Analyzing Eight Months of ICD-10

Article by Mary Butler. This article was originally published on the Journal of AHIMA website on June, 2016 and is republished here with permission.

Teachers know there are two kinds of students. Student A studies and does their assigned reading throughout the whole semester, earning extra credit where they can. When the final exam rolls up, they’re anxious but prepared—they did their homework. Student B, on the other hand, has the same IQ as Student A, but lives for deadline extensions and the hope that their teacher will reschedule the final exam. Naturally, Student B does all their studying the night before the exam—even when the teacher has already moved the exam day at least once.

The scores of both students’ final exams might very well end up being the same, but the wise teacher knows which style of preparation pays off in the long term. The student who didn’t bank on delays and prepared when they should have will still tell you what they learned in that class one, two, or three years later, while the other student graduates a year after they should have.

The eventual implementation of ICD-10-CM/PCS on October 1, 2015 proved that hospitals, physician practices, health plans, billing clearinghouses, and coding professionals have more Student As in their ranks than laggards. Indeed, the go-live of ICD-10 came and went without the scores of rejected claims, cataclysmic financial losses, mass physician retirements, or chaotic coding that some in the healthcare industry forecasted. Some stakeholder groups, like the Cooperative Exchange, which represents the healthcare clearinghouse industry, declared October 1 a “non-event,” according to a statement they released,¹ while the Centers for Medicare and Medicaid Services (CMS) and AHIMA’s Sue Bowman, MJ, RHIA, CCS, FAHIMA, compared the go-live to the Y2K hubbub in 1999—well hyped but in reality no big deal.²

“There are some surveys that showed some [claim] denials, and some [providers] were having issues. I don’t think we can say nobody had any problems, that would be a stretch, but I think we can say the scale of them and the significance of the problems was much less than many people predicted,” says Bowman, AHIMA’s senior director of coding policy and compliance.

As the industry enters eight months of compliance with the new code set, it’s still too early to judge ICD-10’s impact on quality of care and healthcare costs overall. However, there are plenty of lessons learned to be shared as the industry prepares for the end of a five-year code freeze later this year, as well as the end of a 12-month grace period that temporarily allowed unspecified codes on certain Medicare claims.³ The way all industry stakeholders handled the ICD-10 transition—as well as the implementation delays—can shed some light on what the industry can expect as familiarity with the new code set grows.

Top Five ICD-10 Coding Weaknesses and Strengths to Date

Even after only eight months, coding auditors have been able to spot some coding weaknesses and strengths in ICD-10 coding. H.I.M. ON CALL’s Manny Peña shared the following ICD-10 coding areas in which his organization has seen coding professionals excel and struggle:

Five Strengths in Coding

1. Endocrine, nutritional, and metabolic diseases(E00-E89)
2. Diseases of the genitourinary system (N00-N99)
3. Diseases of the circulatory system (I00-I99)
4. Mental, behavioral, and neurodevelopmental disorders (F01-F99)
5. Diseases of the eye and adnexa (H00-H59)

Five Coding Weaknesses

1. External causes of morbidity (V00-Y99)
2. Symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified (R00-R99)
3. Injury, poisoning, and certain other consequences of external causes (S00-T88)
4. Congenital malformations, deformations, and chromosomal abnormalities (Q00-Q99)
5. Diseases of the blood and blood-forming organs, and certain disorders involving the immune mechanism (D50-D89)

Tammy Ree, BS, RHIT, CCS-P, CHC, CPC, with vendor CIOX Health, has also noticed recurrent mistakes. She says she’s seen:

• Simple misunderstandings in the acute hospital setting surrounding the application of the seventh characters from trauma and fracture codes. There’s a DRG difference between applying the seventh character for an initial visit versus a subsequent visit and vice versa. Ree isn’t seeing enough specificity around injuries because the documentation detail isn’t being seen in the record yet.
• There have been issues when the procedure code drives the DRG, especially if it’s not clear if the procedure was therapeutic or diagnostic.
• Auditors have seen confusion around using respiratory failure as a principal diagnosis and around sepsis coding.
• There tends to be mistakes around the use of guidance tools, such as fluoroscopy and ultrasound, and whether dye was used. Facilities should address their requirements in the coding policy so the application of codes for these services is consistent.
• Ree continues to see lack of documentation around devices, components, and grafting material used, such as bone versus synthetics. She’s seen a lot of back and forth discussion between clinical documentation improvement specialists, coders, and physicians to get the correct information, which is time-consuming and could affect overall productivity.

Coding Departments Reap What They’ve Sown with ICD-10

When the proverbial switch was flipped on October 1 last year, the providers that planned for the worst and hoped for the best came out on top, says Barbie Hays, CPC, CPMA, CEMC, CFPC, coding and compliance strategist at the American Academy of Family Physicians (AAFP). The AAFP is a member organization representing 70,000 primary care physicians. While some provider groups advocated forcefully for legislative and regulatory delays for ICD-10 prior to 2015, the AAFP stayed neutral during the debate and provided extensive educational resources to its members.

Hays says that for the vast majority of the academy’s members, the transition from ICD-9 to ICD-10 was “pretty seamless,” except for a slight productivity decline and some minor delays in payment.

“What’s been reported back to me from doctors and office staff is that there wasn’t a huge delay in payment,” Hays says, noting that it took coding professionals and physicians more time to code claims but not nearly as much time as some stakeholder groups warned.

Hospitals also fared well in the transition—at least that was the case with the coding department at Brookwood Baptist Health in Birmingham, AL. Chloe Phillips, MHA, RHIA, director of HIM and clinical revenue at Brookwood Baptist Health, says her revenue cycle was not impacted by the transition.

“When we went live with ICD-10 we saw a shift of coder productivity of about 30 percent the first couple of weeks. We have maintained at 10 percent or less loss in productivity since ICD-10 went live, so the productivity loss has been minimal,” Phillips says.

One of the reasons for the dip in productivity was that Brookwood had to move its coding system to a new server because of the volume of new codes, which made pulling up the encoder very slow.

Phillips attributes the minimal disruption during the transition to the extensive training Brookwood provided to physicians and coding professionals before go-live. Her facility did two years of dual coding on the inpatient and outpatient side. They achieved this by only dual coding on lower volume days of the week and rearranging workflows to allow for the practice. Phillips also required all of the system’s coding professionals to be on-site—even the remote employees—during the first weeks of the transition.

Phillips says that support from the executive leadership in her organization has been a “monumental” factor in its successful transition. “We’ve had buy-in from the top down from the get go. If you can get support and engagement from top down it makes things work. You can move mountains when you have leadership support,” Phillips says.

While Phillips made the most of the implementation delays by providing more training and practice, other providers convinced themselves that go-live would never come to pass.

Tammy Ree, BS, RHIT, CCS-P, CHC, CPC, manager, internal monitoring and coding compliance at CIOX Health, says that during the numerous delays to ICD-10 implementation she saw many facilities use their time well with additional education and testing—though not all.

“I worked with a facility that thought the ICD-10 implementation was never going to happen, so all preparation and education virtually stopped. On October 1, they got hit hard. Their AR [accounts receivable] went way up, coder productivity went down, and rework of claims edits increased.” Ree says. “I think the biggest lesson they learned was that you can never be too prepared, whether you believe it [is going to happen] or not.”

Physicians and hospitals weren’t the only well prepared parties. By the end of February, CMS Acting Administrator Andy Slavitt put to rest fears that Medicare and Medicaid weren’t ready to receive claims coded in ICD-10. In a blog post, Slavitt released claim submission and rejection rates of claims coded in ICD-10 in the fourth quarter of 2015 against historical benchmarks.⁴ According to CMS, about 1.9 percent of total claims submitted were rejected in the fourth quarter of 2015, compared to a historical rejected benchmark of two percent. Additionally, 4.6 million Medicare claims were submitted each day during the fourth quarter of 2015, matching historical baselines that were nearly identical.

Vendors also reported relatively few hiccups in the early days of ICD-10 implementation. According to Jared Sorenson, vice president of revenue cycle for 3M Health Information Systems, “the implementation of ICD-10 was a resounding success across the board. And we hear that in the industry, talking to a lot of our customers.”

Sorenson says over 5,000 hospitals use 3M’s code-finding software, which is usually referred to as an encoder. From what he has seen, industry fears of productivity and accuracy rates bottoming out simply did not materialize, Sorenson says. “Today those who are using our software are seeing on average 90 percent accuracy rates in ICD-10. There was a huge expectation of productivity loss, and estimates in the industry were forecasting a 30 to 40 percent drop in productivity,” Sorenson says. “That’s something we measure and look at very closely for our customers. In the first month, according to our surveys, they saw a productivity hit of 20 percent, which was roughly what the expectation was.”

Hoping to get a definitive look at how ICD-10 implementation has impacted coding productivity and accuracy, the AHIMA Foundation and AHIMA has launched a phased series of studies. The first study was conducted in May with 500 phone calls being placed with AHIMA members listing “coding professional” or a related title in their membership profile, and who held a corresponding credential. The data collected from this survey will be used to create an in-depth analysis and real world assessment of coder accuracy and productivity at various care provider settings. This analysis is expected to be released this summer.

Health plans also appeared to have weathered the transition well. Juanita Savage, RN, MBA, CPHQ, CPUM, director of medical affairs, reimbursement strategy and ancillary program management at Blue Cross Blue Shield of Michigan (BCBSM), says her organizations saw very minimal disruptions.

“We had a few issues internally with vendor systems and all of those issues were resolved,” Savage noted in a written statement. “There was no impact to providers, members, and claims. We had processes in place to resolve identified issues very quickly. From an industry perspective, reported vendor issues have been resolved and provider education has increased with increases in claim productivity reported.”

AHIMA’s Bowman says that prior to implementation there were a lot of rumors suggesting that payers weren’t ready. She says she was encouraged by an ICD-10 Monitor interview with George Vancore, director of IT integration for Blue Cross Blue Shield of Florida, in which Vancore reported that his system was seeing a decline in the use of unspecified codes and an increase in accuracy in ICD-10 over ICD-9.⁵

Bowman attributes this finding to the fact that “for a lot of the physician practices, the ICD-9 coding wasn’t that great and a lot of them never had ICD-9 training.” “The fact that they did get training on ICD-10 and the fact they’re probably being more careful with ICD-10 claims has proven to be a good thing,” Bowman says.

Billing Association’s ICD-10 Survey Reveals Mixed Bag

Results of a survey conducted by the Healthcare Billing and Management Association and released in March purports to be a valid snapshot of the revenue cycle industry’s response to ICD-10—and had less than rosy findings.

Thirty-eight billing companies responded to the survey, which revealed, among other findings, that three revenue cycle companies went out of business due to problems with implementing ICD-10. Additionally, it found that 22 companies reported that they experienced up to a 10 percent increase in denied claims when compared to pre-implementation figures. Eight companies reported no increase in denials. What’s more, 22 companies reported coding productivity is still up to 25 percent below levels recorded prior to ICD-10 implementation, while seven companies reported no decreased productivity.

AHIMA’s Bowman, however, found some silver linings in the report. “The fact that 22 companies reported coding productivity up to 25 percent below pre-ICD-10 levels… I think 25 percent is pretty good,” Bowman says. “It’s only 22 of the 38 companies, and seven reported no decrease in productivity, which for the billing companies that aren’t really coding experts, that’s pretty amazing.

“And 14 of the 38 reported that coding accuracy was the same as pre-implementation and 11 reported that it was more accurate. Only two reported a significant increase in coding errors. I thought that was very telling and very positive.”

Post-Implementation Challenges Ahead

The transition was widely seen as a success across multiple health IT stakeholders, but that doesn’t mean it’s all smooth sailing ahead. Many have attributed some of the transition’s success to an agreement the American Medical Association (AMA) made with CMS prior to go-live. The agreement said that for 12 months after ICD-10 implementation, Medicare review contractors would not “deny physician or other practitioner claims billed under the Part B physician fee schedule through either automated medical review or complex medical record review based solely on the specificity of the ICD-10 diagnosis code as long as the physician/practitioner used a valid code from the right family.”

In a written statement to the Journal, AMA President Steven Stack, MD, said this “mitigation period” is working well and helping providers avoid significant spikes in Medicare claim rejections. “This has mitigated some problems that were encountered, including issues with national and local coverage determinations, ensuring that these issues could be resolved and did not interfere with patient care,” Stack wrote.

Some consultants say this policy gave physicians a crutch by allowing them to select unspecified codes for a whole year. This could lead to a rude awakening on October 1, 2016. AAFP’s Hays says some physicians have relied on code assignments offered in their electronic health records (EHRs), which default to “unspecified.” Many physicians don’t realize they can drill down for further specificity.

CIOX Health’s Ree notes that “for many of the larger clients we monitor internally through audit, we’re seeing a lot of good detail.” “The collaboration with [clinical documentation improvement] CDI programs is really helping physicians understand the importance for that detail. But we see many facilities where physicians haven’t bought into it yet, they’re still somewhat indolent in their documentation practices,” Ree cautions. “CMS announced in 2015 that they would not deny ICD claims for the first year [based on specificity]. On October 1, 2016, that exception expires. I suspect providers will start to see a rise in the amount of denials starting around the third or fourth week of October. They’ll see denials that they haven’t seen before.”

Ree’s colleague, Julie Daube, BS, RHIT, CCS, CCS-P, director of external data quality at CIOX Health, notes that providers with robust CDI teams are performing the best in the transition. However, she says she has been getting calls from providers who are seeing their case mix index (CMI) drop and they aren’t sure why. This is having an impact on providers’ revenue cycles. Daube says the CMI, which reflects the acuity of the patients being treated, can fluctuate for a number of reasons such as the loss of a service line or because a new service was added.

Quality documentation of course plays a big role in reflecting how sick a patient is. Daube says that as an auditing vendor, they’re still in the early stages of figuring out what’s causing the CMI to dip.

“Is it documentation? Is it codes not being able to be assigned due to coder’s knowledge or because they don’t have documentation to assign, so they’re losing acuity? We don’t know,” Daube says. “On October 1, 2016, when we can pull that data together and compare it to the past, then we’ll be able to see what’s really going on. Maybe it’s the DRGs that have changed. Maybe it isn’t. It could be very interesting towards the end of the fourth quarter and into 2017.”

Those stakeholders in the business of providing training and contract coders hold a slightly dimmer view of the transition. Manny Peña, RHIA, CEO of H.I.M. ON CALL, launched a web-based coder testing and assessment software tool for ICD-10 that can also act as a quality assurance tool. This tool allows H.I.M. ON CALL’s contract coders and external clients the ability to do a real time accuracy and productivity assessment on ICD-10 coding.

According to reports based on its assessments, and a survey of H.I.M. ON CALL clients and non-clients, Peña says there are lingering concerns about coding productivity and accuracy, which so far are well below ICD-9 rates.

“For the month of March we had 196 coders on the system averaging, on the diagnosis side, 84.3 percent accuracy. On the procedure side, they’re averaging 80.1 percent accuracy, and then on the CPT, averaging 81.9 percent accuracy,” Peña says.

The results of H.I.M. ON CALL’s study of clients and non-clients found that in terms of productivity, there was a 39 percent reduction for inpatient coding, a 21 percent dip for ambulatory, and a 26 percent hit for emergency department coding.

Peña says he worries that some providers haven’t been able yet to accurately assess the accuracy of their coders. “For example, one hospital started receiving a lot of denials all of a sudden, and they feel like health insurance companies have not yet caught on, because they’re having other issues. It’s pretty much what we’re seeing,” Peña says.

Finally, one of the bonuses of getting ICD-10 implemented was supposed to be all the new, richly detailed data it would provide for physician practices and hospitals to use in population health and utilization management, as well as for the federal government, which hoped to use it for research and quality measurement.

AAFP’s Hays says that if providers are using all the technologies they currently have they should be getting a clearer picture of the acuity of their patients. However, it falls back onto the coding professionals and practice managers to do the coding correctly and present reports back to the physicians, Hays explains. These types of reports—for example, how many times a physician practice codes a given procedure—will become increasingly important as the healthcare industry moves from fee-for-service payments to value-based care.

Ree and Daube noted that ICD-10 is already giving providers better data—for example, obstetrics (OB). “The coding options there virtually tripled in ICD-10. We’ve been able to get better data for OB conditions and complications as well as detailed data around the types of services that practices are providing to patients. It’s not one simple delivery code any longer,” Ree says.

Better data analytics from ICD-10 will help providers comply with MACRA legislation—2015’s Medicare Access and CHIP Reauthorization Act (MACRA)—which includes payment provisions that reward payment and value. Specifically, MACRA introduced the Merit-based Incentive Payment System (MIPS) program and will establish incentives for physicians participating in alternative payment models (APMs) such as accountable care organizations, patient-centered medical homes, and bundled payment models.

3M’s Sorenson says it’s probably too early to suss out any major population health trends with ICD-10 data, but the analytics will be able to tell other things.

“As we shift to value-based payment, and more of that patient-centric view—a longitudinal view versus a look at a patient walking in the door—­I think you’ll find that coding becomes less of the basis of pure payment, and becomes more of the basis of tracking accurate information for the purposes of better patient management. I think that shift, ICD-10, enables it,” Sorenson says.

Notes

¹ Spivey, Mark. “Cooperative Exchange Declares ICD-10 Implementation ‘Non Event.’” ICD-10 Monitor. December 17, 2015. www.icd10monitor.com/enews/item/1555-cooperative-exchange-declares-icd-10-implementation-non-event.

² Slavitt, Andy. “Lessons Learned: Reflections on CMS and the Successful Implementation of ICD-10.” Centers for Medicare and Medicaid Services blog. February 24, 2016. https://blog.cms.gov/2016/02/24/lessons-learned-reflections-on-cms-and-the-successful-implementation-of-icd-10/.

³ Morse, Susan. “ICD-10 to get 5,500 new codes, including ones for face, hand transplants, CMS says.” Healthcare IT News. March 16, 2016. www.healthcareitnews.com/news/icd-10-get-5500-new-codes-including-ones-face-hand-transplants-cms-says.

⁴ Slavitt, Andy. “Lessons Learned: Reflections on CMS and the Successful Implementation of ICD-10.”

⁵ Spivey, Mark. “Health Plans Report ICD-10 Success.” ICD-10 Monitor. December 17, 2015. www.icd10monitor.com/enews/item/1553-health-plans-report-icd-10-success.

Mary Butler (mary.butler@ahima.org) is associate editor at the Journal of AHIMA.