Much Ado about . . . Something?

While utilization of upper-level visit codes will continue to increase, risk will likely remain stable.,

There has been so much in the way of education, discussions, arguments, comments, compliments, and complaints about the 2021 evaluation and management (E&M) guidelines that it gets to be a bit overwhelming. And now, there are a lot of questions about efficacy, and whether practices are abiding by the new guidelines. As a computational statistician, I am particularly interested in how the shifts over the last two years impact audit risk and target identification – specifically, how they impact the predictive algorithms used by the government, private payors, and folks like me. 

To study this, I relied upon both the utilization patterns and audit results of near-time data obtained from a base of around 40,000 physicians. I looked at all office visit utilization for 2019, 2020, and the first half of 2021; this totaled around 68 million visits. Then I looked at the audit results for those same providers across the cohort of those office visits, which totaled around 90,000 audits. 

The first thing to note was the significant reduction in the reporting of E&M codes overall for 2020, compared to 2019. In fact, for this provider population, we saw a reduction of just shy of 36 percent. Without doing primary research, it is difficult to form a cause-and-effect opinion; however, I am pretty confident that the public health emergency (PHE) played a huge role. We all heard anecdotal stories about how the lockdown prevented patients from visiting their primary care physicians, and even after restrictions were eased, people were reluctant to visit their providers, except for the most urgent of health matters. Harvard School of Public health did a great study on this in May 2020; however, it lacks the benefit of a more recent analysis and a more robust lookback. Irrespective, they showed that beginning in March, there was a precipitous drop-off of visits to providers, and while telehealth helped to reverse some of that, there was a several-month delay until the impact began to kick in. This is probably because, while telehealth wasn’t new, it also wasn’t overly popular, and it took time for many practices to gear up for the HIPAA technology requirements. 

Interestingly, while the raw volume dropped off, there wasn’t much of a significant change in utilization distribution for most codes, as shown in the table below:

In fact, with the exception of the two lower-level codes for each category, which had the lowest utilization, distribution ratios were fairly stable.

For 2021, the annualized volume was pretty close to that of 2020. Unfortunately, the Centers for Medicare & Medicaid Services (CMS) is a bit behind on the release of the full Medicare database, and as such, I can’t compare our physician cohort against the full complement of physicians reporting to CMS. Part of this may, again, be due to the fact that even though some restrictions started to relax near the end of 2020, it took some time to change behaviors, which likely just started to occur during the first quarter of 2021. This time, however, I did see a more significant shift in utilization, particularly among codes 99204, 99205, and 99215. And again, for reasons I don’t understand, there was a precipitous increase in 99211 and 99212. I am certainly open to ideas about why use of those lower-level codes is on the rise. 

What did surprise me was that the ratio of new office visits to established office visits remained very stable across all three years (19.46, 21.58, and 21.85 percent, respectively). I really expected to see a drop-off in new office visits, as many practices were taking new patients, but alas, it was a poor assumption on my part. The table below shows the bigger picture for utilization variance year over year.

There has been general reporting, including by me, that there has been an increase in distribution of the higher-level E&M codes, such as 99214, 99204, and 99205.

One question that many providers have asked me is this: how does that impact risk assessment? Well, first of all, I think it’s important to note that there is simply not enough data yet (meaning contractor audit results) to build an effective predictive model. We are, however, getting close, and I expect that by the end of the year, I will have enough data to accomplish this. In the meantime, a proxy for that is to analyze internal audit results. And that’s what I did. As noted, I looked at the results of nearly 90,000 audits for office visits. These were performed by the staff for the same providers that reported utilization. Because 99201 made up a relatively small percentage of overall utilization, and because 99201 was eliminated in 2021, I excluded it from my analysis. 

Focusing on the upper-level codes (99214, 99215, 99204, and 99205), I concluded that there wasn’t a significant difference in pass/fail rates between 2019, 2020, and 2021. What does that mean? In general, it means that the risk hasn’t changed. For example, if we were to see an increase in fail rates over time, it would indicate an increase in risk, because more of the upper-level codes would have been coded in error. If the fail rates were significantly lower, it might indicate that risk was also lower, since an audit would result in a null finding. But because they are mostly the same, it would indicate that risk levels are likely the same, as well. What does this mean for the practice? It means that their compliance strategy should remain the same, when it comes to their efforts in risk-based reviews on their providers. 

One notable change was the ratio of under-coding to over-coding year over year for the mid-level codes (99213 and 99203). For 99213, over-coding was recorded 45 percent less than in 2020, with reporting of under-coding increasing by 58 percent. For 99203, while the under-coding reporting was about the same, reporting of under-coding was down by almost three-quarters. I suspect this means that in the next reporting period, we will see an even greater increase in the use of these upper-level codes.

My conclusion to all of this is that while utilization of upper-level visit codes will continue to increase, risk will likely remain stable, if future audits confirm what we have seen over the first half of 2021. All in all, maybe it really is much ado about nothing!

And that’s the world according to Frank.

Program Note: Listen to Frank Cohen report this story live during the nex edition of Monitor Mondays, 10 Eastern.