Despite countless hours of ICD-10 coder training and practice time on the part of their staffs, coding managers often lack access to objective measurements of real-world coder proficiency with ICD-10.
While training systems test coders with scenarios and check-the-box exams, organizations often struggle to predict how their coders will perform in a production environment, using actual discharges and live clinical documentation.
The importance of proactively assessing coder proficiency is even more critical considering ICD-10’s known areas of coding complexity and confusion.
Tackle the Trouble Spots
This article identifies five coding trouble spots revealed through automated assessments using real cases and a consistent answer key from 300 coders trained in ICD-10 during early 2015. The data revealed five areas for constant auditing and monitoring pre- and post-implementation.
Type of encounter—“A” for initial, “D” for subsequent, “S” for sequela
This information, denoted by the seventh character for certain conditions (particularly traumas), requires meticulous attention to document correctly. When auditing cases, look for the following:
- Do coders assign “A” for active treatment during initial encounters only?
- Do coders assign “D” for routine follow-up care only?
- Do coders assign “S” for a late effect of the initial condition only? Note that coders must report the original condition with a seventh character of “S” as well as a code for the late effect, resulting in a total of two codes.
Obstetrical coding in ICD-10 requires codification of trimester — something not required in ICD-9. When auditing, be on the lookout for the following:
- Do coders capture the trimester correctly?
- Do coders understand the changes in time frames for early versus late pregnancy?
- Do coders understand the difference between abortion and fetal death?
- Do diagnosis codes, in general, reflect the highest degree of specificity?
- Did the condition occur during childbirth or in puerperium?
- For multiple fetuses, with which is the condition associated (this information is necessary to assign the seventh character)?
ICD-10-PCS is a known area of complexity in ICD-10. Knowledge of new root operations and changing terminology confuses coders and surgeons alike. Furthermore, delays in obtaining operative reports may intensify during the early days of ICD-10 as clinical documentation improvement (CDI) teams work with surgeons to improve specificity.
A deeper level of coding review and skill assessment should be completed for surgical cases and complicated invasive procedures. When scoring coder skills, look for the following:
- Do coders assign the correct root operation?
- Do coders capture devices, as well as approaches, correctly?
- Did the coder use his or her advanced knowledge of anatomy to identify the correct body part when the specific documentation did not provide the option/term necessary for coding (e.g., splenic flexure = transverse colon; splenius plexus = abdominal sympathetic nerve)?
- Do coders correctly capture new ICD-10-PCS elements such as laterality and type of contrast?
Areas in which coding guidelines have changed
Three other areas where coding guidelines in ICD-10 are different from those in ICD-9 are included below. Some of these changes also result in DRG shifts and are therefore more important to monitor. Pay close attention to the following new ICD-10 guidelines:
- Sequencing anemia when associated with a malignancy, chemotherapy, or radiation therapy
- Time frame for current versus old myocardial infarction
- More efficient diabetes mellitus coding, including separate “code blocks” (e.g., diabetes due to an underlying condition or drug- and chemical-induced diabetes) as well as all-inclusive combination codes and the elimination of uncontrolled diabetes
The requirement to report external cause codes varies by state. However, it’s important to note that the specificity in these codes has increased dramatically in ICD-10. When auditing, monitor for the following:
- Do coders report the correct external cause code, or is another code more appropriate?
- Do coders interpret these codes correctly and consistently? For example, say a person is injured while pulling weeds in a garden. Is the activity landscaping or gardening?
Be Alert for NOS Codes
Through our testing program, we identified continued coder reliance on “not otherwise specified” or “NOS” codes. Further analysis revealed that insufficient documentation continues to be an issue. For example, emergency department (ED) coders assign NOS codes 22 percent of the time due to poor documentation. While some level of diagnostic uncertainty is expected in emergency medicine, this remained a significant finding and area of concern for the H.I.M. ON CALL audit team.
We are currently compiling a new data set for inpatient cases that include NOS codes. Coding managers should monitor the volume of NOS codes closely to determine whether the NOS code is truly appropriate or whether documentation improvement gaps are responsible for their overuse.
Be Methodic About Your Auditing Approach
In conducting automated versus manual coder assessments, we also identified five best-practice steps to take during ICD-10 coding reviews and audits. Consider these general auditing tips to closely monitor coder knowledge, proficiency and performance:
- Ask coders to code the same cases so you have an apples-to-apples comparison using real cases and live documentation.
- Establish and use consistent answer keys with your entire coding team.
- Focus primarily on the accuracy of the principal diagnosis, as this drives MS-DRG assignment.
- Once coders have mastered the principal diagnosis, begin to audit CCs and MCCs.
- Finally, move beyond CC and MCC capture to audit any additional/secondary diagnoses.
Nationwide Code Auditor Shortage Ahead
As more coders require 100 percent review (as expected in the early days of ICD-10) organizations must find new and innovative ways to manage higher volumes of assessments and audits. However, there is already a known shortage of ICD-10 code auditors. Not only are ICD-10 auditors hard to find, they’re also expensive.
Many consultant auditors charge at least $60 per chart, making it unrealistic for managers to contract for a review of 100 charts or more per coder. There simply isn’t enough money in most budgets to undertake this monumental and expensive task. However, technology may provide a solution for the journey ahead.
Training systems are quickly adding capabilities to automate coder assessments, thereby alleviating managerial burdens, streamlining reviews, and providing objective coder scoring data and dashboards. Just as educational institutions have embraced technology such as Scantron® to quickly analyze test data, healthcare organizations can use technology to do the same. Capabilities include the following:
- Software provides immediate feedback regarding coder strengths and weaknesses in ICD-10.
- System includes hundreds of real medical records that, when coded, allow managers to assess individual performance and easily spot trends across the department.
- Consistent and reliable answer keys are included, but are also customizable for each organization’s specific coding guidelines.
- Objective and cost-effective data analysis of each coder is provided, down to the specific code level.
With ICD-10 only a few months away and coder accuracy predicted to drop by 32 percent, now is the time to measure coder skills and proficiency using real cases, real documentation, and real data.
About the Author
Joseph J. Gurrieri, RHIA, CHP, the vice president of HIM Services at H.I.M. ON CALL, has more than 20 years of experience in health information management. Joseph previously worked for the New Jersey Hospital Association (NJHA), where he was the assistant vice president of information services. He has held various positions in hospitals in New York and New Jersey, including assistant, associate, and director of HIM.
Comment on this Article