By John Carter, Vice President, Apelon, Inc.
On February 16, 2012, HHS Secretary Kathleen Sebelius announced an unspecified delay in the U.S. implementation of ICD-10. Reaction to that announcement has ranged from self-satisfied to frustrated. We should ask ourselves what all the fuss is about, and reconsider our priorities.
My company specializes in helping its clients achieve semantic interoperability, with a special focus on terminologies. A few years ago, we considered what to do about the ICD-10 problem. Customers were asking about it, some of our competitors were marketing 9-to-10 transition services, and a host of new companies and products were springing up to help with the transition. In the end, we decided that it wasn’t really our gig. My colleagues and I have spent more than two decades working on data interoperability. Our core interest and focus are on enabling decision support, because we see that as the only significant opportunity to moderate cost increases and address quality variances. We’re not billing specialists, and ICD-10 is (just) another billing system. Sure, it’s got more codes, more detailed granularity than ICD-9, but it’s mostly used to summarize care in order to get paid. We repurposed some software to assist do-it-yourselfers with mappings, and we hope that some of those new customers will become interested in exploring reference terminologies and decision support. We held steady with our long-term mission, and we didn’t bet the firm on the 9-to-10 gold rush.
Instead, we’re focused on semantic interoperability using every tool available. That means that while we are happy to wring every ounce of meaning out of ICD-9 and ICD-10, we’re even more interested in highly expressive clinical terminologies like SNOMED CT and LOINC. ICD-11 will incorporate the ontological framework currently used in SNOMED CT, the largest and most formally structured clinical terminology out there. For all its many flaws and gaps and hiccups, SNOMED CT relentlessly gets better every six months, and the list of governments providing their support also keeps growing. Applying SNOMED CT in one’s daily practice is still rare, but that doesn’t mean it can’t or shouldn’t be done. The recipe is straightforward: start with a well-defined set of documentation and analysis tasks, add in the right set of structured and semi-structured documents, serve in an EMR that provides good usability, perhaps adding a dash of NLP to taste.
There’s still a divide between two models of healthcare documentation and delivery, with ICD-10 (and below) on one side ICD-11 / SNOMED CT on the other. Today, you provide and document care and then, separately, you assign some codes and hope to get paid. With ICD-11, there’s the idea that your clinical documentation is sufficiently structured and comprehensive to get you paid without extra steps or coding gamesmanship. For any provider who can move beyond meeting HHS’s evolving requirements, that’s where the action is: high-quality computable documentation leads to clinical decision support, statistically comparative quality measurement, and the possibility of population-based health improvement.
Tom Sullivan wrote about the pros and cons of skipping ICD-10 and going to ICD-11, and he concludes that it’s probably too big a step to take, like Keith does elsewhere on this blog. I regretfully agree… if ICD-10 is such a burden, there’s no way for the system to even contemplate ICD-11 (since it’s not even available for contemplation at this point). That means the distractions will continue.
The good news is that more and more enterprises are going ahead and implementing better, more structured, more computable documentation systems today, while still supporting whatever billing codes are needed. Systems like Kaiser Permanente, Intermountain Healthcare and Mayo Clinic, among others, have long recognized that good documentation leads to good measurement leads to better care and lower costs. At the same time, health information exchanges are more popular than ever, because even though the business model isn’t clear yet, the value of exchanging clinical information is simply too great to pass up.
The delay in implementing ICD-10 is probably unavoidable, and of course providers must fight to protect their revenue stream. Still, I hope as a nation we’ll continue to find ways to get past the all that and keep moving toward semantic interoperability. As the standards and the terminologies have improved, and as the deployment of electronic health records continues, we find ourselves at a time of momentous opportunity. We have a vision for a health care system that provides decision support, powered by computable documentation, available when and where it’s needed, and the sooner we can put the ICD-10 conversion behind us, the sooner we can get moving!
Following up on John's comments, the only way to move to ICD-11 would be through regulatory action. CMS originally proposed a regulatory process for delaying ICD-10, and then went back to stating "a process" (e.g., sub-regulatory). I suspect they changed their stance after someone pointed out how long a regulatory process could take (just days before, or even after the current ICD-10 deadline). CMS could, without changing the regulation, decide on what they do about enforcement on dates (just as they did for 5010). So, no regulation means no chance for ICD-11, only ICD-10. So, ICD-10 it is no matter what, it's just a matter of when.