October 1, 2014, is a day circled in red on the healthcare industry’s calendar. This is the day that ICD-9 code sets will be replaced by ICD-10 code sets.
ICD-10 has the potential to affect the entire cycle of care delivery. “ICD-10 is similar to Y2K in that systems will need to be remediated and tested. Yet it involves so much more than changing technology,” according to Craig Kappel of Ernst & Young LLP. “The impact on finances, operations, clinical processes, and people is unprecedented. This is not a compliance exercise by any stretch of the imagination. This is an organization-wide business process transformation.”
While ICD-10 is coming quickly, little progress has been made toward transitioning, according to the Workgroup for Electronic Data Interchange (WEDI). Out of nearly 1,000 respondents to a recent WEDI survey, which included EHR vendors, health plans, and physician providers, half of respondents said they were less than halfway through product development, and only 50% of health plans have completed their impact assessments as of March.
By Cindy Dubin, Contributor
One healthcare organization describes how it is getting ready for the transition from ICD-9 to ICD-10 and how speech recognition will improve documentation productivity with EMRs.
October 1, 2014, is a day circled in red on the healthcare industry’s calendar. This is the day that ICD-9 code sets will be replaced by ICD-10 code sets.
ICD-10 has the potential to affect the entire cycle of care delivery. “ICD-10 is similar to Y2K in that systems will need to be remediated and tested. Yet it involves so much more than changing technology,” according to Craig Kappel of Ernst & Young LLP. “The impact on finances, operations, clinical processes, and people is unprecedented. This is not a compliance exercise by any stretch of the imagination. This is an organization-wide business process transformation.”
While ICD-10 is coming quickly, little progress has been made toward transitioning, according to the Workgroup for Electronic Data Interchange (WEDI). Out of nearly 1,000 respondents to a recent WEDI survey, which included EHR vendors, health plans, and physician providers, half of respondents said they were less than halfway through product development, and only 50% of health plans have completed their impact assessments as of March.
Essential to their assessments is how coding and documentation will be impacted. Here, experts emphasize that proficiency in computer-assisted coding (CAC) will increasingly need to become mainstream. Physicians will need focused training on clinical documentation to ensure that a sufficient explanation of patient condition and services is available for the coder to be able to assign the appropriate ICD-10 code. Speech recognition is also getting a lot of attention because it offers productivity efficiency to physicians during point-of-care documentation.
The University of Washington Medical Center in Seattle is currently transitioning to ICD-10, and Health IT Outcomes spoke with Sally Beahan and Melinda Tully about how the facility will use speech recognition and CAC tools in its transition. Beahan, RHIA, MHA, is director of health information management, ICD-10 project adviser, and coding and clinical documentation improvement officer at the university medical center. Tully, MSN, CCDS, CDIP, is the senior vice president of clinical services and education at J.A. Thomas & Associates, a Nuance Company, and a consultant to the university medical center.
Q: How can speech recognition technologies be applied to support a healthcare provider’s ICD-10 and CAC initiatives?
Tully: CAC depends on the documentation that is in the chart. If you’re looking at CAC in the current fashion, then the benefit with speech recognition is to have the documentation available now. Speech recognition allows documentation to occur at the very time the patient is being treated.
CAC only codes what the physician has documented. Physician documentation is generally not known to be complete. Often, it doesn’t tell the true critical story of the patient. CAC can only code what it’s presented with, so if it’s inconsistent or not clinically focused, the CAC engine will be challenged.
Beahan: The physicians have been asking for speech recognition for a long time, even before ICD-10. What we keep hearing is that the EMR has really slowed them down. They are looking for a tool to help them gain back some of the productivity that the EMR has impacted. We’re looking at speech recognition as that tool. Additionally, dictation is expensive, and we have a lot of it. We’re hoping that by expanding our use of speech recognition for documentation that we can offset costs by reducing our dictation volumes.
Q: Where can speech recognition have the most impact on an ICD-10 transition?
Tully: Every facility needs to make sure that its critical documentation explains why the patient is in the hospital and captures the severity of the illness. Speech recognition makes it easier for the physician to actually document and put the documentation in the chart.
Before CAC, critical documentation integrity must come first, and you have to make sure it’s robust and accurate. It can be enhanced with speech recognition and further enhanced if you have clinical language understanding that actually asks physicians for clarification of medical records at the point of documentation — at the point of care — rather than pushing that to a retrospective process.
Beahan: It’s not so much the where, but the who. Our residents do a lot of the documentation, and the attending physicians sign off on it. So, when you have a population of residents that are rotating through the system every few months, managing who will have a license and when to redeploy is daunting.
Part of our challenge is trying to figure out how widespread to deploy it, and so we’re deciding if we will go with a subscription-based model or purchase licenses. Five years ago, we purchased some individual licenses of Nuance’s Dragon Medical 360 as part of our initiative to decrease dictations. We deployed about 300 licenses and found that without specific training for the providers using the tool, some of them got frustrated and didn’t use it. We’ve been recycling those licenses over the past few years as we make determinations about who is using it and who is not. That has added to our challenge of managing speech recognition.
Our future strategy is to have a way to better manage usage, and if someone isn’t using it, we can redeploy that license. Also, we’re looking at providing training up front when they first get the tool. That way we can ensure they truly use it and decrease their amount of dictation.
For now, we are concentrating on our population of attending physicians who use the licenses. We know that Dragon Medical has been successful helping physicians with their direct entry and navigating through the EMR, so that is where we are thinking of focusing speech recognition.
Q: What best practices do you recommend healthcare providers adopt when transitioning to ICD-10?
Tully: The first step is to provide education and awareness to the physician staff. Physicians need to be educated on why ICD-10 is important. If they don’t buy in as to why it will benefit patient care in the long run, then why would they want to bother documenting for ICD-10? Once you start teaching them, then any of the ICD-10 technologies that you give them should enhance their documentation at the point of care. Speech recognition comes into play here. The EHR can have ICD-10 smart notes, templates that facilitate procedures for documentation.
Beahan: Don’t implement too many things at once. If you have EMRs in place, then implementing speech recognition before October 1, 2014, would be a good idea. This will give physicians the tools they need to reduce the impact of productivity loss they are experiencing with EMRs. But couple that with knowing where your documentation gaps exist so you can get templates built to drive the correct documentation. Don’t go live with speech recognition on October 1, 2014, though. We’ve learned that you don’t want to deploy speech at the same time you deploy an EMR. Do it later. Get the physicians used to the tool and then implement speech recognition so you’re not changing too much all at once.
Q: What challenges are healthcare facilities commonly encountering in the transition to ICD-10?
Tully: Part of the challenge is finding somebody to head up the ICD-10 transition team and lead it through the transition. Some have hired a project manager. Success is also dependent on determining a plan for training and education and then executing the plan. Perform a detailed assessment of current-day coding staff knowledge, identify the specific groups that need training (role-based), identify at what level training needs to be provided to all categories of users, and determine a method and timeline for training.
You don’t want to suffer from any drop in revenue as you transition to ICD-10. This means a strategic approach to documentation integrity and coding needs to be established to ensure accurate Medicare severity/diagnosis-related group assignment; appropriate severity of illness and risk of mortality capture; correct hospital-acquired conditions, patient safety indicator, and present on admission reporting; and bullet-proofing documentation against external audits.
Beahan: The biggest area of concern is documentation and making sure we have the specificity to complete our coding and send out our bills in a timely manner. We worry a lot about coder productivity and the impact that ICD-10 will have. We also don’t know how ICD-10 will impact revenue and our bottom line.
We have four hospitals and nine entities as part of UW Medicine, but not all of those facilities use the same EMR. Currently, we rely on documentation templates in the various EMRs, so the goal is to identify what templates are being used and update and streamline them if possible. But it can be difficult to standardize them across all four hospitals.
To help with that, we are concentrating on our specialty areas, starting with orthopedics. We assembled an ICD-10 coding team to look at the documentation templates and identify gaps between ICD-9 and ICD-10. The goal is to incorporate those ICD-10 pieces into the templates so we can gather the level of specificity that we need, but it’s a daunting project.
Q: What other advice can you offer about the transition to ICD-10 and use of speech recognition?
Tully: Begin now. Use available technology that improves physician workflow and improves clinical documentation integrity. Speech recognition technology facilitates clinical data that is intelligently and easily captured as a part of every clinician’s normal workflow. It doesn’t involve asking physicians to change the way they practice.
Make sure your CDI program will support documentation requirements in ICD-10. Use technology that improves coder productivity and accuracy, such as CAC. The transition to ICD-10 cannot occur in a siloed fashion. Once you’ve transitioned, part of the challenge is looking at the data and identifying gaps. It won’t stop on Oct. 1, 2014 — the aftermath will be just as important as preparing for the transition. And once you’ve made the transition, how do you maintain it and make sure you’re doing it well?
Beahan: If you haven’t started yet, you need to. Starting early has allowed us to be thoughtful about our approach. There is a perception among physicians that ICD-10 won’t impact them, so we have had to educate them on exactly how they will be impacted and how we will rely on them to give us the documentation we need to code and bill for the patient visit.
Editor's Note: For more information on easing the transition from ICD-9 to ICD-10, visit the HIMSS/WEDI ICD-10 National Pilot Program, an industry-wide healthcare collaborative that was established to minimize the guess work related to ICD-10 testing and share best practices from early adopter organizations.