Since the industrial revolution humans have looked for ways to speed up processes and improve productivity – essentially get more for less effort.
It wasn’t until the 1940s that the term automation was first used by an engineering manager at Ford. But it was seen long before then – the development of water wheels, pulley systems and the loom, are just a few mechanical tools used to help us produce more output without the requirement to increase human effort and time. In the industrial world we can see automation in action, picture car manufacturing lines for example. The development of the computer in the 1960s saw the start of the digital revolution and a whole new ball game began. From here we have seen the rapid evolution of software applications that can achieve automation using digital technology. Essentially software enabled us to automate repetitive, often time consuming, tasks or processes, which previously would have relied on humans interacting with a computer. But this rapid increase in the use of digital automation into all industries and business units is invisible to us.
Automation in a digital environment
Automation in the digital environment has been established since the 1990s and is tried, tested and trusted. For example, finance departments have used it to compile financial reports; recruitment agencies will use it to review CVs and pick out key criteria for choosing a CV to look at in more detail; and comparison websites are the classic example in the public domain – you put in a set of criteria then software robots search other sites so you don’t have to. These software robots take the user’s request for a product, search other vendor sites and find the cheapest available whether its – insurance, holidays, or loans.
It has been used in healthcare too – although largely by the IT teams to automate back end processes. As an example, digital automation tools have been in the toolkit of software testers to augment, or replace, the often-used test scripts keyed manually by a bank of users. Not only is the technology used to test the applications, it has supported numerous cases of data transfer where vendors are unwilling or unable to process data files.
Process automation in a connected digital environment
As technology has moved on so have the capabilities of automation software. The term we use today to describe this technology is robotic process automation (RPA).
This technology allows users to train a software based “robot” to move a mouse cursor, press buttons, type text, read from screens and images, and interact with key productivity software and existing applications.
The opportunity in health now is for end users, admin staff, and clinicians start to use and benefit from this technology.
A key message of the NHS ten-year plan is interoperability – transferring information seamlessly from one system to another. In small pockets this is being achieved through the transfer of data from standards-based data formats such as FHIR and HL7 which are being processed into front-end input and vice versa. This approach is incredibly useful where the destination solution has a need for information to be processed via rules and validation that would otherwise be missed through direct data transfer.
However, data doesn’t necessarily need rules and validation to be transferred between systems. So why isn’t RPA being used more to support the vision of the ten-year plan when it can support the joining up of our health and care systems?
Why is healthcare on the back foot?
Healthcare seems to be on the back foot when it comes to the adoption of process automation using RPA. Though physical robots have been seen in the pharmacy for around 15 years, the move to process automation across other healthcare settings seems to be hindered. Is trust the barrier, or is there a lack of awareness of the capability, and the availability of the technology?
We all understand the physical, but we can’t visualise it so easily in the digital world. People ask: how can you trust a robot not to make a mistake? The answer – humans are more likely to make a mistake. People are concerned that robots misinterpret information whereas a human has the ability to understand what is there and contextualise. There will always be cases where we need humans to interpret data – can software interpret different handwriting styles? Not all, not yet, but great strides are being made in this area.
But a repetitive task that might take a human over forty hours of a working week to complete could be completed within a little over two full days through RPA. Humans need sleep, they need breaks – an automated task can run 24/7 without a break and multiple instances of these automated tasks can be executing the same processes concurrently, to achieve capacity. And automation will do it consistently every time. Each of us, with our individual nuances and experiences will read a manual and inevitably interpret it differently. The quality of output from human interaction is dependent on this interpretation. Quality of outputs from automated processes are not subject to variances in interpretation and therefore will be consistent all the time.
Another issue for people is that it’s invisible; you can’t see software automation in the same way we can see automation in manufacturing. There is nothing tangible except the process being completed with higher accuracy, speed and potentially with less human input.
Uses cases in health
RPA in the NHS has been offering demonstrably positive outcomes for many areas. Some examples are:
- Automating areas of appointment management has allowed trusts to process cancelled slots in a shorter timeframe, enabling patients who are available to have their appointments brought forward.
- Another example is the processing of deceased updates into local systems for patients who are no longer in the catchment area supported by the service thus reducing the risk of inappropriate contact.
- And finally automating the selective updates of TRUD information into key systems and providing a top-up service between releases.
And these are just the tip of the iceberg. At such an early stage of adoption, awareness of the capabilities of RPA is important whilst we grow more comfortable with it.