From connected cars to intelligent, voice-enabled laptops, it’s no secret that natural language understanding (NLU) has already started to transform how consumers interact with and leverage technology in a more conversational, natural way. What isn’t quite as widely known is that the NLU equivalent in healthcare — Clinical Language Understanding (CLU) — is a critical enabler in helping providers drive productivity in the new digital era of healthcare and maintain focus on what matters most: patient care.
Something’s Gotta Give: Increasing Pressure on Physicians Driving Demand for CLU
The healthcare industry is juggling far too many changes when it comes to the shift from paper to digital health records mixed with a myriad of regulatory changes — all of which is putting serious strain on doctors. As we continue to ask doctors to do more with less, their focus is forced further away from the physician-patient interaction.
More so, the transition to ICD-10, a new and more complex way of describing and classifying the care a doctor provides a patient, holds the potential to further exacerbate the physician-patient divide. We’re essentially asking doctors to capture more detail in the patient record — which is a critical and much-needed step in the move toward population health management. But, at the same time, many hospitals and health systems are not able to take anything off of their plates, and they aren’t sure what type of technologies and education will drive care team productivity in order to make room for this new demand.
CLU to the Rescue: Getting Back to the Art of Medicine
Customized with medical ontologies and clinical patterns, CLU is enabling doctors to more efficiently address major changes like ICD-10 without adding complexity and hours onto their workday. CLU helps physicians do things faster and easier, much like NLU has done for consumers. The following are some of the key ways CLU will help providers address change, transition to value-based care and refocus their attention on patients in the coming year:
- Simplify Interactions with the EHR – What if you could talk to the electronic health record (EHR) and use your voice to easily navigate the digital record and tackle the headaches of day-to-day digital physician workflow? CLU has the ability to act as an intelligent, conversational, automated assistant that can help doctors with things such as computerized physician order entry or faster navigation within the EHR.
A 2013 study reinforced physician demand for intelligent virtual assistants built on CLU with approximately 80 percent of US physicians stating that virtual assistants will drastically change how doctors interact and use EHRs by 2018. In the coming year, as the healthcare industry looks for ways to simplify technology for physicians, virtual assistants built on CLU and artificial intelligence will become more commonplace.
Balance the Need for Patient Narrative and Structured Data – Over the last year, leading EHR vendors have embraced the need to help doctors efficiently capture the patient narrative (dictated notes) and mandated structured data. In addition to technologies that help drive productivity and specificity in the overall capture of patient information — like speech recognition — there’s also a call to action to extract structured data, such as problems and medications, as a result of the Meaningful Use mandate.
Considering the looming productivity loss associated with the transition to ICD-10, over the coming year, more and more providers will integrate CLU with their EHRs in order to help drive clinical efficiencies. With CLU, physicians simply narrate, press a button and then have facts and evidence returned to them on a reconciliation interface for review and approval. In addition to simplifying the data-entry process for physicians, CLU also helps safeguard clinical data integrity which positively impacts the coding and quality reporting teams further downstream.
Increase Documentation Specificity in Real-Time – CLU fact extraction can go one step further in helping physicians get back to the patient by keeping “watch” on what they are dictating or entering into a patient’s clinical record. Language understanding technology can support computer-assisted physician documentation (CAPD) by analyzing narrative text while the physician is documenting, and drawing attention to areas that may be missing detail or are unclear. With CAPD, fact extraction takes place and then that information is run through a knowledge base that addresses what details are missing in the physicians’ documentation, automatically prompting the physician for additional specificity.
By automating queries while the physician is documenting, the physician is able to reflect and adjust documentation in real-time, bringing the workflow much closer to their clinical thought process than traditional query methods which often come to the physician after the patient has left the building.
Drive Coder Productivity – In order to derive the full value from another CLU-driven technology, computer-assisted coding (CAC), providers must first and foremost ensure that the documentation coders are receiving is fully optimized. By arming physicians with the previously mentioned CLU-based tools that drive documentation accuracy from the start, providers can ensure coders have the specific facts and evidence they need to more efficiently code and bill for the care provided.
With the introduction of ICD-10, coders will essentially need to unlearn the ICD-9 classification system and learn the ICD-10 classification system — a process that will adversely affect their productivity. This impending loss of productivity shines a light on the need to further focus on accuracy upfront, starting with the physician. By taking a physician-first approach to ensuring clinical data integrity, providers will help coders lighten their load by lessening the time coders have to spend following up with clinicians to verify documentation.
What’s Next: Laying Groundwork for Population Health
While many of the clinical hardships associated with the transition to ICD-10 have been addressed throughout, one of the benefits worth calling out is the fact that ICD-10 will provide a better understanding of the general population in terms of burden of disease, potential preventable causes of illness and death, and the appropriate allocation of health care resources.
In order to transition to population health management, healthcare providers must first and foremost get a handle on their unstructured data. To shine a light on how massive a problem this is, it’s worth noting that UPMC previously estimated that 80 percent of their 3.2 petabytes of data were in fact unstructured. Increasing the specificity of patient information captured with ICD-10 (yes, that includes walking into a lamppost), enabling health information exchange, and applying intelligent technology like CLU to all of this data would be a solid first step toward enabling providers to access and derive insight from the information they’re collecting.
Much like NLU has helped drive intelligent, natural interactions between consumers and technology, CLU will help re-humanize healthcare. By enabling physicians to focus on the patient, not the technology, providers can begin to embrace a next-generation approach to healthcare that will drive efficient, intelligent clinical decisions that impact each and every facet of patient care. Source