While supplies last! Free 2022 Coding Essentials for Infusion & Injection Therapy Services book with every RACmonitor webcast order. No code required. Order now >

Frank Cohen

In prior articles I have discussed the Comprehensive Error Rate Testing (CERT) study and the impact it has on the auditing community. As a refresher, CERT is an annual study conducted by the Centers for Medicare & Medicaid Services (CMS) that was established to comply with the Improper Payments Elimination and Recovery Act of 2010 (IPERA), or Public Law 111-204. Its purpose is to identify cases in which a CMS carrier/FI paid a claim in error. For example, let’s say that you submit a claim to Medicare for “exploration of abdomen” (CPT code 49000) and subsequently are paid the prevailing amount (non-par would get 80 percent of the allowed amount).  Some time later, you get a letter from CMS requesting the chart for that claim, and after a CERT auditor reviews the documentation, the auditor determines that you should not have been paid due to one of the many reasons cited in the study (such as insufficient documentation, improper code, etc.). This would show up as an error (paid in error) in the published study and would contribute to the overall published error rate. So for the 2011 study, the most recent one, the published adjusted error rate is 8.6 percent, or $28.8 billion. This means that the carriers improperly paid out $28.8 billion to providers during the survey year. Note that the unadjusted rate for the study was 9.9 percent, meaning that 1.3 percent (or approximately 15 percent of the total) was reversed in favor of providers on appeal. This is important because in 2009, the federal government decided to use more stringent standards when auditing claims – and even under these circumstances, 15 percent of findings were deemed incorrect on appeal.

For physicians, the unadjusted error rate was 10.5 percent, which was reduced to 9.2 percent after adjustments for successful appeals (an overturn rate of just more than 12 percent). This accounted for $7.8 billion in supposed improper payments. This also included $800 million (0.2 percent) in underpayments, which unfortunately also are counted as errors in the report. 

Last year, recovery auditors like the RACs recovered nearly $1 billion in supposed improper payments. I use the term “supposed” because, since so few practices actually appeal the findings, no one really knows how sloppy the audit results are. In recent surveys I have conducted for practices that did appeal audit findings, I have seen an overturn rate as high as 80 percent, particularly for appeals pushed to the administrative law judge (ALJ) level. It’s important to note that auditors rely on this study to identify which areas and what types of providers are at the greatest risk for overpayment. Remember, findings of overpayment are not always due to something the provider may have done wrong, but often they are a result of the payer not doing its job correctly. I am not saying that it is right to withhold payment due to some arcane administrative rule, but rather that it is very wrong to pay a provider for a service it performs and then later to take back that money because the payer screwed up. 

Under the CERT study, there are specific categories assigned to the finding of improperly paid claims.  For hospitals, for example, these categories include incorrect settings, joint replacements, cardiovascular stents and cardiac pacemakers. For SNFs, the biggest problem was improper documentation, which ultimately results in determinations that services lacked medical necessity. For DMEs, it’s easier to list things that did not constitute overpayment errors, as for the DMEPOS market the error rate was a whopping 61 percent. I’d rather be a Walmart greeter than own a DME company (but that’s just me). For medical practices, typical issues included improper documentation, lack of medical necessity and coding errors. A real kicker was the findings for E/M coding. According to CERT, 12.2 percent of the overall FFS improper payment rate (13.9 percent of the unadjusted improper payment rate) was due to E/M coding, accounting for approximately $4.1 billion. The majority of the improper payments for E/M services were due to incorrect coding and documentation errors, including insufficient documentation. 

This brings me to my second issue: the new U.S. Department of Health and Human Services Office of Inspector General (HHS OIG) study on E/M coding titled Coding Trends of Medicare Evaluation and Management Services, which was released in May. The crux of this study was that the level of E/M coding within specific categories has increased significantly from 2000 to 2010. Again, this provides more ammo for the auditors, even if the results are not factual. The report begins with a check on the cost of the Medicare Part B program. Data indicates that payments for Part B services increased 43 percent from 2001 to 2010, going from $77 billion to $110 billion – and during the same time, E/M services increased by 48 percent. The report also indicated that “E/M services have been vulnerable to fraud and abuse,” going on to talk about a couple of entities that were fined for E/M fraud. It gives the immediate impression that increases in E/M levels equate to fraud, which isn’t the best way to start an unbiased study. If we were to look a bit further, we would find that the number of Medicare beneficiaries increased by around 18 percent in this time, and the relationship between a single Medicare enrollee and the number of E/M services rendered is quite nonlinear. We also know that the median age of Medicare enrollees continues to rise and that older Americans require (or demand) a higher level of service compared to younger Americans. 

What was amazing to me was the OIG’s claims to have eliminated 30 percent of physicians reporting to the Medicare database because they reported fewer than 100 E/M visits. What? Nearly a third of all physicians who submit claims to Medicare (around 650,000) report fewer than 100 encounters with patients for the whole year? The report also indicated that 1,700 physicians consistently billed higher-level E/M codes. What the heck does that mean? Of the 440,000 or so docs in the study, less than one half of 1 percent were found to be billing higher-level E/M visits. The report’s authors did qualify this by noting that their patients were the same as patients for docs who billed lower E/M codes by comparing ages and diagnoses of beneficiaries. But without reviewing the chart documentation, it is anyone’s guess as to the validity of that statement. Their summary indicated that CMS ended up paying, on average, $205 more per beneficiary and $43 more per E/M service to physicians who consistently billed higher-level codes. Again I ask, what is the point? The conclusion is that physicians are billing higher levels of E/M codes and that this, somehow, is wrong.  

Here’s my point (finally!) Auditors look at the CERT reports to determine targets for overpayments to physicians. If certain codes (like 99214 and 99233) are reported as being consistently overpaid, then the auditors target those codes. In the HPMP (the inpatient equivalent of the CERT study), if certain DRGs or surgical procedures are considered to be overpaid consistently, those areas become targets. In essence, the more the payer messes up, the greater the risk to the provider. With the OIG report, the issue revolves around a claim that physicians are now over-coding – and that simply lacks merit, turning what could be a valuable series of studies into more junk science. It’s always unfortunate when those who claim to be researchers or scientists force their own bias into their research. Assuming we accept that the study itself has some merit (which, by the way, I do not), what we have is nothing more than data. What the OIG did was assess a cause-and-effect assumption into the equation without actually conducting any type of cause-and-effect analysis. Hey, my cause-and-effect explanation is that the introduction of EHR programs finally has made it so that physicians are coding to the guidelines – so this increase in levels we see very well may be because, for the past 10 years, physicians have been under-coding! Is that true? Maybe, but if we go with the type of research seen in CERT and this OIG study, who cares?


In 2002, a published study dealt with the problem of E/M coding (Expert Agreement in CPT Evaluation and Management Coding by Mitchell S. King, MD; Martin S. Lipsky, MD; Lisa Sharp, Ph.D, Arch Intern Med. 2002;162:316-320). In this study, 300 certified professional coding specialists chosen at random were given six hypothetical progress notes of office visits (vignettes) and asked to select the appropriate E/M codes. The results were quite amazing, as the level of disagreement among this group of professional coders was between 50 and 71 percent. Think about that; certified professional coders, each with a minimum of eight years of experience coding in a physician’s office, and only somewhere between 30 percent and half agreed with each other on which E/M code was the proper pick. And this wasn’t the only study to report such results. 

So maybe coding variances have nothing to do with coders or physicians; maybe the coding guidelines themselves are deficient. Yet you don’t see that in either the CERT or the OIG reports. In the follow-up study I did to the above study, I found that physicians tended to under-code more than over-code – and in my own study of E/M changes during a five-year period, I found that variances were all over the place in terms of specialty.

So, who audits the auditors? And who peer-reviews the work of those who feed ammunition to the auditors? In 2005, the U.S. Government Accountability Office (GAO) looked at CERT, and while finding a whole lot of problems with the study methodology, the GAO concluded that the study itself was “adequate” for determining the overall level of overpayments nationally. But how does this translate once the data has been stratified by type of carrier? Well, it didn’t do quite as well, yet GAO still reported that the study was “adequate.”

The bottom line is this: practices, hospitals, SNFs and other medical providers, the overwhelming majority of which are doing their absolute best to provide quality services in an overregulated and hostile environment, should not be subject to overzealous, overaggressive and unfair tactics that challenge their financial survivability. Quality is expensive. When I mess up in my business, I have to eat my mistake. When the government messes up, everyone else pays the bill. Instead of increasing the level of stringency with which audits are conducted, how about we consider reducing the rule set and giving our medical providers a chance to do what they do best: provide quality care at a market price.

About the Author

Frank Cohen is the senior analyst for The Frank Cohen Group, LLC. He is a healthcare consultant who specializes in data mining, applied statistics, practice analytics, decision support and process improvement.

Contact the Author


To comment on this article please go to editor@racmonitor.com

Cardiac Procedures and Medicare Regulations


You May Also Like

Leave a Reply

Your Name(Required)
Your Email(Required)