A week ago, we used PAMA raw data to show not just the median price for codes, but the actual distribution of payments (normal vs bimoda vs skewed left or right, etc). Blog here. In a later blog, we wondered if accusations (from employers) that United Healthcare was overpaying some providres, might be related to the most extreme prices seen on the far-right column of our bar charts. Blog here.
I've now had Chat GPT re-run the raw data. In each triplet of charts, the top chart (all in blue bars) is the number of paid claims in each "bin" of price ($100-200, $201-300, etc.) Price distributions are very broad; for 81408, from $5 to $2100 per unit test (median $477). The ratio of highest-paid claim to lowest-paid was 40X.
Then, for the second and third bar charts, the Y-axis is no longer test volumes, but actual revenue dollars. (For example, 10 units of a $100 test is 10; but the revenue of the test is $1000).
Finally, for the third bar chart, we arbitrarily assumed that 35% of the median price (the CLFS price) is a ballpark number for COGS. We subtracted that, so in the third bar chart the Y axis is still dollars but now as contribution margin (revenue - COGS).
81513 (Example)
For example. for code 81513, "Aptima BV," price bins run from $0-$50 up to $200-plus. Bin $200-plus is only 6.9% of unit volume, but brings 38% of revenue (payments), and brings almost half (48%) of contributio margin. That means: 7% of claims paid at top-end prices are half of profit.
 |
| click to enlarge or download PDF set |
See a full set of examples in PDF format HERE.
