I have heard from some plans that they relish the flexibility that ambiguity grants them to “sail closer to the wind” when it comes to policy decisions regarding HCC coding. It permits them to harvest more diagnoses and the revenue that they bring, which does all kinds of good things for maintaining rich plan benefits and lower premiums to compete in the market. Isn’t that a good thing? Well, maybe not.
When it comes to walking it all back from the charts in a contract level RADV audit, what will the failure rate be when the tally is done? CMS has yet to announce the long-promised extrapolation penalty and the FFS benchmark, so are these plans betting that (a) they will not be tagged among the 30 MAOs audited each year, or (b) even if they are among those being audited, their validation failure rate will not compare so badly to the presumed FFS benchmark?
The stakes around these policy decisions have recently escalated, however. HHS has published a schedule of estimated levels of improper payments for Part C of Medicare Advantage in 2014 amounting to $12.2B or 9% of total payments. On Part D, this is estimated at 3.3%, adding up to $1.9B. The aim is to recover this amount of money from the MA and Part D plans. https://paymentaccuracy.gov/tabular-data/projected-by-program/237
In contrast to previous statements by CMS regarding the plan for RADV extrapolation, it appears that these figures presume no FFS benchmarks to dampen the error rate by comparison. It seems that any chart validation error is an overpayment and, consequently, needs to be recovered. Here is a point of divergence between the HHS OIG and the CMS program, which has enormous financial implications for MA plans: zero tolerance for any chart audit discrepancy. That is what the industry had originally feared: the “nuclear option” for extrapolation, which would be financially catastrophic.
Next, consider the Department of Justice enforcement actions taken against a physician in Florida contracted with Humana, charged with bilking the risk adjustment system by deliberately upcoding diagnoses for his Humana members. This action puts into perspective the fact that the providers originating the patient encounters are also at risk for False Claims Act prosecution based upon their contributions to an erroneous flow of data factored into health plan risk scores.
At the healthplan level, we have to find ways to achieve the most accurate snapshots of our members’ health conditions without overstating them and thereby running risks of stepping out-of-bounds on compliance. We use “RADV” as code for this particular worry now. Yet given the additional emerging issues above, it is no longer limited to risk exposure under a CMS contract level RADV audit: the scope and scale of these risks has dramatically expanded. The OIG can conduct their own independent RADV audits.
We operate in an area with certain ambiguities require decisions that draw bright lines where none really exist. When push comes to shove as to what combination of diagnostic codes and specific chart documentation will actually pass muster under RADV audit conditions, we enter into a very gray zone. While training by CMS in the past has outlined the general framework of what is and is not acceptable, we often find ourselves at a loss for clarity when we get into the particulars. For example, at RISE, we hear an ongoing debate about where and when past medical history is acceptable and under what circumstances. CMS has steered clear of giving guidance on specific instances, instead referring to ICD-9 guidelines and The Coding Clinic for interpretative support. But even so, these resources have not provided the definitive answers to some of the nitty gritty “for instances”. This resulting ambiguity creates the proverbial vacuum which nature abhors, and on a plan-by-plan basis, our policy decisions attempt to fill. Consequently, when asked, coders from different firms around the country are all over the map on what they would and would not accept for retro-chart audits.
My argument is that this ambiguous situation creates a lot of confusion, at a minimum, but it also invites those that are more risk tolerant to drag the whole industry across a line that the OIG and DOJ might be happy to draw. I suggest that, if we can land on more concrete guidance about what will ultimately validate under a RADV audit, we would benefit as an industry and the specter of large scale payment recovery efforts would be lifted. Essentially, we should all be on the same page, including the OIG and DOJ. I suggest that the time for convergence is now.