NHS England - Transformation Directorate

Understanding RPA

What is RPA?

Automation is used to refer to a cluster of technologies including Robotic Process Automation (RPA).

About ‘automation’

The term ‘automation’ describes a wide range of technologies that reduce human intervention in processes. Human intervention is reduced by predetermining decision criteria, subprocess relationships, and related actions - and embodying those predeterminations in software or machines.

Defining automation in the context of operational and business process improvement

In the context of business process improvement, automation capabilities have progressed along a continual spectrum as a variety of technologies have evolved and matured over recent decades. These technologies can be clustered into three distinct groups based on actions they enable, and the level of sophistication and degree of complexity of technical solutions used. These clusters are Robotic Process Automation (RPA), Intelligent Automation (IA) and Artificial Intelligence (AI).

Definition of automation clusters

Robotic Process Automation

RPA is a technology that enables the build, deployment, and management of software (robots) that can be programmed to emulate human actions and interact with digital systems in order to automate basic manual and repetitive tasks.

Intelligent Automation

IA refers to the integration of robotic and intelligent systems from various emerging technologies, thereby increasing the scope of automation beyond simple rule-based tasks.

Artificial Intelligence

AI is the simulation of human intelligence or cognitive processes such as problem solving, visual perception, speech recognition and decision making by the computer systems.


Increasing technology and process complexity

Automation can support and enable staff to digitise and or enhance clinical and business processes across all levels of the organisation.

RPA

Robotic Process Automation imitates activities carried out by humans. It can automate high volume, rule-based, repeatable tasks, delivered just like its human counterparts. However, RPA can only handle structured and digitised data.

Example use cases

  • Front office: patient administration, appointment scheduling
  • Middle office: operational and service management, report generation and distribution
  • Back office: corporate functions like HR and Finance, claims administration 

IA

Intelligent Automation uses more sophisticated technologies than RPA for structured decision making. It can simulate rule-based decisions to automate more complicated tasks. It mainly handles structured data, but some IA technologies can digitise unstructured data to further enable RPA. 

Example use cases

  • Front office: FAQs customer assistant - Med Sec, OP Call centre
  • Middle office: patient enrolment and eligibility, theatre scheduling
  • Back office: physician credentialing

Example technologies

  • Intelligent Content Recognition / Extraction
  • Natural Language Processing

AI

Artificial Intelligence refers to computer software with the ability to think. It allows examining of large, unstructured, varied data sets to uncover hidden patterns, trends, customer preferences and other useful data that can help inform better decisions.

Example use cases

  • Front office: patient data analysis and triage to assist referrals – eConsult, eTriage
  • Middle office: fraud detection and risk management
  • Back office: medical imaging analysis support – clinical admin of diagnostic support services

Example technologies

  • Natural Language Generation
  • Machine Learning


Unpacking RPA

RPA deep dive

RPA should always be considered as part of a wider, people-focused, transformation that will enable efficient work delivery in the NHS.

Today's technology impact

Emerging technologies used sporadically across the health and care system with ability to scale proving to be a significant challenge.

Continued struggle with volume of work vs continually increasing demand sometimes leading to poor outcomes and substandard experience.

COVID-19 driving existing backlogs, but also accelerating availability and use of technology across the sector.

The way work is delivered is beginning to change creates an opportunity for improving patient and staff experience.

Future technology impact

Digitally enabled staff using technology to improve care quality, efficiency and maximising time with patients – adding value to patient care, getting it right the first time, with the right clinician, at the right time.

Digitally engaged patients with greater autonomy over their health and wellness – personalised care and empowered patients managing their own care and care plans.

The boundaries of where work is delivered and care is provided are changing as models of care move outside of hospital and care settings; Integrated Care Systems – end to end pathways with seamless handoffs and care with the right professional.

Seamless, real-time access to information in a single view at the point of need.

Drivers for change

  • Changing expectations – connected staff and patients
  • Increased connectivity – added value
  • Availability of data – better insights


RPA is the simplest of technologies from the automation clusters

RPA is the automation of processes run today by humans. This automation is undertaken by ‘robots’ or software that mimics human actions. These are not physical robots, rather, they are simply programmes that do what they are told to do. 

RPA is a digital worker

It accesses systems and applications the same way a human does (with its own set of unique login credentials). The robots carry out processing in exactly the way they have been coded to do, defined by business rules and schedule established by process experts.

This means that robots can create reports, enter or move data on systems, update dashboards, send emails, or indeed perform entire processes in the background (such as joiners, movers, leavers, invoicing, patient bookings). The primary purpose of robots is to support humans in the workplace by taking away mundane and repetitive tasks. 

The four back-office steps of robots. Step one: the coordination software tells the robot when to act. Step two: the robots log onto systems. Step three: robots undertake processes in business systems and applications. Step four: robots produce process outputs, whatever those might be.

The scope of RPA application is expanding beyond back-office operations

The traditional scope of RPA was expected to be within mainly back-office functions like human resources, finance and accounting, though this image is now shifting. RPA is increasingly being used in other creative ways alongside other technologies such as computer vision, machine learning, and even to augment existing system capabilities where integration between applications is not possible. For example, in clinical settings robots could flag only the tests that are out of range for the GPs and consultants so that they can avoid reviewing the entirety of tests reports.

RPA in the context of screen scraping, for example, smart card controlled systems without Application Programming Interfaces (APIs)

Screen scraping is one of the capabilities RPA bots can deliver where there might not be any APIs available or are costly to implement. Traditional screen scraping tends to be fragile, needs constant changes and can sometimes require bypassing built into the security controls. NHS approved guidance is that screen scraping should be seen as a temporary solution which should be replaced by properly secured APIs once available. It must also be reviewed and approved to ensure it meets internal security standards.


Benefits of RPA

RPA excels in taking away repetitive, manual work from employees, such as scheduling activities, copying and pasting data, and booking timesheets. In addition to operational and cost efficiencies, RPA unlocks the capability of organisations by augmenting their staff. Within the context of the NHS, this will mean freeing up valuable staff time – both clinical and non-clinical, so they can focus on value adding activities that improve patient care and outcomes.

Cost reduction and ROI

Robots are cheaper, faster, available 24/7 and can improve productivity and data quality, resulting in lower operational costs and hence better value for communities. Most organisations report 20-30% cost reduction and 30-50% Return On Investment (ROI) on RPA projects.

Auditability

Robots collect information on everything they undertake, allowing for full, retroactive inspection on every transaction they have undertaken.

Productivity

Available 100% of the time 24/7 – the robots will never need to sleep, they will undertake their work whenever required, giving back time for clinical and non-clinical activities.

Decoupling growth from labour

Robots increase the capacity of organisations allowing them to do more with less/same resources, which then allow teams to tackle care backlogs faster.

Employee satisfaction

By giving robots the mundane tasks, employees focus on the things that people do best (thinking, deciding, producing, and creating). This improves staff resilience – more time to do transformational work and adopt new ways of working.

Speed

RPA undertakes tasks 4–10 times faster than a person, which means more patients can be seen in the same amount of time.

Reliability

RPA robots only do what they are told (no human errors) and will never mis-key, miscalculate or have a bad day; provided input data and business rules are correct, output data will be correct and consequently improve patient safety.

Flexibility

Robots are easy to schedule and assign to automations once they have been created. They can also be updated relatively quickly if the process requirements change, increasing responsiveness for patients.

Light touch

Robots work with existing applications and systems that an organisation has, which enable fast-tracking to digital transformation.

Reduced attrition

Better staff satisfaction results in reduced attrition across organisations. Increasingly, companies are focusing on this as a main benefit they seek from RPA.


Strengths and limitations

RPA is typically best suited for areas where process or business objectives could be outlined with simple rules.

RPA is a relatively straightforward solution which is best at highly structured actions. RPA robots can work effectively alongside humans automating manual, rules-based tasks, freeing up time for their human counterparts to do more transformational and creative work.

  • Humans

    • Creative and innovative
    • Verbal conversation
    • Subjective thought
    • Unstructured information
    • Emotion and compassion
  • RPA Robots

    • Structured work
    • Repetitive tasks
    • Accuracy and consistency
    • Logical processing
    • All-hours operation

Different types of RPA

RPA robots come in two formats, each with its own capabilities. Teams can decide on the best fit solution depending on the process/task requirements, frequency of the process, and level and frequency of human intervention required.

  • Unattended

    • Unattended robots are triggered by a specific event or are scheduled to run at a given time – for example, automated appointment reminders for patients.
    • They run in the background without impacting ongoing process performance – for example, reviewing appointment slot issues and waiting list management.
    • They are used to eliminate human input for swathes of processes – for example, waiting list report or specialty reporting.
  • Attended

    • Attended bots are triggered by a person when they want it to run – for example, linking appointments with pathology and diagnostics.
    • Increases human productivity by finishing an individual’s tasks – for example, long-term condition management of patients to track missing actions.
    • Used specifically to improve customer-patient interactions – such as patient initiated follow ups, booking / polling range for booking appointments.

Typical challenges

There are multiple practical challenges and limitations that may be faced when delivering your RPA programme.

Challenges and mitigations

Set-up isn’t quick

Typically it can be 3-5 months due to IT procedures and processes.

Mitigation: engage IT early with dedicated support from Day 1.

IT change processes

Delivering changes and updates to your RPA solution can be limited through internal change processes and timescales.

Mitigation: clarify and understand your IT change processes to avoid delays.

Process standardisation

Processes are typically more complex than at first glance and can cause delays to successful delivery.

Mitigation: be driven by data, not perception. Identify and engage process experts early and follow a lean-based methodology to reduce variation and exceptions in your existing processes.

Plan beyond POC

RPA Solutions are available on Cloud and on-Prem – with significantly differing costs and capability.

Mitigation: plan early, with a clear future vision of your solution architecture, hosting and security.

Software updates

Can cause robots to fail and impact process operations, including business critical processes.

Mitigation: define your business continuity processes for critical processes - including manual fall-back plans


Identifying opportunities

RPA can be used in conjunction with or independent of other technologies; leveraging additional technologies allows organisations to automate more complicated processes.

  • What RPA software robots can do

    • Log into any application
    • Move files and folders
    • Read and write to databases
    • Scrape data from the web
    • Connect to system APIs
    • Extract content from documents, PDFs, emails and forms
    • Open emails and attachments
    • Make calculations
  • What RPA software robots cannot do

    • Read handwritten or scanned paper documents
    • Understand, interpret or make decisions without machine learning and AI
    • Process unstructured data such as emails, images, video, audio and text
    • Work on systems, applications or websites that continually change user interface

Specific characteristics of use cases

Processes that make good candidates for RPA have some or all the following attributes outlined below. That is not to say that processes that do not possess some or all these attributes or features cannot be automated – but in those instances, project or delivery teams should proceed with caution.

Rule-based

Activities that can be performed by following well-defined rules are a good fit for RPA.

Consider adding case management or process workflow in the mix for more complex processes and decisions that require human judgement.

High volume

The higher the volume and frequency, the higher the potential for saving staff time and reducing risk and human error.

In some cases, low volume tasks can also be a good fit if there are needs for reducing human error to improved compliance and to manage risks.

Low exceptions

Tasks with limited variation and fewer exceptions are a great fit.

Consider case management or enterprise workflow solutions for dynamic processes.

Stable and well-defined process

Tasks that mature and stay relatively unchanged are a great fit.

Processes that change often require changing the RPA scripts. The resulting overhead may defeat the purpose of automation.

Low system change

Processes that require limited or no changes to existing systems are a good fit.

If the underlying system needs change, then it defeats the purpose of automation.

Structured data and readable electronic inputs

Tasks that require working with structured data and readable electronic inputs (Excel, Word, PDFs) are a good fit.

Consider adding OCR (optical character recognition) and other IA/AI technologies to the mix if the data is unstructured or in a format that is not readable, such as images.

Prioritising opportunities

Typically, teams will identify multiple RPA opportunities that will require prioritisation. An example framework is described to align opportunities with organisational priorities.

The four-boxed framework diagram describes the outcome of high and low solution complexity, high and low operational value, and high and low operational efficiency. Low complexity and repetitive admin tasks take staff away from patient care. To deliver a well-defined and validated benefits case which outlines clinical and non-clinical operational impact, identifying the high volume, low complexity, high value opportunities will be imperative. RPA initiatives also require the right ownership and engagement from key stakeholders at the right levels. Immediate wins can be achieved by going for volume.
About the framework
  • Low complexity and repetitive admin tasks take staff away from patient care. Being able to give that time back in both clinical and non-clinical settings can have significant operational impact and can improve patient care and outcomes.
  • Depending on what stage an RPA programme is in, setting-up costs might require significant investment. Hence, it is critical that there is a well-defined and validated benefits case which outlines clinical and non-clinical operational impact. To deliver that, identifying the high volume, low complexity, high value opportunities will be imperative in driving the success of the programme.
  • Like any other transformation programmes, RPA initiatives also require the right ownership and engagement from key stakeholders at the right levels. Right type of opportunities that enable successful delivery of the benefits case are the key to driving the success of the programme and gaining stakeholder buy-in. If in doubt, go for volume. 
  • Projects and programme teams will need to work hand in hand with clinical and non–clinical functions to blend the functional expertise with technical knowledge to identify the right opportunities.


Current use cases

Applications of RPA in the NHS today within front, middle and back office are summarised below.

Front office

Appointments: freeing up and matching capacity

  • Rostering
  • DNA/CNA – reminders and rebooking
  • Bed management

Supplementing clinical judgement*

  • Imaging
  • Case note change tracking
  • Coding discharge letters
  • Key controls (for example – end-of-life, drug seeking, child protection)

Transition between different care providers

  • Out-of-hours
  • Anticipatory care plans*
  • Medication mapping / reconciliation
  • Lab results and blood tests

Improving data quality

  • Patient records – registrations updates and reconciliation
  • Patient record analysis for proactive care*
  • Immunisation records
  • Clinical evaluation forms

Note: *based on scope and output of the use case, Medical device regulations might be applicable.

Middle office

Strategy and planning: analysis of reports, legislation and contracts

  • Used alongside cognitive technologies to help with:
  • Report intelligence
  • Compliant clauses
  • Contract leakage

Budgets and reporting

Used alongside AI and analytics to better manage:

  • Gathering, cleaning, processing and interpreting data
  • Predictive budgeting and forecasting
  • Approval workflows

Risk management

Used alongside Intelligent Automation to improve:

  • Risk factor monitoring
  • Counter fraud processes
  • Decision making

Programmes and projects

  • Monitoring and responding to data to drive triggers
  • Automated report generation and distribution
  • Support to real time analytics
Back office

Human resources 

  • Joiners and leavers (account creation/privileges) 
  • Temporary staff management
  • Employee information maintenance 

Finance and accounting (including payroll) 

  • Accounts payable & invoicing 
  • Reconciliation 
  • Operational cost management
  • Reduced approval times

Procurement and supply chain

  • Automated sign off and approval workflow
  • Order confirmation
  • Supply replenishment and inventory control
  • Supplier performance (fulfilment)
  • Inventory management

Informatics and reporting

  • Monitoring and responding to data to drive triggers
  • Automated report generation and distribution
  • Support to real time analytics

Mythbusters

Breaking down the most commonly held misconceptions about RPA:

Robots are humanoids that will replace humans

Robotic Process Automation, or robots, or bots, refers to software code that can perform manual processes and create streamlined workflows. RPA only automates repetitive human activities freeing them from rote administrative work to focus on innovative and creative aspects of their work.

Robots are perfect

Bots can eliminate human errors and greatly reduce noise in statistics, however they are only as good as the information that is put into them. Bots are programmed to execute the formulas they are fed and as such, if there are errors in the logic of their code, they will continue to replicate those errors indefinitely.

RPA is expensive

One of the biggest advantages of deploying RPA is instant results and quicker ROI compared to other transformation initiatives. The other important thing to remember is that RPA does not require replacing existing systems, instead it adds automation to existing systems to mimic human behaviour.

Automation is owned and driven by IT

RPA programmes are a collaborative effort between business and IT. IT support is required to provide necessary resources and oversight for the platform, applications and RPA software to function smoothly

RPA is an AI virtual assistant, specialised to do a narrow set of tasks

RPA is rule-based algorithms which can be used to capture, process and interpret streams of data, trigger appropriate responses and communicate with other processes. However, it cannot learn on its own – a key trait of an AI system.

To use the RPA software, one needs to have basic programming skills

For business and functional users, there are no programming skills required. They just need to understand characteristics of processes that fit automation. Although, now there is low code software than can allow these users to build their own automation without any requirement of programming skills.

RPA is fully automated and doesn’t require human supervision

Humans are indeed required to programme the RPA bots, to feed them tasks for automation and to manage them. There’s also the efficiency factor which comes into play – the RPA systems are fast, and almost completely avoid faults in the system or the process that are otherwise caused due to human error.