Case Study

AI-powered Automated Medical Coding Platform

A major US-based healthcare tech company automates the medical coding process to transform provider and patient experience.

A leading healthcare tech company in the US enables providers to deliver a superior patient experience by automating their medical coding processes that ensure the most appropriate care delivery.
Industry: Healthcare IT/Healthcare Tech, USA

Technology Stack: Web UI: Django, Web UI,
Database: MongoDB Image-to-Text, Conversion: OCR, ML Models for Classification, Data Flow: JSON Parsing, AI Models: Named Entity Recognition, Sentence Detection

Client Overview
Our client is a leading healthcare tech company in the US empowering providers to leverage new-age IT solutions that help them deliver quality outcomes and derive maximum value.
Business Challenge
Delivering the best care and patient experience was the top goal and also the biggest challenge for most providers approaching our client. As per the healthcare tech company, providers were struggling to ensure appropriate care delivery due to the usage of legacy processes. Manual medical coding, in particular, seemed to be a common pain point. Inaccurate ICD-10 code mapping was impacting the right care delivery and positive health outcomes. Incorrect patient risk scores assigned due to the shortcomings of manual coding were standing in the way of providers achieving their patient-centric goals. This required immediate attention as it was impacting providers’ financial performance.
Solution Offered
When the healthcare tech company approached us with this challenge, we started with the discovery exercise as the first step in the service design and delivery process. We had done our research around understanding the problem statement in-depth – what they had been struggling to achieve, how they had been functioning so far, and what was their broader business objective.
The in-depth evaluation revealed that it was the manual and tedious ICD-code mapping process that was leading to errors and inconsistencies and causing inappropriate care delivery.
Combining our experience and expertise in healthcare with the insights drawn from this detailed evaluation, we proposed a strategy that defined how the client could overcome their business challenge. Following this, we presented a proof of concept to demonstrate the potential value of our solution.
Assured by the evaluation and proof of concept provided, the client awarded the project to us. Our solution was an AI-powered automated medical coding platform that enabled healthcare providers to empower their medical coders to code more efficiently and accurately.
Powered by Natural Language Processing, our solution was able to understand the content of the EHR document and compare it against that in the ICD Tables to find the most relevant ICD code for the diagnosis that is presented in the EHR document. Additionally, leveraging the cognitive automation capabilities, it could extract medical data automatically from discharge summaries, test reports, free-text clinical notes by doctors, and other such narrative data. This saved both time and effort for the medical coders.
Highlights of what the new AI-powered medical coding solution delivered –
  • Optical Character Recognition (OCR): Extraction of medical data embedded in various types of files such as jpeg images, tables, charts, pdfs, etc.
  • Information Mining: Automatic classification, identification, and extraction of ICD code-related information from the key phrases referring to the ailment in the EHR document.
  • Automatic ICD Code Mapping: Mapping of extracted medical phrases to relevant ICD-10 codes through a presentation of the top N contextual matches with confidence scores in a Ul.
  • Quality Assurance: Identification of medical coding gaps, inaccuracies, and inconsistencies, giving medical coders the opportunity to make necessary rectifications and assign accurate patient risk scores.
Value Delivered
  • Improved turnaround time for information extraction and mapping to ICD-10 Codes, making the entire coding process more accurate, agile, and efficient.
  • Helped providers to document accurate patient risk scores on claims submitted to health plans.
  • Ensured appropriate compensation to patients and positive value delivery, in line with the provider’s patient-centric goals.

Want to learn more about our AI capabilities and projects?
Connect with our experts.

AI-powered Automated Medical Coding Platform