A custom-built natural language processing system helped one hospital network reclaim 16 hours of clinical staff time per week by automating the extraction of diagnoses, medications, and procedures from unstructured physician notes. The solution, called MedLex AI, achieved 92% accuracy on clinical entity extraction and reduced patient record analysis time by 4.2 times compared to manual review.

What Problem Does This NLP System Solve?

Clinicians at the hospital network faced a familiar challenge: critical patient information was buried in unstructured text scattered across multiple electronic health record systems. A drug allergy noted in an old discharge summary or a diagnosis buried three pages into an intake form could easily be missed during patient appointments. Medical coders also spent hours manually cross-checking documentation against ICD-10 and CPT codes, the standardized billing codes used in healthcare, before every billing cycle.

The core problem wasn't a lack of data. It was that the data existed in a form humans had to read, interpret, and manually organize. "Hospitals don't need another dashboard, they need someone to read the notes for them," explained Mangesh Gothankar, Chief Technology Officer at Signity Solutions, the company that built the system.

How Does MedLex AI Extract Information from Clinical Notes?

MedLex AI uses a multi-layered approach to turn unstructured clinical documentation into structured, searchable data. The system combines three core components:

• Clinical Named Entity Recognition: A specialized engine trained on medical terminology that identifies and extracts diagnoses, medications, allergies, and procedures from physician notes and discharge summaries.
• Retrieval-Augmented Generation Search: A semantic search layer that allows clinicians to query patient records using plain language instead of keyword matching, cutting record retrieval time by 3.6 times.
• Coding Assistant: A tool that flags candidate ICD-10 and CPT codes with confidence scores for human review, improving coding efficiency by 54%.

The system integrates directly with the hospital network's existing electronic health record system, so results stay within familiar clinical workflows rather than requiring staff to switch between platforms.

What Were the Measurable Results?

The deployment delivered concrete, quantifiable improvements across multiple dimensions. Clinical staff recovered an estimated 16 hours per week previously spent on documentation review, a 67% reduction in manual work that the client's finance team tied to approximately $182,000 in annual administrative savings.

Beyond time savings, the system improved data quality and workflow efficiency. The semantic search capability cut record retrieval time by 3.6 times, allowing clinicians to find relevant patient history faster. The coding assistant improved coding efficiency by 54%, reducing the back-and-forth between clinical documentation and billing departments.

Importantly, the system maintained human oversight. "The entity extraction and coding suggestions still get checked by a person. The first pass just takes minutes instead of hours now," Gothankar noted. This human-in-the-loop approach meant clinicians retained control over critical decisions while automation handled the time-consuming initial review.

Why Does This Matter Beyond One Hospital?

The success at this single hospital network points to a broader opportunity across healthcare. Amit Dua, Co-Founder and CEO of Signity Solutions, observed that the unstructured-data problem MedLex AI solved exists in nearly every hospital's records room. "MedLex AI proves the pattern works at one site. The bigger opportunity is what happens when hospitals stop treating clinical notes as paperwork and start treating them as data," Dua stated.

Amit Dua, Co-Founder and CEO of Signity Solutions

This shift in perspective represents a fundamental change in how healthcare organizations can leverage the information already embedded in clinical documentation. Rather than viewing notes as compliance artifacts that must be filed and occasionally retrieved, hospitals could treat them as a rich data source for operational efficiency, billing accuracy, and clinical decision support.

Signity Solutions built MedLex AI to meet HIPAA, GDPR, and SOC 2 compliance requirements, addressing the regulatory concerns that often slow AI adoption in healthcare. The company has delivered more than 1,000 projects across healthcare, fintech, retail, and other industries, with teams in Mohali, India; North Brunswick, New Jersey; and Auckland, New Zealand.