Category Archives: artificial intelligence

AI and Audits

Medicare audits growing rapidly. Artificial intelligence is being used to replace medical judgment. Cost of audits dropping rapidly. Health care providers forced to allocate ever-increasing amount of resources to audits.

BIG DATA AND THE FUTURE OF MEDICARE AUDITS

Edward M. Roche, Ph.D., J.D. — Barraclough NY LLC

Part III — Artificial Intelligence and the Audit-Free Future

In Part I of this series, we discussed the exploding number of Medicare claims and the inability of the current appeals regime to handle the workload.  We also reviewed how special computer algorithms are being used to down-code Diagnosis-Related Group (DRG) claims, and argued that these actions are not really “audits” because artificial intelligence (AI) algorithms using statistical comparison are being substituted for medical judgment.

In Part II we examined the emerging arguments being made in “algorithm law”, but suggested that this area of litigation will need to be developed further, and the type of experts needed in the appeals hearings will change dramatically, because they will need to be familiar with artificial intelligence (AI).

In this closing part of the series, we will examine scenarios for the future. But in looking at the future, we must make a few reasonable assumptions.

The number of audits will continue to increase, and one reason for this is that due to automation the cost of audits is dropping rapidly.
The ability of the appeals system (re-determination; reconsideration; Administrative Law Judge; Medicare Appeals Council) will remain under pressure to handle the litigation workload.

The quality of audits, which most agree is very poor, will not improve, primarily because there is no incentive for the RACs do do so.

Health care providers will be forced to allocate an ever-increasing amount of their already scarce resources to dealing with audits.

Given these assumptions, there are a number of scenarios that seem reasonable ten to fifteen years hence.

Future Scenario One

More of the same. The system will continue as it is, but will simply become worse for the health care provider. The burden of audits (uncertainty, claw-backs and litigation expenses) will continue to grow. Health care will become a sector that few will wish to enter into as a career. More providers will become bankrupt.

Future Scenario Two

Change in appeals procedures. CMS already recognizes the backlog problem in appeals and has started to take action. In these proposals, there is little discussion aimed at re-thinking the overall auditing process. The primary change is in improving the capacity of CMS to handle the litigation.

There are many variations of Scenarios One and Two. But lets take a look at the future using an “out of the box” approach.

Future “Out of the Box” Scenario Three

In this scenario, algorithms using artificial intelligence continue to be used, but the provider’s medical information system will be designed to intervene before the claims billing stage. Here is the logic: If it is possible to find a different coding solution looking backwards, as current auditing approaches now do, then it should be possible to apply the same algorithms to prevent bad claims from being filed in the first place.

The optimum solution would be to replace the auditing system, and instead insert the artificial intelligence algorithms between the health care provider and the government.  Instead of being brought in after the fact, these algorithms will be injected into the space between the provider and the claims system beforehand. (See Figure) The AI system would simply stand as a front end for claims processing. It would correct deficient claims and prompt for additional information as needed.

racmonitor-cyber-audits-exhibit-001

The standing algorithm could be standardized across the United States, and as we know, today’s technology allows constant updating to the algorithm software, much like computer security updates today are pushed out from vendors.

And what would happen to the RACs? We don’t want these poor people to lose their jobs. They would transition into working for the health care providers and operating the algorithm engines. In so doing, they would focus on making sure that the AI reflects sound medical judgment, and not merely the desire to extract as much money as possible out of the hide of the provider, which is the case now.

This would eliminate the need for auditing altogether, and end this scourge of litigation and chaos that sits on the shoulders of the provider. Perhaps this type of solution might be considered by public policy makers and perhaps CMS needs to think about a more intensive R&D program. Carpe Diem — the future is there for the taking.

Note: Prior to entering law, Dr. Roche served as the Chief Research Officer of the Research Board (Gartner Group), and Chief Scientist of the Concours Group, both leading IT consulting and research organizations.

This article was originally published in RACmonitor.

 

 

 

 

Tyranny of Algorithms

BIG DATA AND THE FUTURE OF MEDICARE AUDITS

Edward M. Roche, Ph.D.,J.D. — Barraclough NY LLC

Part II — Defending Against the Tyranny of Algorithms

In Part I of this series, we reviewed how the number of Medicare audits has increased by almost 1,000% in the past five (5) years, and how virtually no decisions by ALJs are being handed back within the statutory time frame.

We discussed also how RACs have started to rely on big data mining of hospital claims to generate large numbers of Diagnosis-Related Group (DRG) downgrades. This is costing hospitals plenty, not only in the reduction in payment revenue, but also in the constantly increasing cost of defending against audits.

The use of computer algorithms has drastically reduced the cost of conducting audits, but there has been no corresponding reduction in defensive costs for hospitals, and this is an example of what military people call “asymmetric warfare”, where the cost of defense is always disproportionately greater than the cost of offense. It is an impossible game to win.

We will now examine a few of the legal issues that are presented by the need to defend against not an audit, but against an algorithm.

The MPIM specifies that the decision to conduct an audit is “not reviewable” in a hearing. This means that even if a provider is being profiled or targeted through an artificial intelligence algorithm, they are fair game, no matter what the reason.

This lack of review-ability does not extend to the review itself. That is handled by the appeal system. The typical appeal has little success in the first two levels — reconsideration, redetermination — so the grass gets mowed with the Administrative Law Judge (ALJ). Appeals generally are based on a claim-by-claim argument regarding each patient or procedure, combined with a refutation of the statistical extrapolation, which is almost always based on shoddy work.

This litigation profile will change. Why? Rather than challenging the expertise or judgment of the audit reviewer who rejected a claim, the argument instead will be aimed at dis-crediting the algorithm responsible for the claim rejection.

But since these algorithms do not make decisions based on medical logic, but only on a pattern of statistical probabilities, the arguments against them will by necessity be couched in quasi-mathematical terms. To do so will require resort to an entirely different type of expert, and understanding of what we might call “algorithm law”. Yet, for the most part, many of today’s health law attorneys are ill-prepared to litigate this type of case.

This article appeared also in RACmonitor.