Automating Administrative Work Page ContentThe Opportunity Robert is a finance officer in a hospital and his job involves processing bill adjustment requests. Having performed this role for many years, he is extremely familiar with the tasks – a series of necessary but mind-numbingly repetitive steps. He felt that his time could be better spent at other higher value-added tasks.He picks up his first case of the day – a father’s request to deduct the payment from his MediSave instead of his child’s. He takes 10 minutes to complete the change; such requests are not uncommon – he gets about 20 every month. His next seven cases are more complex and he spent the next two hours completing them. Such is a typical day in the life of our colleagues in the Finance community: catalysing business operations through manual and repetitive tasks. Could technology help?The SolutionThe answer is YES! On 1 November 2019, the National University Healthcare System (NUHS), in collaboration with IHiS, went live with its first Robotic Process Automation (RPA) initiative for Outpatient Billing Adjustment.The RPA has been trained to perform what Robert would have done, and is capable of doing so at greater speed and accuracy. Robert can now focus his efforts on more complex requests. Overall, this would allow NUHS to address patients’ requests faster and more accurately.How does RPA work?RPA is a well-established term in the industry. Do not be confused by the terminology “robot” – these aren’t physical robots in the usual sense of the word, but more like a computer software that uses other software the way a human would.In simple terms, RPA is process automation made possible by software robots that replicate the tasks which humans would have undertaken; these can include running applications, initiating responses and communicating with other systems for information. By focusing on mimicking human actions, it can automate most software in a non-invasive way and with minimal disruptions.What processes are suitable for RPA?Robots follow rules. Where a procedure requires interpretation and skill in judging an outcome, a robot may not be suitable. A common technique in re-organising tasksteps is to ensure that steps requiring human judgment are dealt with collectively.In this way robots can handle bulk rules and hand-off to humans whenever human judgment is needed. The typical criteria for processes suitable for RPA are:Processes that are highly repetitive and have low cognitive requirements (i.e. not complex and do not involve subjective judgements)Transactions that occur in high volumeTasks where there is a high probability of human error (as evidenced from historical data)Processes with a limited set of exception-scenariosHow RPA helps in Central Claims Processing System (CCPS)After a patient had completed surgery for both eyes on the same day, a staff submitted his payment claims in CCPS. As CCPS is not able to differentiate between left and right eyes, it deems the submissions as duplicates and rejects them. Staff would thenhave to spend time to verify and rectify. However, RPA has been taught to handle this process and would be able to confirm that it was not a duplicate, saving staff from having to check!Implementing RPAAfter IHiS identified a suitable RPA solution, it had to map out how the solution would fit into NUHS’ systems, as well as the steps involved in processing a billing adjustment. During discussions with NUHS, IHiS realised that different staff had slightly different ways of processing a finance request. An analogy is this: everyone brushes their teeth, but no one brushes exactly the same way. As the robot could only be taught one way, IHiS and NUHS did a deep dive together to map out the work flow to find the most efficient way, before translating it into computer code to teach the robot. IHiS also helped to pitch the RPA programme to the Healthcare Digitalisation Committee for funding for this project. RPA development would contain many reusable components that would be sharable across the various clusters, making it easier to implement and helping them to increase their productivity in certain areas. In subsequent phases of the RPA roll-out, NUHS will also be automating the structured error claims in the Central Claims Processing System (CCPS) and Medical Claims Proration System (MCPS), as well as the reconciliation of counter collection.Savings and ResultsRPA is gaining popularity internationally and NUHS is believed to be among the first in Southeast Asia to implement it in the public healthcare sector. NUHS estimates that RPA would free up the equivalent of eight full-time employees to perform those functions annually. This not only helps them resolve the manpower crunch, but also allows their staff to focus on higher-value work.