I posted that data Friday night. Since many people see this blog as a news feed, I'm letting you know that I have redone the analysis comparing CMS decisions with expert panel voting.
See the updated blog at the original location - here.
For example, CMS and Panel agreed 48% of the time. While Panel recommended 28 gapfills and CMS recommended 26 gapfills, only half the codes chosen matched between those sets of 28 and 26. All the codes getting gapfill were PLA codes. See more at the link.
For another example, CMS and Panel both recommended crosswalk 16 times. I randomly sampled 6 cases. If CMS offered its preferred codes irrespective of price, the CMS code could be randomly higher or lower in price than the Panel choice. In fact, 6/6 of the CMS choices were lower price than the Panel choices (p<.02). CMS pricing ran about 65% of Panel pricing.
 |
| click to enlarge |
