Text Analytics - a game changer in claims and clinical review
Text analytics or text mining is an artificial intelligence (AI) technology that uses natural language processing (NLP) techniques to examine and analyse unstructured text in documents and databases, extract relevant information from it, and finally transform it into structured data that can be used for analysis or to drive machine learning (ML) algorithms. The structured data finally created can be integrated into data warehouses or business intelligence dashboards and used for prescriptive or predictive analysis. Text mining zeroes in on facts and relationships that would otherwise stay hidden. Advanced text mining applications have been developed in a variety of fields such as insurance (fraud detection) and medical research. In fact, text analytics in the healthcare segment is the fastest-growing segment because of the constantly increasing demand to process big data and procure valuable insights from it.
The healthcare industry today is overloaded with data, and almost 80 per cent of it is unstructured. It includes a variety of clinical notes, electronic health records, medical publications, emails, medical imaging reports, audio, and video files, and lots more. Efficient utilisation of this data can improve operational effectiveness and patient outcomes across the world. Claims management is another pressing challenge of healthcare. Digitising every step of the claims process can potentially streamline the industry, boost its accuracy, lower costs, and offer customers better experiences.
End-to-end digitising is, however, not yet a reality, but the ability to digitise large parts of it is a step in the right direction. The roadblocks to digitisation include privacy and data security concerns, complications because of multiple stakeholders and regulatory constraints. However, more and more consumers are demanding digital and other user-friendly platforms that allow them to manage every stage of their healthcare journey. This also allows companies to stay networked and connected with consumers.* And this demand is not being made by the millennials and gen Z only; even gen-Xers and baby boomers are asking for it.
Optical Character Recognition
AI and ML are key to meeting most of these demands. Optical Character Recognition (OCR) is an AI-backed tool that supports interpretation of unstructured data such as handwritten text or scanned images, extracts relevant data and presents it in an understandable form. Clinical notes, patient intake forms, discharge summaries, medical history notes and test records are sources of unstructured data. AI-OCR uses a variety of Deep Neural Networks (DNN) techniques to extract the relevant data and feed it to an NLP pipeline for further analysis.
Uses of text analytics in the healthcare industry
- Control fraud and abuse: Text data is analysed to establish patterns and any abnormal or unusual patterns or claims are easily detected. Fraud and abuse can be identified, and inappropriate prescriptions, referrals, and insurance and medical claims can be effectively highlighted and controlled.
- Healthcare management: Text analytics is used to track high-risk patients and patients with chronic disease and the information can be used to design appropriate interventions so that hospital admissions and claims can be reduced. Text analytics can also be used to assess hospital readmission data and resource utilisation, and to compare that data with current scientific literature. This can lead to the development of better diagnosis and treatment protocols.
- Treatment effectiveness: Text analytics can be utilised to evaluate the effectiveness of medical treatment by comparing the causes, symptoms, and courses of treatment. Steps to reduce the risks of affliction can be determined too.
- Patient profile analytics: Text analytics can also be used to extract patient data and draw up holistic views of patient information. Such profiles can be used to determine patients who need lifestyle changes and those who would benefit from proactive care.
7 basic steps of text analytics
Text analytics engines break down phrases and sentences in text data before analysing. Very briefly, the seven steps involved in the process are:
- Language identification: Since every language in the world has its own specialties, the language of the data must be identified first. Language defines the whole process for every function so language identification is crucial.
- Tokenisation: In this step, text documents are broken into tokens or individual units of meaning. Punctuation and other language-specific tokens are identified too.
- Sentence breaking: Identifying where exactly a sentence ends is important. For example, there’s a difference between the period after ‘Dr.’ and one at the end of a sentence.
- Part of speech tagging: The part of speech every token is aligned with is identified in this step. And no, it’s not as simple as it sounds.
- Chunking: Also called light parsing, this step breaks up a sentence into its component phrases.
- Syntax parsing: One of the most computationally intensive steps in text analytics, syntax parsing determines the actual structure of a sentence.
- Sentence chaining: This is the final step in preparing unstructured text for deeper analysis.
These steps are followed by using text intent recognition programmes to classify text according to what the writer intended to achieve. This is an essential component of chatbots and is used in customer support by many industries, including the claims management industry.
Need of the hour
Healthcare-related organisations and AI-organisations must work in tandem to develop more sophisticated AI- and ML-based tools to take advantage of the large quantities of unstructured data that is collected and stored. All data are invaluable sources of information for future benefits.
* For organizations on the digital transformation journey, agility is key in responding to a rapidly changing technology and business landscape. Now more than ever, it is crucial to deliver and exceed on organizational expectations with a robust digital mindset backed by innovation. Enabling businesses to sense, learn, respond, and evolve like a living organism, will be imperative for business excellence going forward. A comprehensive, yet modular suite of services is doing exactly that. Equipping organizations with intuitive decision-making automatically at scale, actionable insights based on real-time solutions, anytime/anywhere experience, and in-depth data visibility across functions leading to hyper-productivity, Live Enterprise is building connected organizations that are innovating collaboratively for the future.