Healthcare, which has always been based on the doctor-patient interaction, is nearing the end of Stage 1 meaningful use, and as the industry increases its reliance on electronic health records, it faces a new data challenge.
That conundrum, says Nick van Terheyden, MD, and CMIO at Nuance Communications, is how to reconcile the need for standardized structured data capture with the importance of narrative in patient-doctor interactions.
"When a patient walks into the office...they want the attention of the clinician and unfortunately what the process of data entry and data capture has done is defocus that interaction," van Terheyden told Healthcare IT News.
The start of meaningful use Stage 2 in 2014 for many organizations in will usher the implementation of more rigorous standards for data capture, as well as new rules for using EHRs to collect a wider array of structured data, including demographic information, encounter diagnosis, medications and medication allergies, and lab results.
Documenting data in a structured format to meet the standards for meaningful use Stage 2 will be important for allowing physicians to capture detailed and accurate information on their patients, but also for maintaining that detail and accuracy when moving patient data to other settings. This means that when patients have to move from provider to provider within the healthcare system, they can do so without having to input their information repeatedly whenever they enter a new clinical setting. The idea is to provide better and more efficient care.
However, physicians are used to taking their own free-text notes from patient interactions, and it has been their preferred mode of data capture for a long time.
Viet Nguyen, MD, and CMIO of Systems Made Simple, notes in a June 28 interview with Healthcare IT News that doctors have felt that additional structured documentation was an onerous process. But, "when they recognize - and they are recognizing - the value of this codified information for retrieval and aggregation of data, they'll adopt it more."
Having patient data codified under the SNOMED CT or LOINC terminologies, both of which are called for in Stage 2, allows doctors to query databases and recall patient information quickly. Structured data capture can streamline a physician's workflow, allowing them to give more informed care via efficient technological interaction.
The intent of this portion of the Stage 2 requirements is that physicians will become increasingly better at meeting the structured data objectives, which means maintaining more complete, informative patient records.
Nguyen noted that doctors tend to feel a sense of attachment to their patient records since they used to manage their own paper charts. However, "The reality is, physicians and clinicians are the stewards of the data of the patient, and so we need to do whatever we can to make that data useful for the care of the patient, and that means we want to be able to share it with authorized and appropriate other parties by using structured data," he said.
Patient stewardship must also always mean focused interaction with the patient. While physicians maintaining their personal records isolated patient information, technological interaction that is too complex could depersonalize it, turning patients into bits of data, rather than cohesive stories with context and nuance.
Patients "don't enter in with a multiple-choice questionnaire printed out on their chest and we [the clinicians] check boxes on it. They come in with a story," said van Terheyden. Clinicians should operate in a narrative, or unstructured, world.
Doug Fridsma, MD, chief science officer and director of the Office of Science and Technology at ONC, says both structured and unstructured data are important and physicians should include both on medical records. "Medicine is filled with examples of when narrative and free text is more expressive in describing the clinical encounter," he said in an e-mail response.
That narrative documentation could be useful in successive patient interactions, regardless of whether the patient is revisiting the physician who took those notes or is meeting a new clinician, suggests that bridging the gap between discrete data capture and patient narrative may prove beneficial for both patients and physicians in terms of preserving as detailed a patient report as possible to ensure the integrity of treatment, regardless of setting.
Bridging the gap is the crux of van Terheyden's and Nuance's work with clinical language understanding, an improved version of natural language processing technology oriented toward processing clinical terminology embedded in free text into actionable patient information.
To address the issues of technological difficulty and interoperability, or lack thereof, in EHR usage, van Terheyden has looked into "the potential to offer a tool that transcends all these systems," referring to the diversity of EHRs currently in use, "that would allow you to interact with them in a way that is perfectly natural and all the complexity is hidden from you...I'm referring to speech."
To that end, Nuance has developed Florence (with a tip of the hat to the iconic nurse, Florence Nightingale), a virtual assistant focused on helping doctors simplify their interactions with technology so they can spend more one-on-one time with patients. The Florence prototype utilizes speech recognition technology and CLU, making it possible for physicians to engage the EHR with voice commands.
For van Terheyden, the answer to allowing physicians and hospital staff to better interact with EHR technology, even when transmitting data across disparate clinical systems, is "creating systems that are so easy to use that you just can have them all being used and you don't need a lot of training." "We're obviously at a significantly distant point from that utopia," he said.