Clinithink’s Chris Tackaberry and Peter Johnson explains 5 key elements of successful clinical documentation solutions to unlock unstructured clinical data.
There is no question that clinical documentation is becoming increasingly complex, as new quality measures and coding practices are introduced into healthcare. It’s ironic, in fact, that such reform efforts designed to bring clarity to healthcare services have proven to be nothing short of cumbersome and confusing. There are technologies out there, that when used together with clinical documentation solutions, can make the process easier, however. You just have to know what to look for, according to Clinithink’s Chris Tackaberry and Peter Johnson.
“Clinical documentation can be a powerful tool,” said Tackaberry, CEO and co-founder of the UK-based software company. “The problem with a lot of clinical documentation solutions out there, however, is how they structure clinical data. Naturally, the data has to be structured in a way that’s easily communicable for computers, but it also has to be adaptive and intuitive for clinician use—therein lies the challenge.”
As physicians with extensive experience in HIT, Tackaberry and Johnson have long understood the quandary of capturing the data produced during the point of care. The narrative that’s recorded during those clinical practices is often chock-full of valuable information for decision making or diagnosis, for example. However, that information and its value often don’t go beyond a dead-end data field due to the rigid structure of today’s documentation tools.
Johnson, Clinithink’s co-founder and clinical terminology director, explained: “As the care of a patient continues, that unstructured narrative data remains hidden and inevitably trapped in free text that comes from a variety of sources, including transcriptions or blobs, within a structured EMR/EHR. There’s no means of applying further meaning or knowledge with that data from that point on. In our opinion, that data is too valuable to be overlooked.”
Seeking a solution to capture that powerful, yet untapped clinical narrative, prompted Tackaberry and Johnson to set out and build one of their own. Thus, the pair founded Clinithink in 2009 and developed their patent-pending clinical language indexing technology known as CLiX.
CLiX technology offers healthcare organizations the benefits of clinical natural language processing (CNLP) to help them unlock unstructured medical data in order to access the information needed to optimize revenue, improve quality, enhance processes, and improve clinical outcomes.
Having gone through the trial and error of building a comprehensive clinical NLP tool, Tackaberry and Johnson know what it takes to make solution like CLiX effective. We at HIT Consultant asked them for their insight on what key elements contributed to their solution’s success. Take a look at what they had to say to see how your solution stacks up.
Here are five key elements to look for in your clinical documentation solutions:
Starting January 2014, as a proposal for Stage 2 of Meaningful use, EMR/EHR users will be required to use the SNOMED-CT coding system (Systemized Nomenclature of Medicine- Clinical Terms) as part of their documentation processes in the US. The system of medical terms (which provides codes, terms, synonyms, and definitions covering diseases, findings, and procedures) aids in reducing the variability in the way data is captured, encoded, and used for clinical care and research.
Integrating SNOMED-CT into a system will improve documentation accuracy, but only if you can find what you’re looking for.
“Most clinical systems using SNOMED-CT only allow users to pick single (or pre-coordinated) concepts from a list of available terms,” said Johnson. “However, SNOMED-CT was designed to be used in a compositional way, where more complex (post-coordinated) expressions are dynamically created by the system.”
Consider a physician describing a “recurrent sprain of left ankle” for example. A system that only allows users to pick from available concepts would offer “Sprain of ankle” as the best match. Using post-coordination, however, CLiX can automatically combine three concepts to create “Ankle sprain – recurrent, left”. In the post-coordinated version, no information is lost and the meaning is fully processed by the computer.
According to Johnson, only systems having the ability to take discrete ideas and turn them into complex expressions will experience the full advantage of SNOMED- CT and drive maximum clinical utility. Having the ability to take free text and turn it into these powerful SNOMED expressions is something that only CLiX can do, however.
[See also: Understanding SNOMED CT - The What and Why?]
It’s no surprise that having a clinical NLP tool that can take advantage of that all-important narrative is high on Tackaberry’s and Johnson’s list. The pair sees true potential in solutions that can offer clinicians, who are busy and trying to meet challenging -demands in high volume environments, an easy and intuitive means of capturing clinical data.
However, when seeking out the right solution, the system used has to be as efficient as it is dynamic, according to Johnson. “Solutions that can use the clinical narrative at the point of care will only be useful if the technology can seamlessly and swiftly integrate into the documentation process,” he said.
Bearing that in mind, it’s best to steer clear of technology that you can’t use in real-time scenarios.
“If you have to wait a long time for the technology to work, you’re not going to gain efficiency by using it,” Johnson added.
While many documentation solutions put an abundance of data at a clinician’s disposal, it’s not always organized or packaged in clear and consumable fashion. Therefore, ensuring that your documentation tool includes a query framework, equipped with the sophistication to select and collate relevant data, is essential to further analyze and aggregate your data appropriately.
Johnson explained: “Most vendors in this space typically leverage standard reporting tools in order to query and report on data that has previously been structured. So, they might make queries such as, ‘count all instances of code x for patients ages a-b’. Although this might be useful in some cases, it becomes far more powerful if you can ask ‘count all instances of code x, representing disease group y for patients ages a-b.’ Having the right query framework allows you to view the data from a deeper perspective and draw more meaningful conclusions.”
Not unlike SNOMED-CT, ICD-10 coding requirements are just around the corner. Having access to technology that can map from clinical expressions to the more than 68,000 codes of ICD-10 will become an essential element for any sound clinical documentation tool.
Johnson continues to stress, however, that a tool that simply maps to these codes may not be enough to gain efficiency.
“Most clinicians don’t usually think or speak in the constrained language that is often affiliated with coding,” he said. “Again, we think you need a tool that can pull those pieces from the more casual clinical narrative often used at the point of care. Bottom line: The structure to any mapping output should not create more work for the clinician.”
If Johnson and Tackaberry have brought anything to light, it’s that searching for the right clinical documentation tool doesn’t have to be complex if you know what you need to be effective. By using the important elements mentioned above as their guide, they have developed a promising NLP solution with CLiX.
Clinithink gave us four key elements to explore, but the fifth and final element at play here is clearly one of human understanding.
Regardless of what solution you choose, it should be one that is intuitive and adaptive to human interaction, not the other way around. That’s the lesson to take from Clinithink.
“Any time you overlook important human factors when it comes to technology, it’s a mistake. You could be overlooking serious value, as is often the case with clinical narrative. Any information that can help you do your job better is important and not to be ignored,” Tackaberry concluded.