This website is best viewed in portrait mode.
2 Ways to Optimize the use of RPA in Healthcare
“Process is king, a multiplier that turns the human plus machine into a transformative advantage.” - Garry Kasparov, in the foreword to Hyperautomation
In our previous blog post, we outlined the uses of Robotic Process Automation (RPA) in healthcare, and how RPA is revolutionizing healthcare in multiple aspects.
For all its benefits, RPA has a significant drawback: it's dumb. It is based on the simple principle of “if this, then that”. So it cannot learn from its environment, which creates a natural constraint for use cases. Where and how can you best use what is effectively a low-level mimic? Additionally, human intervention may be required to correct the mistake if the bot encounters an exception.
This concern is particularly salient for healthcare. No two patients are the same, nor are they affected by an ailment in exactly the same way. The same individuality extends to drugs and medical devices as well. Large healthcare organizations cite knowledge and resource crunches as a major reason for avoiding automation. The reverse is also true. The program simply doesn’t know enough to assist with each case on its own merits.
So if RPA is to be deployed in ways that go beyond rule-based support, it has to blur the line with AI. This entails listening to conversations and processing them for value.
Let us consider a couple of the ongoing attempts to improve the use of RPA in healthcare.
Hyperautomation is arguably the hottest trend in automation at present. In its simplest terms, it is a broad-based strategy that combines individual RPA solutions with other technologies like AI and analytics to automate complex processes. This is salient for businesses that started by automating individual tasks separately but found it hard to maintain each process in its silo. A mature automation strategy should take the bigger picture into account, and for that, the program requires a robust grasp of how the business functions.
Last year, Gartner went so far as to call hyperautomation “a condition of survival” in an RPA webinar. They had previously listed it among their Top 10 Strategic Technology Trends for 2020, which made much of the industry sit up and take notice.
In an era where medical technology devices are constantly pushing the state of the art, the use cases for hyperautomation are higher than ever. Medtech now involves software testing, hardware maintenance, documentation harmonization. Therefore, the idea of an integrated and proactive approach makes sense which is essentially hyperautomation.
To cite one example, unstructured data accounts for almost 80% of enterprise data, making it hard for MedTech firms to accurately organize or use their data. This is particularly crucial in an industry that relies heavily on percentages and fine margins. A strategy that combines NLP-based learning with regular RPA will help structure this data and enable the proper conclusions. A Deloitte study broadly pointed to the three areas where a Medtech company could benefit from automation. Control, Compliance, and Capacity enormously boost efficiency and process quality across the company. And that’s before we take into account the standard internal functions of any organization – Finance, HR, etc., which can also be looped into the same hyperautomation toolbox.
Another fusion of RPA and AI is termed Intelligent Process Automation (IPA). If hyperautomation is a toolbox, IPA is an advanced tool, one that eliminates many of the deficiencies of traditional RPA. For context, if RPA is the autopilot, IPA is closer to being the driver. It combines RPA with:
- Natural Language Generation (NLG) to convert data inputs to written text
- Machine Learning (ML) to help detect patterns in structured data
- Cognitive agents to bring the former two together to create a ‘virtual workforce'
- Smart workflow to integrate different stages of the process, human and robotic, smoothly
Pharmaceutical businesses might be especially interested in this technology. Take for example its effects on pharmacovigilance, which were researched in February 2021 by professionals across pharmaceutical MNCs. Their conclusions were broadly optimistic.
“Intelligent automation can contribute to the quality and consistency of case processing and assessment, leading to a timely assessment of safety signals,” they wrote in the study. “The potential advantages of AI-based systems are that more complex associations and patterns sometimes unknown to humans can be considered by a model, resulting in greater flexibility and ability to handle variable input (e.g., NLP for the handling of unstructured text).”
The authors also drew an interesting distinction between ‘static’ and ‘dynamic’ systems. Existing validation methodologies tend to be rule-based and static, i.e. they depend on batch-learning techniques that teach them the entire dataset at one go, as opposed to ‘dynamic’ AI-based systems, which continuously absorb data to update their models. The latter might seem more cutting-edge, but might also be less stable. As one might expect, the quality of the input data plays a crucial role in the accuracy of the outcome.
Pharmaceutical businesses may also find a use case in the ongoing, FDA-mandated migration from Computer System Validation (CSV) To Computer Software Assurance (CSA). This means a shift in emphasis. The focus is now on critical thinking and testing, rather than the associated documentation. More time is spent testing each drug according to its own merits, which means more human labor. This is just the kind of problem that intelligent automation is meant to solve.
Finally, there is the provider perspective. The uses of RPA in reducing the administrative workload. Patient appointments, follow-up bookings, treatment plans, etc. have already been documented in our previous RPA blog post. But AI can take it a step further. As Robotics and Automation News puts it, “Adding a layer of artificial intelligence and intelligent automation can improve patient interaction and improve appointment follow-throughs without increasing manual hours” through chat-enabled, AI-driven RPA bots.
It is easy to get carried away by the boundless possibilities of new technology, hence a note of caution is in order. These trends are still evolving, and their full scope for business isn’t fully clear yet. And RPA usage is highly asymmetric across industries. Some companies are putting their first RPA systems in place, others are already integrating ingestion engines. There is also the possibility that the current fervor will cool off once the pandemic abates, and organizations will try and revert to a status quo ante they are more comfortable with.
However, trends like IPA and Hyperautomation emerged before the pandemic, and they seem to hold the long-term potential to fundamentally restructure an organization.
The healthcare sector’s much-lamented slowness to digitize may prove advantageous in this context. More automated sectors have to invest in upgrades and rework around pre-existing ‘technical debt’. However, medical businesses can directly leapfrog to advanced RPA. They can bypass the intermediate stages, avoiding the legacy problems that are now surfacing in early adopters.
Examples from other industries may illustrate this point. The likes of Rakuten in Japan and Jio Telecom in India benefited tremendously from being late starters. Jio in particular built its entire operation on 4G, overtaking better-established rivals who took years to upgrade from 2G. As RPA continues to evolve, it may just give the healthcare sector its biggest technical leap in decades.
Authored by: Mangesh Deshpande and Karthikeya Ramesh
Note: The opinions that may be presented in the article are that of the authors