- The rise of big data in the healthcare industry is setting the stage for natural language processing (NLP) and other artificial intelligence tools to assist with improving the delivery of care.
NLP algorithms have already proven valuable in this venture, largely showing potential in simplifying clinical documentation and enabling voice-to-text dictation.
As health IT tools become more advanced, however, the potential of NLP to improve the care continuum will only grow.
A recent report from MarketsandMarkets indicates that the NLP market is expected to grow at a CAGR of 16.1 percent until 2021, resulting in a $16 billion market opportunity.
What does the future look like for NLP, and what are some key use cases for healthcare organizations looking to leverage these tools?
Enhancing provider interactions with patients and the EHR
Improving the provider EHR experience is a high priority for healthcare organizations.
Attempting to give patients their undivided attention, while also trying to complete burdensome documentation requirements, has left many clinicians feeling drained and dissatisfied.
The issue has become a healthcare epidemic. A recent survey found that 83 percent of clinicians see physician burnout as a problem at their organizations.
NLP tools, such as voice recognition, may offer a viable solution to EHR distress. Many clinicians already utilize this technology as an alternative to typing or handwriting clinical notes.
At WellSpan Health in Pennsylvania, providers are using voice-based dictation tools to improve patient-provider interactions and reduce EHR frustration.
“I’m a primary care provider by background, and when I dictate my notes in front of the patient, he or she gets to hear what I’m saying and make sure that it’s correct,” R. Hal Baker, MD, Chief Information Officer and Senior VP of Clinical Improvement at WellSpan told HealthITAnalytics.com.
“It’s a much more cooperative approach – not to mention a more efficient one. I can talk to both the record and the patient at the same time, so I don’t have to walk out of the room and recount the entire visit again at some later time. That lets me spend a greater percentage of my time in the patient’s presence.”
The organization has found that this approach also improves the quality of the documentation, which may make it more useful for analytics downstream.
In the future, voice recognition tools may go beyond clinical dictation to receive and carry out directions from providers.
Virtual assistants like Alexa, Siri, and Cortana have already made their way into healthcare organizations as administrative aids, helping with customer service duties and help desk tasks.
As the industry refines its capabilities, these tools may soon enter the clinical side of the healthcare industry, taking on roles as medical scribes and ordering assistants.
Google recently began recruiting individuals to help develop voice recognition tools that record clinical documentation, indicating that virtual medical assistants may soon become a reality.
Improving patient health literacy
In addition to easing EHR difficulties for providers, NLP tools may contribute to smoother interactions between patients and health IT tools.
With more organizations using patient portals, patients can now access their health data, make more informed medical decisions, and keep their health on track.
However, the benefits of patient data access are lessened if patients can’t make sense of what their data means.
A 2016 poll found that although 60 percent of patients could access their EHR data, 15 percent had trouble understanding the information, and just 22 percent used their EHR data to make medical decisions.
The issue of limited patient health literacy weighs on providers as well. Physicians must often spend extra time defining terms for patients and soothing the anxieties of those who may have misread a diagnosis or lab test result.
By applying natural language processing to EHR data and integrating the results into the patient portal, providers could improve patients’ understanding of their health information.
In a 2017 study, researchers used NLP tools to match medical terms from clinical documents with their lay-language counterparts. The algorithms outperformed baseline systems in precision when presented with unlabeled evaluation data.
In another recent study, researchers developed an NLP tool to link medical terms to simple definitions to improve patient EHR understanding and the patient portal experience.
After collecting physician feedback, the team made several usability and clarity changes to the system, which significantly improved the algorithm’s ability to recall medical definitions.
While neither study developed a system that could be applied to patient data in a real clinical setting, the initial results of both demonstrate the potential for these algorithms to boost patient EHR understanding in the future.
Contributing to higher quality of care
NLP tools may also offer a more efficient way to evaluate and improve care quality.
Measuring physician performance and identifying gaps in care is a critical competency for organizations making the switch to value-based reimbursement.
Researchers have shown how NLP can simplify the process of benchmarking the professional skills of physicians, automating the evaluation of free text and reducing the amount of time and human effort typically required to complete this task.
A 2017 article from the Journal of Medical Internet Research describes how researchers applied NLP to free-text questionnaires filled out by providers’ peers and found that they agreed with human assessments of the same documents 98 percent of the time.
NLP algorithms could also help providers identify potential errors in care delivery.
For example, in a 2017 study, a research team applied an NLP tool to unstructured data to identify adverse drug events (ADEs) in medical literature and social media postings.
“Discovery of ADEs has gained great attention in the health care community, and in the last few years, several drug risk-benefit assessment strategies have been developed to analyze drug efficacy and safety using different medical data sources, ranging from EHRs to human-health–related social media and drug reviews,” the team explained.
The algorithm achieved 92.7 percent accuracy and 93.6 percent precision, outperforming traditional big data analytics tools and demonstrating its potential to improve care and ensure patient safety.
Additionally, a study conducted in 2018 showed that NLP could help providers measure the quality of inpatient care and monitor adherence to clinical guidelines.
Researchers developed an NLP system designed to extract relevant EHR data and identify whether clinically relevant medications were prescribed to heart failure patients upon hospital discharge.
The system outperformed manual data extraction in sensitivity and could improve the efficiency of quality measurement and enhance guideline-concordant care.
Identifying patients in need of improved care coordination
Machine learning and NLP tools have also shown potential for detecting complex patients who may benefit from enhanced care coordination.
Non-clinical factors such as housing instability and food insecurity can make it difficult for patients to adhere to treatment protocols, and may also make it more likely that these patients will incur more care costs in their lifetimes.
Mental health and substance abuse disorders can exacerbate these issues, resulting in poor health outcomes and increased healthcare spending.
However, data detailing patients’ social determinants of health is often harder to access than their clinical information, and is usually in an unstructured format.
NLP algorithms can offer a solution. By extracting meaningful information from large datasets, these tools can provide clinicians with the information they need to detect complex patients.
For example, researchers at Massachusetts General Hospital applied NLP techniques to the EHR to help providers identify key terms associated with the social determinants of health.
The team found that 22 terms provided enough specificity to reliably identify patients at higher-than-average risk of psychological, social, and behavioral impacts on their health.
NLP can also be beneficial in improving care coordination for patients with behavioral health issues.
Beacon Health Options, a behavioral health management service provider, is using machine learning and NLP tools to mine unstructured patient data and identify those in danger of falling through gaps in the healthcare system.
“Our goal is to move from being a reactive model that solely looks at what has happened historically to being a much more predictive, proactive, and targeted service provider,” Dr. Emma Stanton, Associate Chief Medical Officer for Beacon Health Options, told HealthITAnalytics.com.
“It’s an opportunity to bridge the siloes that exist in the healthcare delivery system, and it’s an example of where machine learning can help to bulldoze through those traditional barriers to make progress for an incredibly vulnerable segment of the patient population.”
While the healthcare industry still must refine its data capabilities before NLP tools are widely deployed within clinical organizations, these techniques have a significant amount of potential to improve care delivery and streamline provider workflows.
In the future, NLP and other machine learning tools could be the key to better clinical decision support and patient health outcomes.