I believe this March 2018 blog was actually conservative, and I was basing it on full year data collection for ADLT prices. That would mean that low prices pushed into the future, would merely count against the lab the next year.
In fact, later CMS educational documents suggest that CMS will only collect half-year ADLT prices, meaning that low payments in Q1/Q2 that are appealed into final payments occurring in Q3/Q4 will never be counted in any year. See here.
Pricing rules for ADLT allow payment at list price in Year 1 (actually for 3 quarters) and then payment at the average price of Year 1 in Year 2. CMS gets a recoupment when the CMS list payment in Year 1 has proven to be more than 130% higher than the actual observed commercial median payment in Year 1.
I've made a model system where an ADLT has 900 sales to Medicare and 900 sales to commercial in Year 1, and 6,000 sales to medicare and 6000 sales to commercial in Year 2.
List price is $4000. Actual commercial payments in all years are 1/3 $1000, 1/3 $2000, 1/3 $3000.
In Year 1, the lab accepts payment for all its $3000 commercial cases, but appeals all the $1000 or $2000 payments into Year 2. Thus, its commercial median payment in Year 1 is $3000, which sets its Medicare rate for Year 2.
Since $4000 CMS paid list price in Year 1 is 133% of the later observed commercial median of $3000 in Year 1, CMS recoups 3% x (900*$4000) = 3% x ($3.6M) = $118,800 in Year 2.
Total CMS payments in Year 1 are 900*$4000 = $3.6M, 6000*$3000 = $18M in Year 2, and with the small clawback, this is net of $21,481,200 in the two years for 6900 cases, giving an average CMS payment of $3,113.
Had CMS paid all cases at the actual commercial median for both years, which was always about $2000, CMS would have paid the ADLT lab 6900*$2000 or $13.8M instead of $21.4M. Thus, despite the clawback rule, which applies only to Year 1, it is fairly easy to get net payments over 2 years that are 150% or more of the actual commercial medians.
|click to enlarge|
|click to enlarge|
Spreadsheet in cloud here.
For an OIG report on "gaming" of hospital quality metrics, here.
For an OIG report on data problems in PAMA data, here.