November 15, 2011

ICD-10 Conversion: New Tools, Strategies Are Keys to Success

By

As if hospital administrators didn’t have enough to worry about these days, they’re now staring down the significant risk of productivity and revenue losses that could come as a result of ICD-10 implementation – unless they start preparing now.

With denied claims expected to more than double, going from 3 percent to 6-10 percent, a 250-bed hospital could take a financial hit of as much as $850,000 in 2014 due to lost productivity, denied claims and undercoding (see Figure 1) resulting from the use of more complex ICD-10 codes. Add in the intensifying coder shortage and many healthcare organizations are at a loss to identify where to begin solving this impending crisis.

 

Many providers, with the October 2013 deadline looming, are undergoing initial assessments to determine the costs and approach they will need to make the ICD-10 conversion as smooth and efficient as possible in a two-year time frame. There is growing awareness of the need to implement new computer-assisted coding (CAC) technology that goes well beyond the capabilities of existing coding tools. However, many are struggling to find solutions that will address all of the most critical ICD-10 pain points.

To ensure a successful ICD-10 conversion, healthcare organizations should consult the following checklist and assess new tools and strategies:

Do your CAC homework: a CAC can read and interpret an entire set of records and recommend a short list of codes within seconds, resulting in increased coder productivity, improved coding accuracy and enhanced coding compliance. However, some of these software systems require extensive coder involvement in selecting codes while others shift much of that burden to the software itself, transforming the roles of coders to auditors or reviewers. The right CAC tool will help providers decrease administrative costs, time to revenue and denials due to inaccurate coding.

Check the NLP in your CAC: recent advances in natural language processing (NLP) technology are powering CAC solutions that scan and interpret contents of clinical documents directly and extract discrete, meaningful pieces of information. But not all NLP is created equal. The standard measurements of NLP accuracy are precision and recall. Recall is a measure of coverage; enhanced recall corresponds to fewer missed codes in documentation. Higher precision corresponds to a higher number of accurate answers (codes). Understand the precision and recall exhibited by the various NLP technology choices, as well as NLP’s expected impact on ICD-10 preparation. Less-sophisticated NLP technology may struggle with the sheer scope of ICD-10 and exhibit low precision, producing a large number of matches but with many errors. Of the many approaches to NLP, providers should look beyond simple medical dictionary and pattern matching, instead aiming for a process that utilizes a combination of statistical modeling and symbolic rules. This will ensure that the CAC will decipher the meaning and context of facts within medical records and subsequently recommend the appropriate codes, truly saving time and delivering financial impact.

Understand your need for speed: ask how much time it takes for any solution to interpret patient records with highly granular details such as laterality (right or left), severity, acuity, exact body part affected, etc. Solutions should be able to scan and interpret an entire medical record within seconds to alleviate the productivity challenges of ICD-10. Even some of the first implementations of CAC solutions demonstrated a 20 percent increase in the number of inpatient charts coded per hour. Others have seen an up to a 90 percent increase above average ED coder productivity standards.

Review for accuracy: superior recall means that a CAC tool uses an extensive knowledge base to classify a variety of medical concepts in patient documentation correctly, which is critical to avoid undercoding and missed charges. This can help with the ICD-10 concern that the increased volume of code choices may make it difficult to find the correct code. There may even be varying degrees of “correct,” with the most specific generally being the most accurate. With a focus on accuracy and coding quality, providers can reduce external audit recommendations and associated costs significantly while reducing revenue leakage.

 


 

Rework the workflow: HIM leaders will need to evaluate how use of a CAC will impact their coders’ workflow and ultimately impact the number of coders they require. The experienced judgments of coders still will be essential to the process, as the complexity of both inpatient and outpatient encounters and the common use of hybrid records require human expertise to ensure that finalized codes are complete and accurate. As a result, and as recommended by an AHIMA-sponsored CAC committee, coders and coding managers will transition into “coding editors” and “coding analysts,” respectively. At the same time, an automated coding environment involving CAC will offer significant process improvements – supported by the right CAC tool, HIM leaders can cut coding workflow steps by half (see Table 1).

As hospitals brace for the changes they need to make to adopt ICD-10, it will be very important that they create an implementation plan that helps them choose the software that best matches their needs and exhibits a strong track record. Equally important will be that they plan appropriately for training of their coders, change management and follow-up. The result will be fewer bumps in the road when the new coding takes effect, and ultimately an increase in the overall quality and efficiency of their revenue cycle operations.

 


 

 

 

Read 38 times Updated on March 16, 2016
Mark Morsch, MS

Mark L. Morsch is the co-inventor of the LifeCode Natural Language Processing (NLP) engine with three patents on NLP technology for computer-assisted coding (CAC). He serves as vice-chair of the HIMSS Health Story Project and was co-chair of the AHIMA-sponsored CAC Summit in 2011 and 2012. Mr. Morsch is an author and speaker on the effective use of NLP in healthcare and has over 20 years of experience developing NLP solutions. Morsch is also an AHIMA-approved ICD-10 trainer was named an Optum Fellow in 2012. He holds Master of Science and Bachelor of Science degrees in electrical engineering from Clarkson University.