The Magazine

Code Chaos

Another nightmare for doctors, courtesy of the federal government

Mar 10, 2014, Vol. 19, No. 25 • By STEPHEN F. HAYES
Widget tooltip
Audio version Single Page Print Larger Text Smaller Text Alerts

Jacksonville, Fla.
Ever considered suicide by jellyfish? Have you ended up in the hospital after being injured during the forced landing of your spacecraft? Or been hurt when you were sucked into the engine of an airplane or when your horse-drawn carriage collided with a trolley?

Dave Malan

Dave Malan

Chances are slim.

But should any of these unfortunate injuries befall you after October 1, 2014, your doctor, courtesy of the federal government, will have a code to record it. On that date, the United States is scheduled to implement a new system for recording injuries, medical diagnoses, and inpatient procedures called ICD-10​—​the 10th version of the International Classification of Diseases propagated by the World Health Organization in Geneva, Switzerland. So these exotic injuries, codeless for so many years, will henceforth be known, respectively, as T63622A (Toxic effect of contact with other jellyfish, intentional self-harm, initial encounter), V9542XA (Forced landing of spacecraft injuring occupant, initial encounter), V9733XA (Sucked into jet engine, initial encounter), and V80731A (Occupant of animal-drawn vehicle injured in collision with streetcar, initial encounter).

The coming changes are vast. The number of codes will explode​—​from 17,000 under the current system to 155,000 under the new one, according to the Centers for Medicare and Medicaid Services (CMS).

The transition to ICD-10 was planned long before Congress passed the Affordable Care Act in 2010. But Obama administration officials say it is a critical part of the coming reforms. “ICD-10 is the foundation for health care reform,” said Jeff Hinson, a CMS regional administrator, in a conference call about ICD-10 for providers in Colorado.

It will affect almost every part of the U.S. health care system​—​providers and payers, physicians and researchers, hospitals and clinics, the government and the private sector. That system​—​already stressed with doctor shortages, electronic medical records mandates, and the broader chaos of Obamacare​—​is nowhere near ready. And that has lots of people worried.

Health care professionals use ICD codes to talk to one another. The codes record diagnoses and services provided, and third-party payers​—​government, insurance companies​—​use the codes to determine reimbursements and to deter fraud. Coding errors can mean unpaid claims or costly audits​—​or both.

Virtually everyone agrees that the transition will mean decreased productivity and lost revenue, at least for a time. Some experts, dismissed as alarmists by ICD-10 enthusiasts, are predicting widespread chaos in a sector of the economy that can little afford it.

“I’m very nervous about whether once we flip that switch on October 1 this is all going to work,” says William Harvey, an assistant professor of medicine and the clinical director of the Division of Rheumatology at Tufts Medical Center in Boston.

But nobody really knows just what to expect. And remarkably, despite the embarrassing failures of HealthCare.gov, until recently the federal government had no plans to conduct end-to-end testing of the system before the launch this fall.

In a letter to CMS administrator Marilyn Tavenner on February 18, 2014, four Republican senators pressed for comprehensive testing. The senators​—​Tom Coburn, Rand Paul, John Barrasso, and John Boozman​—​are all physicians and expressed deep concern that CMS is planning only one week of “front-end” testing. After receiving the letter, CMS hastily announced that it will offer limited end-to-end testing to “a small group of providers” at some point in “summer 2014” and promised that “details about the end-to-end testing process will be disseminated at a later date.”

That’s hardly reassuring. One health care consultant, a longtime ICD-10 proponent, put it this way: “This is probably going to be the most painful year we’ve seen in the history of U.S. health care.”

On a foggy Thursday morning in early January, 30 medical coders gathered in a nondescript meeting room on the third floor of the downtown Hyatt Regency in Jacksonville. They paid between $585 and $985 each to attend a two-day “boot camp” on the new codes taught by Annie Boynton, from the American Academy of Professional Coders. On the black cloth covering each table were the day’s necessities: a Hyatt Regency pad of paper and pen, a coffee cup and saucer, a jar full of hard candy, a glass and a sweating metal pitcher filled with ice water. At each place, students found a thin spiral book​—​the “ICD-10-CD General Code Set Manual” for 2014​—​and a six-pound, phone-book-thick “ICD-10 Complete Draft Code Set.” 

Boynton began by asking the students to introduce themselves, to describe the practice that employs them, and, as an icebreaker, to tell everyone the first album that they’d purchased. She started in the back of the room, where, in an effort to remain unobtrusive, I had chosen to sit.

Left with no choice, I told the class that I was a journalist working on an ICD-10 story, and admitted, reluctantly, that my first album was Asia by Asia. (The signature song of that debut album, “Heat of the Moment,” played in my head for the rest of the day, as it may now do in yours.) Others in the class—with one exception, all of them females—came from a variety of fields that will be directly affected by the coming changes. There was an obstetrics coder, a Medicare contractor, a hospital administrator, and an owner of two urgent care clinics (Britney Spears, Def Leppard, the Monkees, and Michael Jackson, respectively).

Boynton, whose first album was Tiffany, is a native of northern Maine who now lives in Boston. The computer she uses for her PowerPoint presentation features a large “Eat Lobster” sticker, and her favorite descriptor, not surprisingly, is “wicked,” used as both a positive and a negative qualifier.

Boynton knows her stuff. She is the director of communication/adoption and training for UnitedHealth Group and she helped write the ICD-10 curriculum for the group sponsoring this course, the American Association of Professional Coders. A list of her credentials, displayed on the large screen at the front of the room, contains more letters than the alphabet: BS, RHIT, CPCO, CCS, CPC, CCS-P, CPC-H, CPC-P, CPC-I. She’s been working on the ICD-10 transition for nearly a decade.

She began the session with a straightforward question: “How many of your practices have begun to prepare for the transition to ICD-10?” Just three hands went up. Boynton smiled and shook her head in amazement. She’s not surprised. “I gave a speech to providers in California last month and only 7 of the 300 doctors in attendance had begun preparing for the transition,” she tells the class.

A survey of physician practices released in mid-January backs her up: Seventy-four percent of those surveyed reported that they’d done nothing at all to prepare. (Despite this lack of preparation, most expressed confidence that they’d be ready.)

“How many of you work for a physician who doesn’t think ICD-10 is even going to go live?” she asks. Almost everyone raises a hand. “If I had a nickel for every one, I’d be on a beach somewhere with a fruity drink in my hand. It’s 5 o’clock somewhere, right?”

Boynton launches into a brief history of ICD-10 and the debate surrounding its implementation. The current coding system, ICD-9, has been in place for nearly 30 years. Although it has expanded gradually, with additional codes to reflect new diseases, the latest innovations in treatment, and improvements in medical technology, it is nearing something close to its capacity. ICD-10 proponents​—​and Boynton is one of them​—​say there is no choice but to move to a more sophisticated code set.

Other developed countries began their implementation of ICD-10 some 20 years ago, after the World Health Organ-ization released its basic version of the new code set. But their versions of ICD-10 won’t be nearly as complicated as the U.S. version. Boynton says that only 10 other countries use the codes for reimbursements​—​one of the main functions of ICD-10 in the United States. And payment systems elsewhere are far less complicated, in part because there is usually just one payer: the government.

The multiplicity of payers in the U.S. system partly explains why ICD-10 will be vastly more complicated here. But, paradoxically, if government explains the simplicity of ICD-10 codes elsewhere, government largely explains the complexity of the ICD-10 codes here. And those codes are complex.

“If you sustain an injury falling off a toilet seat on a spaceship in Jacksonville after this class, there’s probably a code for that,” says Boynton.

There are codes for those “bitten” by a crocodile, “struck” by a crocodile, and “crushed” by a crocodile. There is also a code for injuries sustained through “other contact” with crocodiles. “I just don’t want to know about ‘other contact,’ especially with farm animals,” says Boynton, to sustained laughter. “That joke doesn’t fly in Montana.”

Boynton’s “personal favorite” is code V9027XA: “Drowning and submersion due to falling or jumping from burning water-skis, initial encounter.”

It’s the favorite of many who have studied ICD-10 codes (and the “white whale” for others). In the reporting that I had done before attending the ICD-10 boot camp, I’d had no fewer than five people mention it to me. The obvious question: Has anyone ever drowned because he’d jumped from burning water-skis? Do we need codes for things that have never actually happened?

For the answer, I turned to experts at the USA Water Ski Foundation and Hall of Fame. I was introduced to Lynn Novakofski, who was described to me as “a walking history book of water-skiing.”

His answer seems to confirm suspicions. “In my 60 years of skiing, I am not aware of a drowning caused by ‘burning skis,’ ” he told me. “Back in the ’50s, a popular act in water-ski shows was to pour gasoline on the water in front of a ski jump, light it on fire, and a ‘daredevil’ skier would jump over the flames. I have even seen the ski jumper kick off his skis in midair, dive headfirst into the flames, and swim under water​—​while everyone in the audience held their breath​—​to surface a safe distance from the burning oil. More recently, I have on occasion seen skiers, usually barefooting, skimming along with a water and oil soaked towel on fire billowing out behind them. This has a bit more potential for singeing the skin, but all the skier needs to do is drop into the water and the flames are quenched.”

Even if no one in the United States has drowned after jumping or falling off of burning water-skis, it’s possible such a tragedy has occurred overseas. I checked with Dr. Lorenzo Benassa, chairman of the medical committee at the International Waterski and Wakeboard Federation, who reported, after consulting “literature from the past 20 years” that he found “no cases” of “burning water-ski injuries.” He added: “In our experience, we have never heard of something similar.”

What about an injury short of drowning? Lynn Novakofski allows that there may have been some “minor injuries” as a result of stunts like the ones he’d seen years ago. But he didn’t recall hearing of any.

In any case, ICD-10 has those covered, too. There is a code for a mere “burn due to water-skis on fire” (V9107XA) and for someone being “hit or struck by falling object due to accident on water-skis” (V9137XA) or jumping from “crushed water-skis” (V9037XD). More generally, there’s “other injury due to accident to water-skis” (V9187X) and “other injury due to other accident on board water-skis” (V9387XA). And there’s the rather inexplicable code V9227XA: “Drowning and submersion due to being washed overboard from water-skis.”

“An injury from your water-skis catching on fire?” says Senator Tom Coburn, a physician who is leaving Congress later this year. “Eighty percent of these codes will never be used.”

How do these kinds of injuries​—​real or imaginary​—​get their own codes? This is one of the great mysteries of ICD-10. No one from any of the U.S. government agencies responsible for ICD-10 regulation and compliance would agree to an interview for this article, despite more than two-dozen requests over the course of two months.

That’s odd, since the Department of Health and Human Services (HHS) and the Centers for Medicare and Medicaid Services (CMS) are in the middle of a major public awareness campaign on ICD-10. If you follow CMS on Twitter, your feed is bombarded with tweets conveying the urgency of ICD-10 compliance. “Next CMSeHealth Summit on #ICD10 will be held on Feb 14. Register to attend via webcast here,” read a tweet from @CMSGov on February 3. The next day: “ICD10 is only 239 days away. Check out this CMS blog post on the last year before ICD-10” and “Need an overview of #ICD10? Check out the Intro Guide to ICD-10.” And the day after that: “Are you in a small or rural practice preparing for #ICD10? CMS has a fact sheet with tips for your practice” and “Not sure how your clearinghouse can help you with #ICD10? Read this to find out.”

Despite this urgency, public affairs officials from HHS, CMS, and the National Center for Health Statistics (NCHS) at the Centers for Disease Control all declined repeated requests for interviews. A spokesman for the NCHS provided this overview of the process on background: A contractor developed a prototype of the U.S. ICD-10 code set after reviewing recommendations from the World Health Organization; NCHS offered “enhancements” to that code set and revised it further after consulting with physicians, clinical coders, and other users of the previous version, ICD-9.

The objective was greater detail, more specificity. The new code set introduces the concept of “laterality” to ICD coding, allowing physicians to identify in code, for instance, whether a hand injury is a right-hand injury or a left-hand injury. But ICD-10 also adds thousands upon thousands of new injury codes​—​some 37,000 new musculoskeletal and injury codes all together, according to an ICD-10 expert who consults with CMS.

“There were 9 codes for bites in ICD-9,” says Boynton. “There are over 300 in ICD-10.”

Virtually every conceivable malady or injury has a code. There’s code V9102XA for someone who is “crushed between fishing boat and other watercraft or other object due to collision, initial encounter.” Or T71232A, “Asphyxiation due to being trapped in a (discarded) refrigerator, intentional self-harm, initial encounter.” If you are hurt in an abattoir, there’s code Y9286, “slaughterhouse as the place of occurrence of the external cause.” Code F521 is “sexual aversion disorder,” not to be confused with code G4482, “headache associated with sexual activity.”

Some codes appear to be anticipatory rather than descriptive. Has anyone in the history of mankind ever attacked another human with frog venom? Or sought contact with the same for the purposes of intentional self-harm? Probably not​—​and not just because frogs don’t produce venom.

But code T63813A is “toxic effect of contact with venomous frog, assault, initial encounter.” I asked Dr. Kyle Summers, one of the world’s leading experts on poisonous frogs, about this. He told me that frogs do not produce “venom,” and therefore, while some are poisonous, none are “venomous.” Summers further explained that while members of the Embera tribe of western Colombia have used batrachotoxin from the skin of frogs in the genus Phyllobates on the tips of blow-darts to kill monkeys, he did not know of any incidents in which the darts have been used on human enemies and had “not heard of anyone intentionally hurting themselves by contact with a poison frog. But,” he added, “I have not researched the issue.” Other codes describe occurrences that would seem unlikely to result in any kind of injury at all, such as code W20XXA, “contact with non-venomous frogs.”

Back in Jacksonville, Boynton moves from a general discussion of ICD-10 to some specifics. She explains in tremendous detail how the new codes offer several different ways of codifying engagements with patients—“initial encounter,” a “subsequent encounter,” and “sequela.” The “initial encounter” in codespeak is not limited to the “initial encounter” as one might understand it in plain English, Boynton explains. There could, in fact, be several initial encounters with a patient, if those subsequent visits involved the initial injury and treatment. Bewildered looks spread across the class like bad herpes (A6000 or one of the other 38 herpes codes), and the resultant confusion led to a series of questions about the meanings of “initial” and “subsequent.” One student asked the question that seemed to be on the mind of everyone in the room: “So a subsequent visit would still be an initial encounter?” And then, after a brief explanation, another question: “Wait, there could be five initial encounters with the same physician?” 

After lunch, the class plunged deeper still into the intricacies of the new coding. Boynton walked the class through “excludes” codes, meant to prevent using two codes that would seem to contradict one another, and the advent of the “placeholder” character, intended to allow coders to fill all seven characters of a code in which not every character has meaning. (“X can be a placeholder, but it can also be a code character.”) Boynton is a very clear communicator and managed to keep the interest of most of my classmates by alternating between code minutiae, issuing stark warnings about the consequences of failing to understand ICD-10, and dropping the occasional codeworld inside joke.

But the system is complex and the scope of change is immense. “Learning these codes makes learning Mandarin seem easy,” she tells a frustrated student. This is what has so many in the health care world nervous.

The introduction of a system with exponentially more codes, and far more complicated codes, will inevitably mean many more coding errors. The default position of payers, whether government or the private sector, will be to deny all claims that are not coded correctly. In many cases, providers will be left with a lose-lose choice: forgo payment altogether or dedicate valuable time and resources to appealing the denied claims. Hospitals, large physician practices, and other big institutions can absorb some of the losses and have the workforce at their disposal to challenge the denials. Small practices do not.

“When you have a provider who hasn’t prepared, who doesn’t know the codes, ​and they have every claim rejected because of improper coding for three months, that’s going to put people out of business,” Boynton tells me over breakfast before the second day of training.

“Most practices in the United States are small businesses,” says Senator Coburn, an obstetrician and family practice doctor from Muskogee, Oklahoma. “This could ruin them.”

An ICD-10 preparation plan from the Health Information and Management Systems Society (HIMSS) advises practices to have a minimum of six months revenue in reserve to help avoid that possibility. Such warnings have been coming for years. Financial institutions have begun offering lines of credit targeted to potential ICD-10 shortfalls. “With potential disruptions becoming more and more probable as the industry hurtles haphazardly towards October 1, 2014, having half a year’s cash or credit on hand may be vital to keeping your doors open,” writes Jennifer Bresnick in EHR Intelligence, a website that tracks news on electronic health records and medical technology.

A 2008 study on the costs of implementing ICD-10 from the health care IT firm Nachimson Advisors warned that “significant changes in reimbursement patterns will disrupt provider cash flow for a considerable period of time.” The study projected that the total cost of the ICD-10 implementation would be $83,290 for a small practice (3 physicians and 2 administrative staffers), $285,195 for a medium practice (10 providers, 1 professional coder, and 6 administrative staffers), and $2.7 million for a large practice (100 providers, 10 full-time coding staffers, and 54 medical records staffers). Boynton says those numbers seem on target five years later.

Coburn believes the new system will require doctors to spend more time coding. “You’re just not going to trust a nurse to do that,” he says. “If they put in the wrong code, they’re going to hammer you. The penalties are getting more severe. If you fail a recovery audit, they don’t just take your money, they penalize you on top of that.”

Coburn’s concerns go beyond the likelihood of a rough transition to ICD-10 to the long-term effect the changes could have on the doctor-patient relationship. The specificity of the codes will require doctors to spend more of their time on documentation. “Let’s say it takes you an extra two minutes per patient to do the coding yourself,” he says. “It doesn’t sound like much. But if you see 30 or 40 patients a day, that’s at least an extra hour you’re spending on this stuff. That minute or two that you’re not spending talking with the patient might be the minute when you learn something critical to your diagnosis or treatment plan.”

His prescription: “Delay it forever. The health care system can’t take another cost, especially right now.”

Coburn has introduced legislation to do just that, but most industry experts believe the prospects for a delay are poor. ICD-10 implementation has already been delayed twice, most recently in April 2012, giving “covered entities” an extra year that expires at the end of September. “I’d be shocked” if there’s a further delay, says Holly Louie, the ICD-10 coordinator for the Healthcare Billing and Management Association.

CMS administrator Jeff Hinson, in his conference call with Colorado providers, offered a stern warning about the October 1, 2014, compliance date. “You need to know that the deadline is firm,” he said. “The deadline is firm.”

That could spell disaster.

Despite desperate pleas from virtually every corner of the health care industry, the federal government has offered no details for comprehensive end-to-end testing of the new coding system before it goes live in seven months. “In meetings over the past two or three years between commercial payers and CMS, we were told that if everything went relatively smoothly, and we just saw the typical hiccups associated with a major transition like this, there could be as much as a one-year disruption in cash flow​—​for both large and small practices,” says Louie. “And that’s when we thought there would be end-to-end testing.”

Coburn, along with the other Republican physicians in the Senate, is trying to force CMS to perform comprehensive testing or to delay the start date. “Given the size and scope of the potential transition to ICD-10, the brevity and limited scope of this test is worrisome,” they wrote in their letter to Tavenner.

Annie Boynton, the ICD-10 trainer, says that 20 of the 50 states have done “nothing” to update their systems for ICD-10. “The sector of the industry that scares me the most is government,” says Boynton. “Historically, they are not great with major regulatory implementation rollouts,” she adds with a knowing smile. “HealthCare.gov was a perfect example. When [Health and Human Services] Secretary Sebelius was sitting in front of Congress, the Energy and Commerce Committee, answering all those questions, I had this really sinking feeling that in 2015, we’re going to be there again.”

Stephen F. Hayes is a senior writer at The Weekly Standard.

Recent Blog Posts

The Weekly Standard Archives

Browse 18 Years of the Weekly Standard

Old covers