If you have ever used Siri or Alexa, you have already seen the efficiency of Natural
Language Processing (NLP). NLP is an area of computer science used to extract
meaning from everyday scenarios, turning them into actionable insights. To do this,
NLP converts large amounts of speech and text into the coding language that a
software system can understand. Then, this new type of data is analyzed to return
insights to the user, or make predictions for future applications. One exciting use
of NLP is in the healthcare experience, especially to improve EHRs, mitigate risks
and to educate the public. MarketsandMarkets predicts that industries buoyed
primarily by NLP and its applications will with be worth $16.07 billion by 2021; with healthcare a key player in this estimate.
Electronic Health Records (EHRs) and Medical Coding
During a patient visit, doctors take notes in the EHR system which end up being lost
in the sea of patient information. However, with NLP, this information is translated
through language recognition to provide physicians with a summary of the patient’s medical records. Not only can NLP be used to analyze old health records to extract overlooked information, but it can improve future clinical documentation as well. For example, doctors can dictate their comments and have their speech converted to notes in real time! This improved quality of documentation frees the doctor to focus more on the patient, and less on the note-taking.
Along with the increased quality of EHRs, comes increased depth and complexity. This creates an area of need - to make the search of health records quicker and more efficient. With access to every detail of a patient’s medical history, doctors can diagnose the patient correctly, and give the correct medical code. In the future, NLP will be able to take these detailed medical records and apply the proper code to the diagnosis with ease.
NLP analysis is also being used to identify and manage risk factors, as well as forecast complications, for multiple chronic conditions. Using NLP, a pilot program identified 8,500 patients who were at risk for developing congestive heart failure at an 85% accuracy rate. Another study similarly used NLP to analyze free-text from the previous 24 hours, to predict readmission and mortality rates for hospitalized heart failure patients and increased specificity from 82.7% to 97.5%. NLP data in hospitals can also expedite the patient identification process for clinical trial enrollment to provide clinicians with the best possible candidates.
Due to NLPs ability to understand massive amounts of language, the system can be used to summarize lengthy journal articles to provide patients with educational materials. The information can also be specifically tailored to the patient and their needs. Also, doctors can share the consolidated EHR materials that they obtained through NLP directly with patients so that they can see their medical records.
The development of natural language processing for healthcare applications continues to grow, despite the risks of potentially costly grammatical and semantic errors. Future NLP will incorporate intelligent systems that learn from experience, so that the more they are deployed, the more accurate and useful the system will become.