Sharing our learning on Robotic Process Automation

The purpose of this report

This report collates the learning we have gathered about Robotic Process Automation technology, implementation and benefits, as an emerging technology that appears to have the potential to make a significant contribution to releasing NHS capacity.

It complements the national RPA guidance document issued by the NHS Transformation Directorate’s Digital Productivity Programme.

Who is it for?

This report is aimed at NHS senior leaders, board members, digital leads and transformation leads.

You can browse the report online or download a PDF copy of the full report here.

What is Robotic Process Automation?

Top tips for writing an RPA business case

How is RPA being used in the NHS?

Case study 1:

Outpatients

Case study 2:

Migration of data

Case study 3:

Acute trust

Case study 4:

Referral-to-treat cancer pathway

Case study 5:

HR services

Case study 6:

Finance

Case study 7:

Primary care

Lessons learned

References, resources and contacts

What is Robotic Process Automation?

Robotic Process Automation is simply a computer programme, or script, that runs in the background to automate a task or set of tasks that had previously been done by a person through a series of keystrokes.

The computer programme code runs in the cloud, often referred to as the robot, or ‘bot’. The computer programme is sometimes called ‘Process Automation’, which removes the misconception that robots are an inherent part of this technology.

In the NHS, RPA can be used in clinical and non-clinical settings to automate administrative tasks performed by clinicians or non-clinicians. A number of real example case studies are described below.

Automation can be enhanced with other features, such as character recognition and translation services.

Is RPA safe?

RPA is a computer programme and will therefore run a set of instructions and then stop. It has no ability to action anything that isn’t written into the computer programme, and RPA scripts will include instructions on what to do should certain conditions be met. It is therefore specified and controlled and would be implemented following testing and validation.

For example, automating the finance process of emailing debt reminders to customers would require the following steps as illustrated below:

This script could be set to run monthly. It would not need any human intervention unless there were system changes, such as a new password for the report system, or a requirement for the final report to be sent to a different email address. In this scenario, the script would be paused, updated, tested and validated and then re-activated.

An automated process may require a login and password, which would be securely stored in line with NHS cybersecurity and Information Governance requirements.

As RPA scripts don’t involve humans, there is no human error. The script will run repeatedly as required. It doesn’t fatigue, get distracted or suffer from boredom – factors that can lead to human error.

What are the benefits of RPA?

RPA enables departments to increase capacity without recruiting additional staff.

RPA releases staff time by automating their administrative tasks. These staff are then available to carry out higher value tasks to offer improved services or fill other vacancies.
Releasing time from a clinician by automating their repetitive administrative tasks, creates clinical capacity without employing additional clinical time.
Releasing administrator time by automating their repetitive administrative tasks can help with staff retention as staff can be developed to move into more satisfying roles.
As increasing numbers of processes are automated, some roles may not be required, resulting in departmental efficiency and cost reduction.
RPA performs tasks much faster than humans, and can scale up to high volumes at pace.
RPA can run processes 24/7 and overnight, so work is not limited to office hours.

RPA can improve safety and assurance as it reduces human error from automated processes by removing the human factor from the task.

Additional benefits can be seen in terms of supporting the workforce through reducing the burden of administrative tasks, improving patient outcomes from increased capacity as clinical and non-clinical staff have more time for patient-facing activities, and waiting times may be reduced or prioritised.

Those who have implemented RPA have also reported benefits from working collaboratively, including reviewing the collective effectiveness of existing processes.

What are the costs?

To futureproof this document, costs have not been included. However, an investment in terms of finance and time is needed to benefit from RPA. This includes:

Procuring the cloud-based technology (robots, or bots) from a vendor. There are a number of options. Open-source technology is free of charge.
Training or procuring a technical developer or team to work with your business leads to identify processes to automate, and write and implement the automation code.
Time from organisational and business leads in engaging staff to prepare for automation, communicate the case, create an improved vision and plan for the business that automation will enable.
Time within the department to manage the business transformation processes of implementing this technology, realising the benefits and creating new pathways and workflows.

What’s the difference between RPA, artificial intelligence and machine learning?

RPA replicates computer processes that were previously carried out by a person using a keyboard and mouse. An RPA programme carries out a set of instructions and repeats these as often as scheduled. It does not learn or make any judgement-based decisions. A poorly designed automated process does not have the ability to self-improve.

Artificial intelligence (AI) is an umbrella term that describes technology which enables a machine to mimic human cognitive functions, such as learning and problem-solving, using maths and logic to simulate the reasoning that people use to learn from new information and make decisions.

Machine learning is a subset of AI which allows a machine to automatically learn from past data without direct instruction. This enables a computer system to continue learning and improving on its own, based on experience.

Top tips for writing an RPA business case

Provide clarity on key drivers for the business case. These could be financial, operational, patient outcomes, staff wellbeing and mental health or all of the above.
Align your business case to national, regional and local strategy. Include any specific issues that need to be solved.
Build a roadmap for future processes to demonstrate vision for change and the strategic approach across various departments.
Give evidence of achievement from other NHS organisations using RPA.
Ensure the ‘total cost of ownership’ is captured.
Clearly define all the clinical and non-clinical benefits.
Ensure the financial return on investment is realistic and achievable; consider if this is cash releasing, cost avoidance or time released with associated financial benefits.
Include specific processes and metrics that will be monitored to demonstrate the return on investment.
Access support and advice from the RPA community as needed.

Case studies: how is RPA being used in the NHS?

RPA is used in different settings across the NHS. The seven case studies that follow describe some of these.

The range of applications gives an indication of the actual and huge potential this technology has to ‘release time’ from clinical and non-clinical roles at a time of enormous workforce pressures.

Read our RPA case studies:

You can also download the full report as a PDF to read all seven case studies.

Lessons learned

Top tips and challenges in implementing RPA, kindly donated and collated from those who have been there and got the T-shirt.

Early engagement

Identify your key stakeholders in the business including service leads, senior management and potential funders. Engage with them early and as often as possible. An RPA implementation will succeed when those affected understand why it is being introduced and play an active role to make it work with local processes.
Work with the willing; don’t underestimate the power of champions.
Involve those who will be affected at all levels including staff whose job roles are likely to change as their tasks are automated; reassure them of job security early and repeatedly.
Engage early with third parties for permission to open up their systems to bots, if needed.

Change management

RPA is a change management tool, an enabler of change. Ensure that implementation is made with both an operation and IT mindset.
Articulate a clear vision of the overall change objectives for the service without focusing on the IT or the ‘time saved’.
Engage staff with co-creating the vision and what additional value they can add with their ‘released’ time.

Implementation approach:

Think about sequencing your automation changes. Starting small to instil confidence may reap more benefits in the longer-term than going for the ‘biggest bang for your buck’. Other considerations will be volume, impact, time saved, and where the process sits in the overall workflow.
Break larger processes up into sub-processes to simplify automation implementation and roll-out. Bear in mind there will be process that cannot be automated, which require manual input. Creating exceptions is realistic and pragmatic. Full automation may not be possible; striving for it may result in missing simpler but impactful opportunities.
Don’t blame the bots! They just do as they’re told.
Once your RPA implementation is understood and underway, demand for automation may increase or even snowball. Create a method to approve and schedule new automations to manage expectations. Bear in mind that it can be a challenge to know where to stop; the possibilities are potentially endless!
It is immensely satisfying when the first process(es) go live.

Identifying processes to automate:

Factor in time to hold stakeholder workshop(s) to discuss as a team which processes could be automated, and process-map these.
Use the change process to review existing processes before automating them – there may be room for improvement. Don’t automate broken processes.
Consider using RPA to automate processes that should be done but no-one has time to do.

Metrics

Time-released metrics are relatively easy to track. Tracking other metrics robustly and regularly can be challenging. Clearly identify what you will measure and how; this should or could link to benefits identified in your business case.
Obtaining pure cash releasing savings can be challenging, particularly in the face of high service demand as there will likely to be pressure to re-invest monies saved within the service, rather than ‘release’ it.

Collaborate and share learning:

Engage with people across the wider NHS ecosystem who are considering or implementing RPA to bounce ideas, access guidance, and share experience and learning.
The limitations of implementation are only held back by imagination in applying the technology.

Primary care:

GP practices work in very different ways to each other. When working across a group of practices, focus on a common outcome not practice-specific processes, and identify a standard start and stop point that can interface with every practice.

How can Health Innovation West of England help?

We’d love to hear from you if you would like any further information about RPA.

Please contact us at weahsn.innovation@nhs.net.