Speaking of work units, Dr. Avery J. Knapp Jr. dreams of a different multi-parametric AI work unit for radiology, arguing that wRVUs (and even time-based units) just don’t cut it:
There are at least 30 factors that affect how hard a study is to read. wRVU captures approximately one of them, poorly.
The RUC won’t fix this. It’s the RUC’s job to maintain the current system, not replace it.
PE won’t fix this. PE profits from the information asymmetry between what a study is worth and what a radiologist gets paid.
I wrote about e(ffort)RVUs last year as well. Whether you agree with his first draft of a proposal, it makes useful points for discussion.
He argues that private practices are the best suited to develop an ideal system:
So who builds it?
Radiologists. Independent groups. People who actually sit in front of a PACS and know that an 80-year-old’s postop MRI and a 30-year-old’s screening study are not the same job.
Just turning the dials in a raw time-based fashion for a CPT code creates winners and losers in an ultimately unsatisfying way. Improvement on RVUs or not, it certainly doesn’t capture the whole picture. Those who were/are frustrated to see their favorite exams changed in a system like TBWU might be happier if the hard exams paid better and were counted fairly, and the easy exams were too, and they practiced in a system where those reading complicated cases weren’t punished for doing yeoman’s work.
We all know that there are hard and easy versions of everything (not to mention people can typically still cherry-pick with the tried and true method of previewing a case and then skipping it if it looks bad). So, no, it doesn’t always balance out in an average. And even if the time to interpret a scan across 52 million exams was a neat bell curve, it doesn’t mean every radiologist’s workload will be right in the middle. RP for example says they removed “outliers” from the calculation, which is arguably important to not inadvertently capture the impact of dietary fiber when a case is paused mid-read, but not all outliers are spurious either. None of this is actually easy. The need to add “incremental seconds” in their system tells us that open/sign times in voice recognition software don’t tell a complete story.
If a new system feels unfair to people—not just because of sour grapes, but because it truly doesn’t reflect the effort it takes to read exams—then it would follow that a simple time-based work unit is insufficient to capture effort and that an ideal system would likely be more complicated. Now, RP certainly doesn’t need to attempt an ideal system, and we could surmise that an AI-derived work unit would be even more of a black box, potentially more contentious, and certainly more expensive to create, and that may all be true. But, it might also work.