Natural Language Processing Technology

Natural Language Processing Technology Melchor Abejon Artificial Intelligence (AI) is a branch of computer science devoted to emulating the human mind (Wager, Lee and Glaser, p.305, 2013). Today, one common use of AI is integrated into the Google search engine which can suggest alternative keywords when a user types a misspelled word in a chain of keywords. Another form of AI is the Siri for the iPhone. In the field of healthcare, the rising use of Electronic Health Records (EHRs) in many health care facilities entails the use of innovative technologies that likewise incorporate AI to facilitate the transition from paper-based health records to EHRs. One type of AI technology that can be found in healthcare settings is the Natural Language Processing (NLP) technology. The purpose of this paper is to: Define the Natural Language Processing (NLP) technology. Determine the possible benefits the NLP technology will provide to medical professionals and health organizations. Determine the possible disadvantages that could occur to using the NLP technology. Determine the barriers to using the NLP technology. Definition of NLP Technology Wager, Lee and Glaser (2013) define NLP as a program that takes human language (typed as text or input as voice) and translate it into a standard computer instruction (p.306). It is a fact that majority of the clinical documentations that can be found in healthcare today are unstructured and are buried within EHRs. These unstructured information are faulty and redundant and can obstruct the healthcare industrys goal to establishing an efficient and data- driven clinical decision making. And with the emergence of the NLP technology, the extraction of valuable information can be leveraged to create a more informed clinical decision making to improve the quality of patient care and at the same time reducing the healthcare cost. Benefits of the NLP Technology NLP enables meaningful use. NLP integrated in EHRs can significantly assist health providers to easily capture specified health facts such as vital signs, allergies, smoking status and health problem lists via narrative description or voice. These health facts are difficult and time consuming to capture via the EHR system alone. The capture of specified health facts with the aid of NLP technology enables providers to qualify for incentive payments by the federal government. NLP enables predictive care. A more advanced Clinical Language Understanding (CLU) solution or technology can provide immediate feedback to health providers at the point of dictation whether they are using a mobile phone, digital recorder or PDA.ÂÂ   An example is, when CLU technology is integrated and running in the background of an EHR system, the system can notify a physician for adverse drug reaction during the documentation of prescription for a patient, and it would even recommend an alternative medication treatment for the patient. When NLP is integrated in an EHR system, it can develop opportunities for a more efficient Clinical Documentation Improvement (CDI). NLP can assist CDI specialists to quickly perform a comprehensive data mining. Identification of patients for clinical trial enrollment can be accelerated. NLP can assist organizations to quickly identify patients who maybe qualified for immunotherapies and clinical trials and research. NLP technology can help organizations comply with the core measures. Immediate assessment of documentation upon admission, and close monitoring of a patient is made possible with the use of the NLP technology. It also enables quick review of program notes and problem lists. NLP technology can provide real-time patient data. NLP can be used concurrently to monitor treatment of patients during their stay in the hospital. The alerts and reminders generated by NLP can help the providers to monitor their patients to mitigate the risk of acquiring infections. NLP technology enables effective billing system. NLP can improve the documentation process thus alleviating a lot of pain from the billing process for health providers and coders. Disadvantages of the NLP Technology Generic searching can be very difficult. Problem with synonyms. The abundance of synonymy in the medical field can be a problem in the use of NLP. Problem with homographs. Homograph refers to words that are spelled the same but has different meanings. Problem with polysemy. Polysemy refers to a word or phrase with many possible meanings. Ambiguous. Difficulty in identifying all of the possible meaning of words or combination of words. Non-standardized and not very compact. User needs to think of own search terminology and or synonym. Barriers to Using NLP Technology NLP is not a new technology in the field of healthcare, but so far its programs have met with limited success. Before it can reach the expected reliability performance, computer and NLP experts are expected to perform a tremendous job to address the barriers to integrating NLP tools in the clinical care. Though NLP is already embedded in products for some EHR vendors, the unstructured narrative texts and clinical notes still pose a major challenge for computer experts. According to Townsend (2013), a clinical text which is often ungrammatical and consists of bullet point telegraphic phases with limited context and lacking complete sentences poses a major challenge to using NLP. Other barriers to using the NLP technology are the poor standardization of data elements, inadequate policies on data governance and the never ending variation in the programming and designs of EHR systems. Conclusion As computer experts continue to seek refinement of the NLP technology, NLP will continue to deliver an important role in the management of health population and data analytics by extracting valuable health information and making them into actionable data to improve healthcare outcomes. Unstructured data in healthcare will remain a major challenge, but as efforts continue to build a stronger information governance and better standardization of data elements, the future of the NLP technology in the healthcare industry looks promising. References Bonney, S. (2015). Practical applications of NLP in healthcare: Enhancing the value of electronic data. Retrieved February 12, 2017 from http://porterresearch.com/practical-applications-of-nlp-in-healthcare-enhancing-the-value-of-electronic-data/ HealthIT Analytics (n.d.). What is the role of natural language processing in healthcare? Retrieved February 12, 2017 from http://healthitanalytics.com/features/what-is-the-role-of-natural-language-processing-in-healthcare Nadkarni, P., Machado, L.O., and Chapman, W.W. (2011). Natural language processing: An introduction. Retrieved February 12, 2017 from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3168328/#b29 Petro, J. ( 2011). Natural language processing in electronic health records. Retrieved February 12,2017 from http://www.kevinmd.com/blog/2011/09/natural-language-processing-electronic-health-records.html Townsend, H. (2013). Natural language processing and clinical outcomes: The promise and progress of NLP for improved care. Retrieved February 12, 2017 from http://bok.ahima.org/doc?oid=106198#.WKJo100rKUm Wager, K.A., Lee, F.W., Glaser, J.P. (2013). Healthcare information systems: A practical approach for health care management (3rd ed ). San Francisco, CA: Jossey-Bass