The HSE Digital Academy in partnership with DPER and hosted by Deloitte, held a symposium on January 30th 2020 to showcase Robotic Process Automation (RPA) and its potential in automating administrative work and freeing up workers to do higher value added work.
Below is a summary of the event which included several NHS leaders presenting case studies and key learning’s from the application of RPA in the UK Health Service as well as a pilot case study from HBS.
Robotic Process Automation Symposium
January 30th, Deloitte, Dublin
Suzanne McDonald, Health and social care lead with Deloitte opened the HSE’s Robotic Process Automation (RPA) symposium. She explained that Deloitte are partners with the public sector for RPA and she welcomed the opportunity to raise awareness and develop new views on RPA. Suzanne thanked Martin Curley and his team for providing the inspiration which was behind the event and then introduced the line of speakers and gave a brief outline of their expertise and what they would be discussing.
Suzanne then outlined three key objectives she hoped the day would deliver
The first was to take the mystery out of RPA, while there are hype and advertising regarding process automation she felt it was important to see beyond the hype and understand exactly how RPA works. This would demystify RPA and make people feel more comfortable around what it means for them, allowing them an opportunity to look at RPA and decide how it could be brought forward within the HSE.
The second objective was to use a series of exercises during the day to determine which areas within the business might be suitable candidates for RPA. For attendees, this would allow them to determine where the most appropriate use of their time and effort might be. The hope being that people might have identified candidate process’s within their area before the day concluded.
The third and final objective was for people to learn how to get started. Suzanne advised us that by the end of the day people may have identified candidate processes but they would need guidance on how to progress towards an RPA solution and this would be provided by Deloitte during the symposium.
Dr Lucy Fallon Byrne
The first speaker was Dr Lucy Fallon Byrne from the Department of Public Expenditure and Reform.
Dr Byrne is the assistant secretary of the reform division within DPER. She revealed to us the work she is currently doing on the next 10-year plan for the Civil Service and the work she plans to do on the upcoming 10-year plan for the Public Sector. A key theme underpinning both plans is around digitizing all transactions in the public sector. Automating routine processes will be a key component of this, driving efficiencies for the next ten years.
Dr Byrne welcomed the delegates who had travelled to the symposium from the NHS and said how she looked forward to hearing from them as the NHS was more advanced in its process automation and engineering than we are in Ireland. She spoke of the great possibilities RPA offered the HSE and told of how Paul Reid was a great supporter of the RPA process.
Dr Byrne finished by outlining how innovation was a key component in RPA and in delivering greater efficiencies in the public sector. She gave details of the public service 2020 innovation fund which has just been launched. She invited people with innovative ideas to apply and informed us flyers had been left on each of the tables with all the information required to put in an application.
The next presenter to step to the podium was Phillip McGrath. Phillip works for DPER where he leads out on public sector innovation. Phillip laid out their reform plan and gave an overview of their development and innovation framework. He explained how DEPR view RPA as a tool and a lever to drive innovation. RPA also supports two of the key pillars of the reform plan namely delivering for the public and innovating for the future.
Phillip gave a brief run-through of the history of DEPR’s involvement with RPA. Initially the opportunities RPA presented for on sharing of problems and cost savings were known by DEPR, but they wanted to see how this could work with the public service in Ireland. In order to do this they sent an expression of interest out to all civil service bodies looking to see if anyone was interested in a pilot on robotic process automation. The invited interested groups to nominate processes and people DPER would train. They view the training of in-house people to becoming proficient in RPA as a key component of RPA growth in the public sector.
After reviewing the applications for RPA they picked the National shared office, People Point, Public Jobs and Revenue. They used Blue Prism as their tool of choice.
They purchased a license for each of the organisations for a year, used local installation of the license rather than on the cloud. DEPR provided training; mentoring and coaching assistance and Dr Fallon Byrne oversaw the governance board for the projects. The requirement for each group was that they automate several processes.
After 12 months of the pilot project, 24 processes were in production or the final stages of development. The return on investment had been achieved and a significant saving, but what impressed DEPR most was the marked improvement in the quality of the process’s being executed. There was a massive reduction in human errors and business information had improved significantly. All steps of the process were also auditable which is another benefit.
Processing times speed increases varied from twice as fast, to twenty times faster as the old system. When processes were moved to a virtual machine in the National Shared Services office, the increase of 25 times was achieved. An example of the benefits automation was one particular process in people point, When RPA was introduced to their process, the original Queue of three weeks was brought down to seconds
Some of the main findings from the pilot Phillip detailed were:
- Exceptions rates were .3% at the lowest and 9% at the highest
- RPA is effective for appropriate processes
- Well embedded processes make better candidates
- There is a need to put a little bit of money upfront
- There is a need to agree with SME’s on uniformity on how each process should be done
- Some degree of programming and coding may be needed but not in most cases
- Start small, think big and grow slowly.
For public sector bodies looking At RPA, Phillip disclosed that trade unions have been engaged at general council level in discussions on RPA and there is nothing being hidden or delivered without full disclosure and consultation. RPA is not a panacea for all problems, but it can be a good interim solution while awaiting a bespoke solution. It can turn the process around in on average 8-12 weeks. For those looking to implement ROA they will have to invest in people’s time. DEPR are very happy to do internal consulting as will Deloitte to assess process’s suitability for RPA.
A digitization memorandum to government went in last December and on foot of that the government has decreed all public service bodies must examine their backend processes for automation using RPA and drawdown from the DEPR RPA framework. Phillip put it quite straightforwardly telling the attendees that RPA is now a requirement and not a luxury.
The current Framework for RPA was published in December 2018. Awarded in July 2019
The key components of the Framework are, Deloitte is the single provider and Blue Prism, Nice and UI Path are the RPA software Vendors. For those looking for more information a full user guide with pricing is available from DEPR
Services provided by the Framework are the supply renewal and installation of RPA licenses, RPA training and development, technical support and the provision of consultancy. Phillip welcomed anyone to call to him with any queries and he promised to answer himself or failing that, have the right people respond.
Professor Martin Curley
Professor Martin Curley spoke next about how healthcare was behind other industries in digitization and Irish healthcare arguably another five years behind the healthcare norm.
Martin situated RPA in the space of incremental innovation taking existing processes and applying RPA to them. Martin draws inspirations from reengineering guru Michael Hammer, one of the fathers of process reengineering who famously said: “Don’t automate, obliterate”. Martin explained the Health Service has process’s which will not be automated verbatim but require improvement and refinement.
The strategy being pursued, digital transformation in the HSE is anchored in the concept of small islands of coherence which will transform the whole system. Will be using RPA to invent processes as well as automate existing ones and in doing so, taking HSE from digital laggard to leader in five to seven years with RPA being a valuable asset in achieving this.
One of the most impressive things Martin found with RPA was the fidelity of process execution and reduction of errors. This aligns closely with the HSE priority of patient safety. Expanding on the safety theme Martin spoke about, one processes aligned closely with patient safety which currently has an error rate of 40-50% that has been identified as a prime candidate to be targeted with RPA. Martin is hoping to move soon from pilot projects to working on some significant work across the whole organisations
Martin next showcased the Digital Academy and some of its objectives as well as the work and training it is overseeing.
A key objective of the academy is to upskill Healthcare Staff to enable them to understand digital health and lead the HSE in digital innovation going forward.
The Masters in digital Health transformation has just been launched and unlike a normal master’s program, there is no dissertation but the student has to complete a multidisciplinary digital health transformation project. This will lead to 12 new transformation projects being launched in the HSE in the next two years.
While the Masters is the flagship project of the digital health academy, Martin listed other programs being developed including a digital health passport initiative and a digital skills proficiency programme.
In order to equip staff to harness the potential of RPA the digital academy intends to run an RPA Boot camp in collaboration with DEPR for five days in April and five days in June. More details will be announced in due course by the Digital Academy.
The digital Academy is creating a network of living labs which incorporate digital start-ups to foster innovation in healthcare and to bring the benefits of this innovation directly to the patients and healthcare workers.
The first examples of these Martin gave was the Health Beacon smart sharps bin for patients who self-inject medications at home. He revealed that use of these was found to increase adherence to prescribed drug therapy from 50 to 70% which is very encouraging. The second example was the mobile x-rays van which brings the x-rays machine to the patient rather than having patients who may be frail or elderly, travel to acute settings for their x-ray. In the pilot program 90% of the people getting x-ray are avoided having to be transferred to Emergency Department. If this system could be deployed nationally it could have a significant impact on admissions to Emergency Departments in Hospitals.
The final example was vital signs automation which uses a tablet to transmit real-time medial information from patients to a central ward station. The use of this tablet adds statistical process control to automate blood pressure, temperature and pulse rates. This enables proactively use of the early warning score system to improve patient safety.
Martin gave two dates for upcoming Digital Academy Forums for people to put in their diaries. The first will forum on March 11th will have Electronic Health Records at the center and in June the next forum will look at data and the opportunities RPA can give to the Health Service.
Martin signed off telling us we are in an early space with automation and have a journey to go to reach maturity but he sees a future where there would be a robot for every person to assist them in their role and ultimately a time where employee-led automation of processes will be as common as work done by RPA developers.
Tim Morthorst stepped up to the podium next. Tim is one of the leaders in Deloitte’s RPA team which advises all areas of the public sector.
First Tim explained that RPA is not robots in the physical sense but is software used to automate processes used on a computer. The exact definition he gave was its an enterprise software tool which automates business processes by mimicking human behavior across systems"
Explaining the definition Tim spoke of how processes could be taking data from MS Word, copying it into Excel and using the data as the basis of an email which is written and sent. The mimicking is when the robot copies some of the actions the human takes on the PC such as reading different fields, inserting text, making calculations and composing emails. The last important thing is the RPA is a fully controllable enterprise tool.
Tim showed a video which displayed an RPA in action. The RPA retrieved data from the Web, inserted the data into a spreadsheet, uploaded information into a Dashboard for viewing and composed and sent multiple emails to multiple recipients based on the data. All of this was done in less than 10 seconds.
There are a multitude of benefits to RPA and some of them outlined were:
Quality improvements: When human perform tedious processes they can make errors, robots can perform tasks quicker without the mistakes a human can make. The second achievement is security, a robot can be told to forget any data and can also be told to store other data
Traceability: The robot can record and check every step in a process
Quantity: A robot can reduce backlogs and anticipate extra workload requirements. It works 24 x 7 – and doesn’t need tea breaks, rest or sleep.
Accuracy and Scalability: A robot will execute a process without error once or 10,000 times. it does not mind how many times it has to repeat a task
Some of the process’s Tim outlined for us which were ideal candidates for RPA were patient reminders and notifications, Reconciliation (Gathering of data) and claims management.
The factors to consider when identifying process’s included is that they must be completed digitally, they must be rules-based (a robot cannot make complex cogitative decisions), they must be high volume and finally, they must have low variability.
Tim’s final advice was to look at what we don’t do or simply don’t have the time to do. He concluded telling us that the Deloitte and the Robots were ready to help.
Mathieu Webster - Transformation Lead NHS Shared Business Services
Mathieu invited us to see how the NHS was embracing RPA and gave us an overview of their journey and the lessons that had been learned. He began by telling us that every time they have automated a process, it has run quicker and with fewer errors and that within the NHS, robotic process automation was initially put forward as a cost-saving measure for back office functions.
While Mathieu worked in one NHS trust, he could see a commonality in all trusts where people were copying patient’s records, searching for appointments and spending a lot of time on other time consuming repetitive tasks. When they examined the processes, they found robots could automate many of these tasks.
When they began in 2017 they chose UI Path as their technology partner. The first wave of processes was in 2018. Initially they chose really simple jobs and tried jobs people did not like doing. An issue they had to face head on was the reality there would be job losses so they tried to take the pain out with reallocating people to different posts.
The first robot they created was called Freddie. This concept of naming the robot was to allow people to relate to the new system. Now, the staff choose the names of the newer robots. This is to get buy-in from the existing staff to robots doing the work previously done by colleagues. The human side of RPA acceptance is really important and has to be managed. With the redundancy of time-consuming processes and the people required to carry them out, the remaining staff will see empty desks and be very aware people are not working there anymore.
One of the first processes they automated did not save money nor was it expected to do so. A member of staff used to spend two hours copying spreadsheets every Monday morning as part of their job. The automation of this did not save money but made the person very happy and proved the concept to other staff.
As time moved on they learned about which tool was best for RPA. They got to have 22 processes automated about June 2018. This was great progress but led to a problem where management of the robots and the multiple processes they were running was needed. The process was paused while they implemented an Enterprise Architecture solution and built a master robot to control the other robots.
Concerning a product vendor, they chose UI Path as they are leaders in the RPA field. One of the benefits of working with UI path was the offline training they provided to staff.
Mathieu tips for a successful RPA implementation are:
- Drip feed information before it started
- Hold Webinars – Pass information down and educate staff to the benefits
- Ensure business cases are detailed and realistic
- Educate people to see what benefits they required
- Engage with IT to ensure they support the concept
- Agree on an operating model.
- Educate senior leaders, The NHS found that once the benefits were explained it was an easy sell.
When they looked at process selection they found they had 800 to choose from. They used a scientific prioritization method which involved isolating single processes and running them through a formula they came up with to determine the priority for automation.
In an example of savings in time and effort, Mathieu gave details of a Cash flow forecasting process which sends information to 270 organisations every day. For this to work, they used to have 42 people working from 5:30 – 10:00 every day. This has been automated and now one robot does all the work in under an hour. The cash flow forecasting occasionally used to contain errors, but this is not the case now it is an RPA process.
When they came across processes which were suitable but unstable they redesigned the processes and automated them. Automation has become so accepted it is now regarded as business as usual.
One word of warning Mathieu had was to have business continuity plans. Robots are not infallible and systems do fail. He concluded by advising anyone thinking of using RPA “If it’s already digital use it”
Philip Brocklehurst -Artificial Intelligence Team NHS Lothian UK.
Phillip approached the RPA issue from the problem solving perspective rather than looking at the technology. The problem they looked at was the referral process and capacity in their specialties. They had seen a 20% increase in referrals since 2015 and this resulted in a 12 week average waiting time but some patients waiting considerably longer.
They looked for a solution whereby the automated processes could increase clinician time spent with patients and also cut down waiting times. One of the very suitable processes was the referral triage process in Gastroenterology.
The original process involved a person visiting their GP, the GP making an electronic referral and a clinical specialist looked at the referral note and letter and deciding whether the person was to be sent to a clinic or be taken for a gastroscopy or colonoscopy or other tests. This final part of the process took an average of two days between referral being received and the clinician making a recommendation on scan or clinic visit.
To automate the process they introduced Artificial Intelligence and RPA to read the GP referrals and the way they did this was by inputting 20,000 historical referrals into the AI system alongside the decisions the clinical specialist had recommended. The algorithm learned to recognize patterns and the outcomes which were recommended by the specialists based on the different symptoms presented. The system was put in place and the benefits of this are a much faster patient triage. Another benefit is the system can extrapolate information once all information is now being processed and spot patterns which can be used for proactive interventions.
Phillip then hosted a question and answer session was where the practical application of a solution such as this into an Irish context was discussed.
Breakout Sessions – A series of breakout sessions were then held where attendees consider the current HSE Organization Structure in terms of RPA suitability and identified possible processes in their areas.
The group I worked with identified several potential processes in the time we held our discussions. These included, referral screening where data could be extrapolated and used to make some current labor-intensive processes redundant, immunization processes were also discussed where we currently have approximately 50,000 babies born each year receiving 15 vaccines and the recording and collating of this data is a multi-step labour intensive process. Other ideas were for automating admin tasks involved in staff levers and starters, the multiple works being done by different departments could be automated and integrated.
Darren Atkins CTO Automation and AI with East Suffolk and North Essex NHS Foundation Trust
Darren Atkins spoke next to bring us through the Journey he and his NHS trust have taken over the last few years with Artificial Intelligence and RPA.
At the outset of the journey, Darren and his team recognized the fear that was out there regarding robots and how they were going to take all our jobs and maybe even take over the world and they set to reframe the argument about introducing RPA. To mitigate against this they took a cultural approach and introduce a philosophy simply called “Time Matters”. They were able to show the savings automation produced using a metric that everyone could understand and able to show how much of it they were saving which could be spent on patient-centred activities.
This philosophy helped staff engagement and buy-in and accept robots in the workplace.
The first process automated was Invoice Processing. The robot reads emails, determines which are invoices, passes them through OCR Analysis and processes them. If there is a problem it passes the invoice back to human with a query, The AI component working with RPA ensures the robot will learn about problematic invoices and eventually not need to pass any back to humans.
Darren believed that automation is easy to do badly, but to do it well requires some more thought. In the case of their first process, it took 3 months to setup.
The next process Darren described to us was GP Referrals for Neurology. The old process involved emails coming in from GP’s initially. These were then printed with relevant files and scanned together to a PDF which was sent to a clinician who read them each Friday.
The new process involves a robot downloading all relevant files, compiling them and sending them to consultants who triage on a real-time basis. The robot does the work in less than six minutes.
Darren spoke from experience when he recommended talking to the staff on the ground to identify processes to automate. He also recommended doing it in house by you for yourselves; he was pleased that Deloitte agrees with this approach.
Concerning automation, Darren says you are only limited by your imagination. He is keen to share his experience with RPA and invited people to logon to www.somethingincredible.co.uk for more information.
Catherine Donaghue – Cross Border Directive Scheme
Catherine explained how the Cross Border Scheme works enabling EU citizens to get healthcare in EU or EZZ. They look after accessing healthcare in Europe which is primarily by means of three main ways which are the overseas department, the cross boarder directive and the treatment abroad scheme
Catherine Donaghue outlined for us the paper-intensive process they hope to automate and streamline with RPA.
She began by explaining that the treatment abroad scheme looks after access to healthcare across Europe. They support access to treatment which is specialized but not available in Ireland. Their remit is under the Cross Border Directive which allows free movement of people to access healthcare anywhere in Europe.
When a patient in Ireland visits with their GP, the GP can now refer patients to community, hospital or another EU country. Access to services may be in the public or private system depending on the person’s level of healthcare cover. It is a complex process involving many forms and covering different aspects of care including travel expenses, extra treatment further, treatment and reimbursement of expenses. The overall process is complex, time-consuming and labour intensive.
The initial process that Catherine has identified as the most suitable candidate for RPA is the establishment of eligibility of candidates through the overseas department. The Overseas department currently issues a number of documents by post to applicants and these are used to establish entitlements of eligibility for treatment abroad. Once the process is completed the applicant is issued with a document which entitles them to treatment.
RPA will be used to streamline this process and the hope and expectation is that it will result in a faster response rate for the applicant and a reduction of workload for staff in the Cross Border Directive department.
Ian Brennan: CIO Laya Healthcare.
Giving a view from the private sector, Ian Brennan told the story of the growth of RPA within Laya Healthcare. He gave advice based on his experience, for anyone else starting their journey with RPA. For the approach, he cautioned to start small. Aim to initially bite off 10% of what you plan to achieve and get proof of concept quickly. He admitted he did a lot of evangelizing to sell the message to end users but it was worth it in that he generated a sense of excitement and hope in what RPA could achieve. Like other organisations, Laya named their robots and in Ian’s case, he named them after science fiction writers. He did admit however that this caused confusion at one stage when someone noticed letters saying Isaac Asimov has reviewed case file X, and anyone that knows science fiction well knows Isaac Asimov passed away almost 30 years ago.
Ian reiterated what previous speakers had said regarding structured processes being ideal for automation. He also showed how it was possible to plug in other technologies to RPA such as Google Vision. Laya found that training dedicated RPA staff was very effective and also they discovered that bringing people with the right mindset who believed in RPA was more beneficial than looking for people with advanced technological skillsets. Other sage advice Ian gave was, choose your pilot projects carefully and notify managers well in advance that user’s manual handling processes are going to expire
Looking at the HSE from the outside Ian felt that RPA could be transformative and provided great opportunities to get a better outcome for patients. Start slowly was his concluding advice and he proposed we do this to by crawling, walking and finally running with RPA.
Kevin Kelly – HSE HBS
Kevin Kelly from HBS closed off the symposium with an overview of the pilot projects HBS are running and their plans for the future. HBS pay 91% of invoices for the HSE and process 35,000 Garda vetting requests annually. They are currently working on two pilot projects.
The first is Hospital and Income Debtors Collection and Reporting and the second is Garda Vetting for Staff. The Current Garda Vetting Process is very manually intensive and HBS have identified two areas where robotic process automation can fit in the overall process. For now, they are working on the second part of the process. This is vitally important to streamline the process as due to recent legislative changes there is going to be a requirement to re-vet all HSE staff in the future. This will result in a huge increase in transactional volume they will need to cope with. They hope the introduction of RPA will allow them to absorb this increase.
HBS can see there is a lot of opportunity for RPA to improve efficiencies and there is a lot of low hanging fruit. Kevin plans to chase the simple tasks first and establish quick wins. HBS plans to create a BOT agency and an HBS RPA Centre of Excellence.
They aim to move eventually from purely automating tasks to complete process re-engineering and while they are looking at automate to end processes, they are very aware that this is a long-term goal and change needs to be incremental.
While Kevin acknowledged that there are some jobs under threat he believes the jobs of the future are with RPA and AI and HBS are working towards facilitating these roles. He advised that HBS are keen to deliver on the challenge of ensuring that Healthcare professional are able to harness digital innovations and that these innovations allow them to spend more time with their patients. Kevin’s last word on automation was that it presented us with a huge opportunity to do things differently and get different results.
Suzanne McDonald closed off the day for Deloitte, telling us they were not sure before the event what the reaction would be to RPA but how pleased there were with the response on the day. The enthusiasm and appetite to work with RPA had come through during the breakout sessions and with the interaction with the speakers and the passion to deliver real change was palpable.
Some of the opportunities identified during the day that Deloitte were looking forward to helping with were work on report consolidation, automating communication, clinical coding and screening services. Other strong candidate areas were the Helpdesk function in IT and HR services where a lot of potential for savings were identified.
The last word was given to Martin Curley who thanked everyone who contributed to a great day full of positive energy. He especially thanked the people who had already started the journey with RPA and had so generously given their time and benefit of their experience to help others who are only learning the potential of RPA and the benefits is can bring to delivering a better health service for the people of Ireland.