Primary Health Care: Open Access 2021, Vol.11, Issue 3, 373.

Research Article

NLP/Deep Learning Techniques in Healthcare for Decision Making

Kamal Jain1*, Vishal Prajapati2
1
Data Science Evangelist, Mentor, United Arab Emirates
2
Department of Information Technology, G H Patel College of Engineering & Technology, Anand, Gujarat, India

Corresponding Author* What is Natural Language Processing?

Kamal Jain Natural language processing (NLP) focuses on analysing text and speech to
infer meaning from words. Recurrent neural networks (RNNs) - deep learning
Data Science Evangelist, Mentor, United Arab Emirates
algorithms play a key role in processing sequential inputs like language,
E-mail: [email protected] speech, and time-series data [1]. Deep learning is a subset of machine
learning having the capacity of learning unsupervised data from unstructured
Copyright: 2021 Jain K, et al. This is an open-access article distributed under the or unlabeled data. On the other side deep learning will be able to learn optimal
terms of the Creative Commons Attribution License, which permits unrestricted features from available data without human intervention.
use, distribution, and reproduction in any medium, provided the original author Natural language processing is used to describe the process of using
and source are credited. computer algorithms to identify key elements in everyday language and
extract meaning from unstructured spoken or written input. NLP requires
skills in artificial intelligence, computational linguistics, and other machine
Received 09 March, 2021; Accepted 23 March, 2021; Published 30 March, 2021
learning disciplines [2].
There are two approaches to NLP.
Abstract
1. A rule-based approach where computer follows pre-defined rules in the
program
Day by day the health care domain is generating millions of records of patients 2. Machine Learning-based approach where we can have supervised and
in a structured and unstructured way. By applying deep learning techniques, unsupervised learning methods. In supervised learning, the computer
it can be converted into a well-structured form i.e. Electronic Health Record learns latent rules through human guidance called annotation while in
(EHR). Decision making is one of the key parts in the Healthcare domain. In unsupervised learning there is no human interaction.
the decision making process doctors must refer to many data like laboratory
reports, diagnosis reports, medical images, demographic information about NLP algorithms, first extract information or concepts from EHRs, then process
the patient, clinical notes, and map it collectively with the concepts of medical extracted information, and finally classify patients into a subgroup as per
science. Here AI especially Natural Language Processing and deep learning rules and learners. Procedures for NLP are complex because it consists of
can be helpful in many ways. multiple techniques together. NLP will map phrases or words to concepts of
The objective of this article is to depict the importance of Deep Learning and interest, and it needs careful pre-processing of text and it will be converted to
Natural Language processing EHRs. Based on EHR, the doctor can take quick document from the natural form [3].
decisions in case of an emergency. Apart from that, it can also be effective A few examples of low-level NLP tasks (text pre-processing) include the
in clinical predictions, detect disease at an earlier stage, forecasting future following:
need of regular check-ups, predictions of hospitalization soon if required. The
provision of an enormous amount of clinical information particularly EHR has 1. Sentence boundary detection, which is usually defined by a period
stimulated the expansion of deep learning techniques that assist within the
2. Tokenization (breaking a sentence into individual tokens)
rapid analysis of patient data.
3. Stemming (reducing word into a root form)
Keywords: Natural language processing, NLP, Deep learning, Data science,
Artificial intelligence, EHR, Healthcare, Electronic health records 4. Lemmatization (process mapping a token)
A few examples of higher-level NLP tasks include the following:
Introduction
1. Named entity recognition
The ubiquitous adoption of electronic health records in hospitals and other
healthcare facilities generates vast real-world information, which is very 2. Setting up negation rules
valuable for conducting clinical research. Over the past many years, electronic
The Figure 1 depicts the text pre-processing and classification of an asthma
health records (EHR) systems are been widely adopted across clinics and
patient.
hospitals. Analysis of this huge data is the foundation to provide improved
healthcare to the patients However, manual review of this vast amount of data Role of Deep Learning/NLP in Healthcare
generated from multiple sources is costly and very time-consuming. It brings
huge challenges while attempting to review this data meaningfully. Hence, Basically, in terms of users, we can categorize the healthcare information
the role of artificial intelligence (AI) techniques is becoming important in coming from below four sources -
enhancing clinical research and care. As a large number of EHR are locked in
clinical narratives, Natural Language Processing (NLP) and Deep Learning (DL) 1. Doctors
techniques have been leveraged to extract information. While computer vision 2. Patient
techniques can be used for medical imaging, NLP can be used for analysing
unstructured information in the EHR, reinforcement learning techniques can 3. Paramedical staff
be used in the context of robotics-assisted operations. NLP algorithms
4. Pharmaceuticals
can be used to identify clinically relevant phenotypes while analysing text
and determining the grammatical relationships between phrases. Rule- Every post-process depends on the diagnosis of a disease. If the disease
based NLP techniques can be used to get high sensitivity (identification is identified properly then one can get proper treatment. Sometimes the
of a large proportion of true cases) and high positive predictive value in situation of patients gets serious due to delays in decision making (Table 1).
clinical records.
Figure 2 represents, deep learning is an approach where there is a large scale
Healthcare is one of the domains where computer science is becoming network that is going to accept a variety of input data types like text, image,
very supportive of varied tasks. AI is increasingly being adopted across the audio, time-series data, etc. for each data type learns a useful features in its
healthcare industry ranging from basic level practices to specialization, and lower level towers. The data from each pillar is then merged and flows through
lots of the foremost exciting AI applications leverage language processing higher levels, allowing the Deep Neural Network to reach to conclusion based
(NLP). The capabilities of these AI techniques potentially enable the on reasoning and evidence across data types.
identification of distinctive clinical characteristics among patients which
further helps in clinical care and minimizing methodological heterogeneity in Natural language processing can help healthcare in information extraction,
medical research concerning various health diseases. unstructured data to structured data conversion, Document categorization,
1
Primary Health Care: Open Access 2021, Vol.11, Issue 3, 373. Jain K, et al.

Figure 1: NLP Algorithms for document-level text processing and patient-level classification [3].

Table 1. Disease stage, process required and the actions to be taken.

Sr. No. Situation Process Tools/Devices/Reports/Actions
Health Awareness Smart sensor devices to provide health updates
1 Pre-Disease Regular health check-ups Devices for health check-ups
How to stay away from disease? Training programs and devices for various exercises
Blood sample, X-Ray Image, Urine Sample, Biopsy, Devices for checking vital
Diagnosis
signs in the body (Oxygen, sugar, heartbeat, pulses, etc.)
Pre-surgery treatments, Minor Surgery, Major surgery, Combined surgery, Post-
2 In Disease Treatment
surgery treatments
Drug dosage based on the nature of the body for a single disease, multiple
Medication
diseases
Follow-ups Prediction of the number of follow-ups or time period of follow-ups required
3 Post Disease Improvement comparison Compare Pre, In and post disease reports
Future predictions

and summarization. Ultimately it will reduce administrative cost by efficient c. Sequencing. By temporally sequencing all data into a patient timeline,
billing and accurate prior authorization approval. It also going to create time-based deep-learning techniques can be applied to the entirety of
Medical Value by providing support for ineffective clinical decisions and EHR datasets for making predictions about single patients.
streamlined medical policy assessment etc. [4,5]. Apart from that it will help in
improving patient interactions with the provider and the EHR, increase patient NLP can be helpful in the neurology domain too. The Figure 4 depicts how the
health awareness, improve care quality, and identify patients with critical care iterative process of the NLP algorithm which can be helpful in identifying 24 brain
needed. imaging phenotypes in two areas of NHS Scotland which had excellent positive
predictive value for cerebrovascular and neurodegenerative phenotypes.
In healthcare, sequential deep learning and Natural language technologies
boosts applications within domains like electronic health records (EHRs) [1,6]. The NLP system, Edinburgh Information Extraction for Radiology reports
(EdIE-R), is a staged pipeline process as depicted in the Figure 4, with XML
An outsized medical organization can capture 10 million patients’ medical rule-based text mining software at its core.
transactions in a decade. One hospital alone generates 150,000 pieces of
data. Aggregation of all EHRs of this scale exemplifies 100 million years of Conclusion
patient data and 200,000 years of doctors’ practice, and it covers excess of
rare conditions and diseases [1]. NLP will play a significant role in accelerating the decision-making process in
the healthcare area. However, the real rewards of developing good algorithms
The Figure 3 depicts various stages of processing will depend heavily on the quality of the data that they acquire and maintain.
a. Unstructured EHR data. Medical records are stored in heterogeneous The faster decision-making process will allow physicians to focus on the
formats as per the working mechanism and data structures of a specific value-added care of patients. NLP with Deep Learning and Computer Vision
hospital. So, it will vary from one hospital to another. can process a variety of data together to take precise decisions. Collaborative
research can lead to a higher level of treatment in Healthcare. Considering
b. Data standardization. Based on FHIR, data will be standardized into the impact of AI techniques, these systems need to be designed and built
the same or we can say homogeneous format by mapping data from very carefully in a larger socio-ecological context of clinical care settings to
multiple sources provide better healthcare to society.
2
Primary Health Care: Open Access 2021, Vol.11, Issue 3, 373. Jain K, et al.

Figure 2: Deep learning can detect features and learn from a variety of data types [1].

Figure 3: Making predictions using EHRs [1].

Figure 4: EdIE-R System Architecture [7].

References
3. Juhn, Y., & Liu, H. “Artificial intelligence approaches using natural language
1. Esteva, A., et al. “A guide to deep learning in healthcare.” Nature Med 25 processing to advance EHR-based clinical research.” J Allerg Clin Immunol 145.2
(2019):24-29. (2020):463-469.
2. Bresnick, J. “What Is the Role of Natural Language Processing in Healthcare?” 4. Rangasamy, S. et al. “Natural language processing in healthcare.” McKinsey
Health IT Analytics (2020) Retrieved from https://healthitanalytics.com/ & Company (2018) Retrieved from https://www.mckinsey.com/industries/
features/what-is-the-role-of-natural-language-processing-in-healthcare healthcare-systems-and-services/our-insights
3
Primary Health Care: Open Access 2021, Vol.11, Issue 3, 373. Jain K, et al.

5. “Top 12 Use Cases of Natural Language Processing in Healthcare” Retrieved (EHR) Analysis.” In: Bhoi, A., et al. (eds) Bio-inspired Neurocomputing. Studies
from https://marutitech.com/use-cases-of-natural-language-processing-in- in Computational Intelligence, vol 903. Springer, Singapore.
healthcare/
7. Wheater, E., et al. “A validated natural language processing algorithm for brain
6. Poongodi, T., et al. “Deep Learning Techniques for Electronic Health Record imaging phenotypes from radiology reports in UK electronic health records.”
BMC Med Informatic Decision Making 19 (2019):184.

Cite this article: Jain, K., & Prajapati, V. “NLP/Deep Learning Techniques in Healthcare for Decision Making.” Prim Health Care, 2021, 11(3), 373.