Contributors include: InterSystems, Wellbeing Software, United Lincolnshire Hospitals NHS Trust and East Suffolk and North Essex Foundation Trust.
AI, data and automation offers enormous potential for healthcare organisations, to save time, support decision making and remove manual process.
In this feature we highlight four examples that are transforming health and care and making AI a reality. The first, the InterSystem platform, it utilises hundreds of thousands of data points to support healthcare professionals at a population level and it helps take away a significant time cost of data management.
It could be a quick win for the NHS
Artificial intelligence (AI) has the potential to transform the way healthcare is delivered in the NHS. By applying algorithms to large data sets, it is possible to make diagnoses or predict the likelihood of individuals developing particular diseases.
The implementation of AI, however, requires substantial amounts of data. In the NHS, data is usually held in multiple clinical and administrative databases, often in different formats, making it hard to access and manage. Data may be incomplete or inaccurate, so that creating a single, coherent, usable data set is time-consuming. Even a task such as deciding whether a care record from two different clinical systems refers to the same patient can be laborious. About 80% of data scientists’ time is spent on “data wrangling” – the job of converting raw data into something useful for AI purposes.
InterSystems addresses this through IRIS, a data platform that simplifies the job of taking data from multiple sources and making it usable. HealthShare, the Regional care record solution from InterSystems, now uses IRIS to amalgamate patient healthcare information from different sources.
It is then relatively straightforward for customers to apply AI algorithms to the data. One US customer, says Jon Payne, the manager for sales engineering at InterSystems, is using population-level data to look at the risk of patients developing diabetes, for example, or of going to the emergency department. The customer is then able to “develop appropriate care plans or look at ways to improve patients’ health.”
Another InterSystems customer, says Payne, is a pathology lab combining genetic data with clinical data to improve predictions about whether a genetic mutation is likely to result in an individual developing cancer. Other customers are using AI to analyse data from hundreds of thousands of smart home drug delivery devices. “These provide information, not only about whether someone has taken a drug, but how effectively they’re using the device, such as flow rates for inhalers,” says Payne. “So you can look at the efficiency and effectiveness of the devices on a wider scale.”
Data protection laws mean that widescale adoption of AI by the NHS may take some time, says Payne. There is plenty of potential, however, to use it for operational purposes, such as predicting the availability of lab equipment. “Introducing change in that context can make a big difference and could be a quick win for the NHS,” he says.
We spoke with Wellbeing Software to understand how AI is making work(flow) better in the healthcare sector
The NHS has announced the launch of an artificial intelligence (AI) laboratory, aimed at finding new ways for advanced algorithms to enhance patient care. It will be supported by a £250 million investment, outlined by Health Secretary Matt Hancock. There could hardly be a clearer statement of the importance the government is placing on the potential applications of AI in the healthcare sector.
Radiology is one discipline where AI has a major role to play. In a recent survey of our UK customer base, 85% of respondents highlighted the advantages of AI integration within RIS workflows.
However, introducing AI to any healthcare context is not as straightforward as simply choosing an AI solution and setting it to work. As with any other new software deployed in the NHS, it is vital for AI tools to be thoroughly integrated with existing technology. After all, the power of AI lies in its ability to analyse and learn from vast quantities of information so it needs the ability to access that information in the first place. And, in order for clinicians to benefit from its implementation it has to fit into their existing processes and workflows.
This is why at Wellbeing Software we have developed our AI Connect platform, which enables the adoption of different AI applications within existing reporting workflows, using a standardised approach. As a technology-agnostic platform, it allows algorithms from a variety of AI platforms to be integrated with any RIS or PACS. Dartford and Gravesham NHS Trust, for example, have used AI Connect to implement Behold.ai’s red dot prioritisation in conjunction with our own RIS software. The combined system is able to categorise radiology scans within 30 seconds, based on more than 30,000 example images, and counting.
Neil Perry, Associate Director Digital Transformation, Dartford & Gravesham NHS Trust, said: “A key piece to the AI puzzle for Radiology is automating the pull and push of diagnostic images based on the required rules for the specific AI service and then managing the results or insertion of the amended images back to the PACS. Our recent AI development for chest radiographs (CXR) with Behold.ai would not be as efficient if not for the Wellbeing Software AI Connect gateway. Wellbeing Software has proven themselves time and time again to be real digital partners across radiology, maternity and wider.”
We are also working with stroke imaging company Brainomix to accelerate the introduction of AI and deep learning into the clinical decision-making process for ischemic stroke patients. Brainomix’s E-ASPECTS solution is designed to assist healthcare professionals to make faster, more informed decisions by applying AI when assessing ischemic stroke damage and can be integrated with existing workflows using the AI Connect platform.
This combination of flexibility and integration, then, is essential in enabling healthcare organisations to truly benefit from the possibilities of artificial intelligence. From the NHS Trust level all the way down to individual hospitals, clinics and other healthcare settings, AI technology needs to be fully integrated with the technology already in place, and it must be able to make use of the vast array of information already available within the health service. Artificial intelligence has huge possibilities for healthcare – but only with the right application.
AI to support screening at United Lincolnshire Hospitals
United Lincolnshire Hospitals NHS Trust is one of the first trusts to take part in a new trial using artificial intelligence to support breast screening.
ULHT is part of the East Midlands Radiology Consortium (EMRAD), a partnership of seven NHS trusts, spread over 11 hospitals, looking after more than five million patients.
Currently all images produced during breast screenings, known as mammograms, are reviewed by two members of the breast screening reading team. With a national shortage of radiologists and with almost a quarter planning to retire within the next five years, there is a clear need to investigate and look for potential alternatives.
The Trust is exploring ways to help develop, test and ultimately deploy AI tools in the breast cancer screening programme in the East Midlands.
The aim is to make the best possible use of scarce resources like radiologists’ time and scanners, and to reduce stress on the clinical and administrative workforce delivering the programme.
The tool uses an AI algorithm to try and diagnose breast cancer. The algorithm has been used on half a million scans from hospitals in Hungary, but it is new to the UK. The UK trial is using scans from ULHT and Nottingham University Hospitals NHS Trust.
The first phase is a retrospective trial where old images have been anonymised and used to see how accurate MIA is at diagnosing scans that need further investigation, compared to the results produced by the breast screening reading team. Already, it is performing better than most humans.
If the evidence shows that it is safe to do so, then the next stage will see the team use MIA to do the first read of all scans before they are then reviewed by a member of the radiology team and the results compared. If there is any difference of opinion then the scan will automatically be sent for a third read.
ULHT Consultant Mammographer and the Trust’s lead on the project, Bernadette Trzcinski, said: “I am really excited to be working on this trial, which may revolutionise how we read scans in the future.”
“Across the country we desperately need something to help us with the current staff shortages, which are predicted to become increasingly challenging as the demand for imaging grows. The success of this project will transform the breast screening service, improving both quality and efficiency for our breast screening population.”
“It is not about replacing radiologists. All scans at the Trust will continue to be read by at least one member of the breast screening reading team. However if MIA is successful, it has the potential to half the amount of time we spend reviewing scans, this is time we could be spending with our patients, improving their overall experience.”
Time matters
East Suffolk and North Essex Foundation Trust (ESNEFT) has embraced a philosophy of Time Matters, seeking to focus all use of employee time on improving patient outcomes. The trust has launched an intelligent automation strategy to eliminate the time staff spend on repetitive tasks and instead increase the time they spend on patient care.
The programme had three specific objectives:
- To streamline GP referral processes within ESNEFT to free up time amongst medical secretaries to focus more on patient care, to reduce costs and allow patients to be seen sooner by clinical staff
- To reduce the rate of missed appointments (DNA – Did not Attend) at Colchester Hospital (part of ESNEFT) to reduce the time and financial costs of missed appointments, ensure more patients can be seen by clinical staff and improve patient care altogether
- At a broader level, to prove the benefits of Intelligent Automation for the Trust itself and all of its staff, and to establish a positive culture of automation across the workforce, to smooth the path to further automation in the future to meet key commercial and patient-focused
Every week the Trust receives up to 2,000 requests for consultations with clinicians through the National GP e-referrals process. Disparate documents and patient data were previously downloaded manually from several systems before being recompiled into a new file, uploaded and sent to a consultant for review.
The DNA rate at Colchester Hospital was marginally over the NHS 5% KPI standard and the Trust needed to address this as a matter of urgency.
Within the first three months of the GP referrals automation programme:
- More than 500 hours of medical secretaries’ time has been redirected to direct patient contact
- Referrals now processed immediately 24/7
- Referral processing time cut from 25 minutes to 5 minutes
- Savings of £220K
- Increased job satisfaction for medical secretaries
Within the first 8 weeks of the DNA automation programme:
- Virtual workers have prevented 1,356 appointments from being missed
- Colchester hospital has saved £216,960 from being wasted through missed appointments
- The estimated cost of an outpatient appointment in Colchester is £160, so ESNEFT is likely to avoid wasting £1.5 million over 12 months
- Freed up vital outpatient capacity for the booking team to actively manage so patients are now being seen more quickly by clinical staff, meaning better patient care, an instant improvement in clinical utilisation and reduced waiting times. The service will have a significant impact on the ESNEFT Patient Tracking list (PTL) performance and the Trust’s DNA rate.
Darren Atkins: Deputy ICT Director, ESNEFT “We’re delighted with the results we’ve realised so far and are hugely excited about the potential benefits of automating more processes across our trust. When you look at the time and cost savings we’ve already banked across just two specific areas of our operations, you start to get an idea of how intelligent automation can drive transformation on a huge scale within the NHS.”
Christine Harvey, Neurology medical secretary, ESNEFT “It took a lot of man hours [before the robots], so it’s the time that we have saved – and paper. It gives you more time to be doing all the other things you have got to do.”
The Trust recognised the opportunity for intelligent automation to absorb time spent on repetitive manual tasks, so that employees across the Trust could dedicate more time to activities that have a direct impact on patient experience and care.
The trust implemented automation technology to deploy ‘virtual workers’ straight from the cloud in order to avoid a lengthy and costly investment in building up a systems infrastructure dedicated to sustaining the automation soft. The technology had to be easy enough to learn how to use in order to build up internal skills on creating and managing a virtual workforce. Virtual Workers needed to be able to access many kinds of systems (legacy and current) and handle unstructured as well as structured data in order to be flexible enough to automate a wide range of tasks across the organisation.
The trust uses Thoughtonomy’s SaaS based platform to access sophisticated AI and RPA technology and train a pool of ‘virtual workers’ to assist its human workforce across the Trust.
Darren Atkins “I chose GP referrals process because it was high in volume. It’s a very segmented process in terms of the different systems that our medical secretaries and consultants need to access. And as you know, consultant time is very precious, as is medical secretaries. We don’t want to waste time logging into different systems. Also this was a paper dominant process and we wanted to change that.”
Virtual Workers now actively monitor incoming eRS referrals from GP patient appointments in real- time, 24 hours a day. As soon as a referral is received, Virtual Workers trigger the automation process and begin processing. Using intelligent optical character recognition (iOCR) capabilities, the Virtual Worker extracts the reason for referral, referral data and supporting clinical information and merges the information into a single PDF document. The Virtual Worker is provided with highly secure access to upload the PDF into the Trust’s administrative systems using virtual smart card technology. It updates all systems, including Kainos Evolve, instantaneously and extracts critical information, which it passes on to the lead consultant for review and grading.
As a result of the programme, the Trust has reduced the time taken to process referral from 20 minutes to just 5 minutes. The Virtual Workers have also released medical secretaries’ time to spend time with patients on the phone, set their appointments and provide advice. In the first 3 months of launch, across 5 clinical specialities ESNEFT released 500 hours of time, reduced spend on agency staff, cut paper usage and increase job satisfaction for admin staff.
On the DNA automation project, ESNEFT deployed the Thoughtonomy Virtual Workforce® platform to reduce the number of missed outpatient appointments at Colchester Hospital. It took ESNEFT’s robotics and integration developer just four weeks to design and put into production virtual workers that can cancel appointments automatically in the hospital’s patient records system.
All outpatients are sent a text message in advance of their appointment which gives them the option to cancel simply by responding with a quick text message. The virtual worker is notified of any cancellation request by the Trust’s text reminder service and can then search for the patient on the Trust’s System C Medway Patient Administration System and cancel the appointment – just like a human would – and notify the Patient Contact Centre at Colchester Hospital. The free appointment can then immediately be reallocated to another patient.
Cancellations are now being accurately recorded in the hospital systems, providing hospital management with a far more accurate and real-time view of DNA performance.
