December 2016

OIG Takes Notice of ICD-10 DRG Assignments
By Elizabeth S. Goar
For The Record
Vol. 28 No. 12 P. 24

What used to be two or more codes in ICD-9 is now a single combination code.

An unanticipated challenge arising from the massive transition from ICD-9 to the more expansive and highly granular ICD-10 code set has been adjusting to the new grouper logic for diagnosis-related group (DRG) assignments. Exacerbating the challenge, DRG issues have also caught the eye of the Office of Inspector General (OIG).

In its 2016–2017 Work Plan, the agency responsible for Medicare billing and reimbursement integrity has called out DRGs in two specific areas as subject to greater scrutiny due to previous findings of inappropriate payments. One is inpatient claims for diagnosis of kwashiorkor, a form of severe protein malnutrition that generally affects children living in tropical and subtropical parts of the world during periods of famine or insufficient food supply. Though typically not found in the United States, "Prior OIG reviews have identified inappropriate payments to hospitals for claims with a kwashiorkor diagnosis."

The second—and far more prevalent—inpatient claim under the OIG microscope involves mechanical ventilation. To qualify for Medicare coverage, a patient must receive 96 or more hours of mechanical ventilation. An OIG audit for 2009–2011 revealed an error rate of greater than 95% in mechanical ventilation billing to Medicare.

"I've identified out-of-date templates where start and stop times of mechanical ventilation are not captured," says Victoria M. Hernandez, RHIA, CDIP, CCS, CCS-P, an AHIMA-approved ICD-10-CM/PCS trainer and founder of Integrity Coding Solutions, an independent coding and auditing organization. "For lengths of stay greater than four days, one must not assume that the ventilation time was greater than 96 hours. This would be a good example of when to query for clarification along with implementing a better process and policy. There has also been inappropriate capture of dependence on respirator status when mechanical ventilation was provided for one admission for an acute condition."

Hernandez describes how the latter scenario typically occurs. "The patient is usually discharged home after being weaned, yet the status code is still being reported, which is inaccurate," she says. "The dependence on respirator status code should be reported if a patient is 'dependent' on the respirator long term and not for an acute condition. This status code falls under a hierarchical condition category [HCC] for risk-adjusted payers and should not be reported inappropriately."

Altered Logic
While the OIG focus is not directly related to the move to ICD-10, it nonetheless increases the pressure on coders to quickly bridge the knowledge gap between the two code sets to ensure appropriate reimbursement levels. Modifications made to grouper logic to account for the differences between the two code sets without changing the assigned DRG have caused much of the confusion. Specifically, clinical concepts that previously called for two or more codes now require the assignment of only a single ICD-10 combination code.

"DRG Grouping and ICD-10-CM/PCS" in AHIMA's HIM Body of Knowledge cites atherosclerotic heart disease with unstable angina as one example. Whereas this condition was previously reported with two ICD-10 codes, it now requires a single ICD-10 code. In ICD-9, cases with atherosclerosis as the principal diagnosis and unstable angina, a complication or comorbidity (CC), as a secondary diagnosis result in the case being assigned to a higher paying "with CC" DRG when applicable. Now, however, there is no longer a separate secondary diagnosis code to cause the case to group to a "with CC" option.

The Centers for Medicare & Medicaid Services (CMS) "made a major effort in the logic of ICD-10 groups prior to go-live, and one of the things they've done with that logic is that a code can be the principal diagnosis and still count as an MCC [major CC] or CC. That was never the case in ICD-9," says Becky DeGrosky, RHIT, product manager with TruCode.

She notes that ICD-10 codes haven't necessarily been a significant challenge for coders because, while the codes are new, the structure is familiar. Where the problem comes in is with PCS, which rolls into DRGs. DeGrosky points to Medicare severity, or MS-DRG, version 34, valid from October 1, 2016 through September 30, 2017, which classifies hospital inpatient encounters into groups based on similarities in resource utilization and clinical similarity. The groupings are classified using principal diagnoses, procedures, other demographic information, and the presence of CCs.

"When we implemented version 34 of the grouper, the biggest part of the logic change was in the PCS codes. Some were moved or renamed," DeGrosky says, adding that the changes addressed needs identified by ICD-10 coders posttransition. "They have a process in CMS that works really well, but we are going to see for the next few years that this is what is going to happen. It's not that things were wrong. Rather, they were moved around [to] simplify the DRG."

Hernandez points to the different coding guidelines between ICD-9 and ICD-10 as another source of confusion. The differences in chapter-specific guidance are readily apparent in diagnoses related to sepsis, anemia in neoplastic disease, hypertensive heart disease with congestive heart failure, and elderly gravida.

"The procedure coding also impacts the DRG assignment, which changed in ICD-10 where more specificity with root operation, approach, device, etc, are spelled out in detail. In ICD-10, coding arterial catheterization often shifts the DRG to an operating room procedure," she says.

However, Hernandez doesn't believe the differences are causing significant challenges for hospitals because ongoing coding revisions and updates are nothing out of the ordinary. "There's always that grace period in adapting to the change and staying abreast with the updates, which should be expected with HIM/coding professionals, but I would not call the DRG difference a significant challenge for health care organizations," she says. "The concept stays the same as long as the documentation supports it and the Official Guidelines for Coding and Reporting are applied."

Nonetheless, Hernandez's audits have revealed several patterns to DRG assignment errors. Missed, overcoded, and incorrect DRG assignments are most frequently related to sepsis, newborn, obstetrics, musculoskeletal, neoplasm, and respiratory cases. Keep in mind that because CCs, MCCs, and procedures performed all impact the DRG, error rates vary.

"Specificity is missed; there are errors in sequencing. Secondary diagnoses impacting severity of illness and risk of mortality are missed, and it's the same for procedures," Hernandez says.

There is yet one more potential reason some coders are struggling with DRG assignments in ICD-10: physicians being pushed by clinical documentation improvement (CDI) programs to document to support higher coding levels. "CDI [specialists] don't have a coding background and don't always understand the nuances. It puts coders in a position where the CDI folks have asked for things that they shouldn't. But that's the nature of the beast," DeGrosky says. "It takes a village, and close interaction between CDI and coders is essential to a successful process."

Bridging the Gap
To accelerate the elimination of errors in DRG assignments, DeGrosky recommends organizations continue ICD-10 training and offer refresher courses that were put in place pretransition. "Coding managers can pull back a bit, but they should still be auditing coders and keeping eyes on everything," she says. "And not just coders—physicians and documentation specialists, too. The same training that was put in place for the conversion should still be in place. Identify, plan, check, ask, then circle back. It's an unending process."

Noting that it's imperative for coding managers to know that personnel are properly following the rules, DeGrosky recommends accessing the guidance and expertise available through AHIMA's community of practice. By doing so, coding teams can stay on top of any significant changes being made to ICD-10's coding rules.

Internal coding guidelines also should be updated, maintained, and communicated, particularly as the number of coders working remotely increases.

CMS provides an online DRG definition manual that anyone can access if they need clarification on anything related to ICD-10's DRG groupings. "If something groups oddly, they can go out to that link and they can manually walk through the diagnosis code and procedure code, and manually calculate the DRG to see why it is grouped weirdly," DeGrosky says.

She encourages coders to play a role in improving the system. "They need to report [abnormalities]. CMS would rather hear from the real coders, the people who are out there in the trenches. They have more oomph," DeGrosky says. "Also, when they publish proposed rules and ask for comment, I encourage everyone to do so. They look at every single comment."

Hernandez recommends coding professionals enhance their skills by coding in all patient settings rather than limiting themselves to a specific specialty or patient class. She also notes that it's paramount for organizations to promote ongoing education, complemented by a robust CDI program.

"Concurrent reviews in addition to retrospective reviews are always beneficial," Hernandez says. "These reviews should also be payer agnostic so that patterns and trends across the realm—not only for the Medicare population, for example—are identified. In addition, HIM and coding professionals would most likely appreciate a nonpunitive approach to their reviews, especially for purposes of education for coders, auditors, and physicians regarding their documentation."

Getting DRG assignments correct is a small but significant part of the entire revenue cycle process. It's important that everyone at every level understand their role in the revenue cycle, Hernandez says. "Many are experts within their realm but may still be unaware of the big picture of the revenue cycle and how their role impacts all. In the end, all the quality data will promote and help achieve better health for all patients," she says.

— Elizabeth S. Goar is a Tampa, Florida-based freelance writer specializing in health care and HIT.