Case Studies

Queue based workflows in Healthcare

How one company saw a 300% increase in employee efficiency

By Kenny Pyatt, Founder & CEO of DevOcho

As you can imagine, healthcare companies don’t like to publicly admit inefficient workflows or processes because we are dealing with healthcare for humans. We also live in a fairly litigious society. With those two ideas in mind, for this case study, I will be intentionally vague. I’ve created some screenshots that are simplified recreations designed to remove branding and change key parts of the project to ensure their proprietary processes are protected. Everything factually discussed in this case study has been publicly disclosed but I’m still being abundantly cautious to protect the interests of everyone involved.

I was privileged to lead a team that built an AI powered system which would review pathology and radiology reports in the oncology space and then prioritize them into a queue. The result of this software was incredible. The software would read through the reports looking for indications of cancer and then sort those cancer positive reports to the top of a queue for faster processing. The time to first treatment is the main variable on survivability for cancers. By being able to process and prioritize the hundreds and thousands of reports flowing through a healthcare system we were able to increase the efficiency of the Cancer Navigators by over 300%. Meaning one Navigator could support three times as many patients. The primary hero of this story was the AI but there was a secondary hero that deserved almost as much credit. The queue based workflow.

The old process for the healthcare system was to use Email and their EMR software to manage and track patients. Someone would be referred into the health system by their primary care doctor, and then the patient would go see a specialist in the system. Depending on the situation a scan would be done and then possibly a biopsy. Throughout the process, patients were waiting on back office processes and appointments. There was also the communication between doctors who are incredibly busy. Navigators helped smooth that process out by facilitating the communication and keeping the patient up-to-date, but they were working from email and spreadsheets so as you can imagine, it wasn’t super efficient and mistakes were sometimes made.

Our new custom software system moved all the pathology and radiology scan reports (the written reports) into a centralized database, regardless of whether they went in for a cancer screening or not, the reports were loaded into the same central database. This allowed our AI powered software to process everything which had the added benefit of allowing the system to find reports that were often a surprise to the patient. For example, if a person went into the ER for a broken collar bone and had a CT Scan. We could read the radiology report. If the CT scan showed something unexpected, instead of referring that patient back to their primary care doctor, the Navigator would be alerted and could reach out to the patient and schedule a follow-up with a specialist and then notify their primary care doctor. This would shorten the process and speed up the time-to-first-treatment significantly. The initial version of our software trimmed 7 days off the average. With additional process changes and software updates it was cut even shorter. IT also kept the patient in the same health system which had a financial impact for their operations as well. It was good for the patient and good for the health system.

So what did this software look like? Well that’s a trade secret. I can show you some parallel ideas that will give you the idea and perhaps inspire you to use queue based workflows in your software.