Medical coding poses as rational exercise. Every service has a code. You document the service you provide. You get paid. On the surface, it seems pretty straightforward.

But add the alchemic CPT rules and the intricacy of treating an actual patient to the mix, and you get a system that tends toward confusion, if not chaos.

The proof? Physicians Practice conducted a simple experiment. We asked four certified coders to identify the best diagnosis and CPT codes from six actual patient charts. In many cases, they gave quite different answers. (See their chart assessments, pages 29-31.)

Read on to find out how dramatic coding variations affect practices' incomes  - and discover some doable solutions for better documentation.

As physicians have long known, coding is as much mercurial art as rational science. It's more like filing a tax return than measuring blood pressure. Or, as professional coder Annette Grady, director of reimbursement at The Bone and Joint Center, in Bismarck, N.D., puts it, the process of coding "is more creative than black and white."

And the impact on reimbursement is significant. Physicians Practice asked four certified coders to identify the best diagnosis and CPT codes from six actual patient charts. In many cases, they gave quite different answers. In the most dramatic terms, if a coder selected all the highest codes, reimbursement for the six visits totals $2,062. If a coder selected all the lowest codes, total reimbursement drops to $1,438. That's a $624 difference just on six visits. Imagine that multiplied over every visit for a year.

Of course, no physician wants to be accused of fraudulent billing, but a changing coding system makes it seem impossible not to code incorrectly from time to time. 

"If the [professional coders] can't [pick the right code], how in the world can [CMS] possibly expect physicians to do any better?" asks Timothy Bopp, an orthopedic surgeon at The Bone and Joint Center who nonetheless codes his own E&M visits, relying on staff only for procedural coding.

"You need to bill for what you do and only for what you do," says Bopp. "And if you've got it documented, you shouldn't bill for less than what you did for fear you might be audited." Logical enough.
At the same time, as our experiment illustrates, simply billing for what you do isn't as easy as it seems - bad news for physicians like Bopp who keep hearing about colleagues being prosecuted for fraud. It's a classic Catch-22.

"Recently, CMS has come out with these fraud guidelines, and we've heard from our public officials about how much fraud is going on in medicine," Bopp says. "There are all these bulletins going around ... about how more and more doctors are getting sued for fraudulent practices. ... It's being publicized a lot more. They've hired a lot more people to actually go out and look for fraud, and if you notice, there are even commercials on TV trying to convince elderly folks to report doctors that they think might be overcharging."

The atmosphere is such that physicians are afraid, even doing their best, that they might still make a mistake - and be punished.
Indeed, undercoding by underconfident physicians is rampant, says Barbara Cobuzzi, who, as president of Cash Flow Solutions, Inc., in Lakewood, N.J., regularly audits practices. "Physicians are so scared of being accused of upcoding that they tend to run 99213s all the time. I call them 'three-spikers.' They run [level] 3s all the time because they want to be safe. Not only are they ripping themselves off now, they are ripping themselves off for the future. They will be measured against what they were billing last year and the year before as an expected pattern. So if they only put 3s, they will always be expected to only put 3s."

In other words, lack of coding confidence is hurting physicians in multiple ways.

Do the documentation

But how can physicians' coding confidence improve when even our coders disagreed so consistently? (See their chart assessments on the following three pages.) Is the coding system really that chaotic? In many cases, we discovered, the problem lay not so much with the CPT codes themselves as with the way different physicians document what they did to arrive at a CPT code. "Physicians have their own language, and it doesn't always match CPT language," observes Grady.

When our panel of coders disagreed with each other, it mostly was because they interpreted the physician notes differently or didn't have enough information to make a definitive choice. Forced to guess what really happened, some coders made aggressive assumptions about what probably occurred. Others played it very conservatively, assuming nothing and coding lower.