What I Wish I Knew About RPA: A Letter from a Former Revenue Cycle Specialist 

I worked in revenue cycle for nearly ten years, and I never once heard the term “RPA,” nor did I even know what that was. When I started at CampTek and began reading numerous articles and case studies, I wondered why I had never heard that term before. Revenue cycle and RPA should go hand in hand because I know firsthand that revenue cycle burnout is real. 

I know there can be many reasons why the healthcare system is apprehensive about RPA, and a major one I have heard a lot is “it has worked for years.” It has, but the revenue cycle is ever-changing, and it is imperative to make changes to adapt to that. I recently read an article from Becker’s Hospital Review about the top 10 processes providers said are getting most regularly backlogged, many of which were part of my old job and could benefit from RPA. These processes are time-consuming tasks that one could argue have been working for years. But they have always been time-consuming tasks that anyone in revenue cycle would appreciate assistance with. 

Challenge: A large Mid-West Healthcare provider with fifteen hospitals across two states had a large document processing problem: importing Patient Consent Forms into OnBase. This is generally not an issue for most DU (Document Understanding) tools because most of the data is structured. In this case these particular documents included a stamp or label of patient demographics added by hand and placed on the document with both structured and unstructured data. This information on these labels then needed to be read and indexed into Epic. The challenge was that these labels were being placed on the documents by clinicians and in most cases were not always positioned in a way that was effectively “read” by most DU tools.

Solution:  As a trusted adviser to the client, CampTek Software brought in several DU vendors to perform a proof of concept for the provider so they could evaluate the ability of several tools to effectively read the skewed labels. Document Understanding solutions such as ABBYY, UiPath and Indico were among those reviewed.

Each product was given a sample set of documents as part of a proof of concept. From there, CampTek would analyze the ability and percentage results to present back to the client. The findings were overwhelmingly positive for Indico. Even upside down, the labels could be read immediately. For others, the test proved that if the label was more than 10% out of alignment, they had poor results. Indico immediately showed a 90% accuracy result. Indico also did not require the use of templates, like some of the other tools did.