Linguamatics NLP-Based Phenome Extraction from the EHR

On October 24 Benjamin Darbro, MD, PhD, Associate Professor of Pediatrics, Stead Family Department of Pediatrics at the University of Iowa, and Alyssa Hahn, doctoral student in the Interdisciplinary Graduate Program in Genetics at the University of Iowa, will present the webinar "The Use of Natural Language Processing to Improve Phenotype Extraction for Precision Medicine.”

How does NLP support precision medicine and improve patient clinical care?

Precision medicine focuses on disease treatment and prevention, taking into account the variability in genes, environment, and lifestyle between individual patients. In order to understand the best treatment pathway for a particular patient or group of patients, it is important to be able to access and analyze detailed information from the medical records of patients, and ideally broader aspects beyond their medical history.


Global Alliance for Genomics and Health (GA4GH) estimates that greater than 60 million patients will have their genome sequenced in some healthcare related scenario by 2025. 

Million Veteran Program is one such example where a research database will be assembled to anonymously study conditions such as diabetes and cancer, as well as military-related illnesses, like post-traumatic stress disorder (PTSD). As of July 2019, more than 770,000 veterans have contributed blood samples and health data. Just think about the ways this precision medicine initiative changes how we could approach clinical care! How did this revolutionary change in medicine get started?

Evolution from peas to Precision Medicine

I don’t believe that Johann Gregor Mendel would have any idea that his work with pea plants could have evolved into a vast area of medicine that had the potential to make such a revolutionary impact on healthcare. It all started in 1854, when Mendel began his work to look at the conveyance of hereditary traits. “Why peas?”, you may ask. Peas have numerous distinct varieties, and generations that could be quickly and easily produced. Of course as with any new theory, Mendel’s work met with skepticism and controversy and it was not until many years after his death that he was crowned with the title “ the father of modern genetics.” 


Physicians at breaking point

Unsurprisingly, physicians who are constantly under peak pressure have the highest rate of burnout with an average of 45.8%. However, the source states emergency physicians claim a whopping 60% burnout rate. I also recently received an unverified Tweet about the life expectancy for physicians in this specialty, and the news just gets worse. It’s almost 20 years less than other specialties. I am unsure if it’s that much however, if you have ever ventured into an emergency department you can see for yourself why this may be true.


NCQA Digital Summit workshop - streamlining HEDIS reporting with NLP

Fractured Fairy Tale - the Price of Quality

Recently, an esteemed colleague pointed out an eye-opening research article to me when we were on the subject of Quality Measures and the expenses that occur in the digital age: "US Physician Practices Spend More Than $15.4 Billion Annually To Report Quality Measures".

This article was published in 2016, however I am willing to wager that this annual expense has not gone down in the past couple of years. This expense is accrued by not only hospital care organizations (HCOs) but by the insurance companies (payers) as well, all in the name of trying to make our population healthier.

We know that time equates to money in the workforce. How much time does this reporting take on the clinical side in addition to required duties for patient care? No wonder we are facing a clinician burnout epidemic. Medscape’s 2019 report determined that 44% of physicians described themselves as being burned out. And this report only mentioned the physicians, on nursing.org statistics nurses only reported a burnout rate of 15.6%. Which almost sounds like a relief until you learn 41% of nurses reported they felt “unengaged”. How frightening would it to be to be under the care of a nurse that was “checked-out” of his/her job? Burned-out and checked out. How do we achieve better outcomes this way? And how do we know payers are obtaining the correct information from the clinical staff?


Patient safety is an issue that healthcare organizations (HCOs) must prioritize – but how can they improve efficiency when it comes to reviewing the 80% of relevant patient information that is locked in unstructured data?

Under pressure to provide value-based care and adhere to quality measures, HCOs are increasingly turning to AI-based technologies such as Natural Language Processing (NLP), which makes unstructured data usable – thereby improving the efficiency of quality initiatives, quality measure reporting and, most importantly, patient safety.

Addressing the healthcare safety and quality challenge with NLP

While the U.S. health system has made progress in recent years, patient safety continues to be a challenge that all HCOs must prioritize. An estimated 1.7 million healthcare-associated infections occur each year in the U.S. leading to 99,000 deaths. Moreover, adverse medication events cause more than 770,000 injuries and deaths each year at a cost as high as $5.6 billion annually, according to statistics cited by the Center for Patient Safety.

NLP workflows can help reduce the likelihood of error and improve patient safety by automating the identification and extraction of key concepts from large volumes of clinical documentation. Findings are transformed into structured data to simplify chart review and speed the identification of high-risk patients.