina 'Katie' Curry, Millennial Specialty Insurance, and Donald Kuk, formerly of BNY Mellon, discuss using automation and BPM

The biggest BPM and data analytics challenge

The biggest challenge around data-driven BPM for NHS organizations across London isn’t monetary or even data access-related – it comes down to driving change on a large scale. “We’ve got data analysts to sit down next to the clinician, identify what the data query is, run it and present it back. The big challenge is about getting the organization as a whole to effect change,” Friend says.

Even with top-down direction to hospitals to advocate and explain process changes, there’s no guarantee that messages are getting through and being heard. Then there’s the issue of individual hospitals’ prioritization levels to impart the changes. “That’s not really a money thing,” Friend says. “It is simply about bandwidth, about how many hours there are in a day to be able to affect certain numbers of changes and track their specific impact.”

Friend identifies that trying to achieve process changes across London or across the country is challenging, describing how the NHS operates within finite resources and is often neither “national, about health or a service,” but rather a series of federated, local organizations that deal with sickness on a reactive basis. True process change is needed to bring care upstream into proactive prevention, supporting the ambition of the NHS’ local leaders focusing on specific community needs, raising clinical productivity and making patient experience seamless.

NHS England is the regulator – it sets the policies and processes that are expected as performance standards, but it doesn’t own the mechanism, Friend says. The best way to implement changes can differ from one local Trust to another, depending on specific factors. “The hospitals and the geography are different in different places and that’s why we have integrated care boards and care systems to design things around what matters locally.”

What does the data tell us?

The ultimate goal is to support business leaders, system designers and clinical decision-makers to ask what the data tells them before making a decision. “That’s the one big change we could make – just simply getting people to be inquisitive,” Friend says. “In process management and improvement, it’s key to think about what the data is telling you before you start.”

In healthcare, it’s common to be penalized for taking risk; doing something that isn’t what you were taught to do, he adds. “You tend to intuitively make the same decision as you made before.” In Winter, for example, the NHS has traditionally opened more beds to avoid running short, but that hasn’t always helped if staffing numbers were the real problem, Friend says. “In those instances, we weren't looking at the data and asking what it tells us. We were doing what we’d traditionally done before.”