The use of technology and data in healthcare decision-making is on the rise. Unfortunately, most systems think like computers, not clinicians. However, machine learning can deliver insights that rules-based prioritization cannot. Discussing this topic, Iodine Software’s Chief Data Scientist Lance Eason and Chief Revenue Officer Troy Wasilefsky joined host of Iodine Intelligence, Hilary Kennedy.

First, Eason explained some of the issues with revenue cycle management, such as not recognizing the clinical cycle. Because hospitals earn revenue via quality of care — not selling products — administrators need to analyze what’s clinically happening with the patient to address the problem.

“Most software is built for workflows, which causes revenue and quality indicator leakage. What it should actually do is solve the underlying problems,” Wasilefsky added.

One example of this is documentation integrity, or lack thereof, while patients are in the hospital, and machine learning could help fill that gap.

The challenge is that the physicians provide context to improve the quality of records while machine learning simplifies it by pre-populating queries with a high degree of confidence that there’s a correction opportunity. Iodine Software’s solution involves sending the notification directly to the doctor’s phone or EHR record.

When comparing rules-based approaches to machine learning, there are clear benefits for the latter. Rules-based has even more issues because it can’t collect the nuances of diagnosis, as diseases can present differently in patients. However, as Eason explained, “machine learning is good at picking up on subtle patterns and looking at multiple different patterns.”

Subscribe to the Iodine Intelligence podcast and visit https://iodinesoftware.com to learn more.