--- title: Leveraging LLMs for Streamlined Data Extraction in Healthcare description: Discover how AutoChart leverages LLMs and RAG to automate data extraction for NSQIP, dramatically reducing manual review time while improving accuracy. image: https://cognome.com/hubfs/AI-Generated%20Media/Images/The%20image%20features%20a%20nurse%20looking%20intently%20at%20a%20screen%20displaying%20a%20patients%20chart.jpeg --- [Skip to content](https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare#main-content) [![Smarter Safer Secure AI for Healthcare](https://cognome.com/hs-fs/hubfs/Smarter%20Safer%20Secure%20AI%20for%20Healthcare.png?width=2000&height=600&name=Smarter%20Safer%20Secure%20AI%20for%20Healthcare.png)](https://cognome.com/) - AI Governance Show submenu for AI Governance - [AI Sniffer™](https://cognome.com/ai-sniffer) - [ExplainerAI™](https://cognome.com/en-us/explainerai-new) - [Data Services](https://cognome.com/ai-ml-services) Show submenu for Data Services - [OMOP Transforms](https://cognome.com/ai-ml-services) - [Self-Service for Research](https://cognome.com/solutions-for-research) - [AutoChart™ - Quality and Population Health](https://cognome.com/autochart) - Insights Show submenu for Insights - [About](https://cognome.com/our-story) - [Blog](https://cognome.com/blog) - [The Learning Health System Podcast](https://cognome.com/learning-health-system-podcast) - [Published Research](https://cognome.com/peer-reviewed-research) - [Partner Login](https://partners.cognome.com/) Open main navigation Close main navigation - AI Governance Show submenu for AI Governance - [AI Sniffer™](https://cognome.com/ai-sniffer) - [ExplainerAI™](https://cognome.com/en-us/explainerai-new) - [Data Services](https://cognome.com/ai-ml-services) Show submenu for Data Services - [OMOP Transforms](https://cognome.com/ai-ml-services) - [Self-Service for Research](https://cognome.com/solutions-for-research) - [AutoChart™ - Quality and Population Health](https://cognome.com/autochart) - Insights Show submenu for Insights - [About](https://cognome.com/our-story) - [Blog](https://cognome.com/blog) - [The Learning Health System Podcast](https://cognome.com/learning-health-system-podcast) - [Published Research](https://cognome.com/peer-reviewed-research) - [Partner Login](https://partners.cognome.com/) - [Contact us](https://cognome.com/contact-us) [Contact us](https://cognome.com/contact-us) Jan 28, 2025, 11:03:13 AM # Leveraging LLMs for Streamlined Data Extraction in Healthcare ![Picture of Adil Ahmed](https://cognome.com/hs-fs/hubfs/Adil.jpeg?width=50&name=Adil.jpeg) [Adil Ahmed](https://cognome.com/blog/author/adil-ahmed) Share: [facebook-f icon](http://www.facebook.com/share.php?u=https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare) [linkedin-in icon](http://www.linkedin.com/shareArticle?mini=true&url=https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare) [Twitter icon](https://twitter.com/intent/tweet?url=https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare) [pinterest-p icon](http://pinterest.com/pin/create/link/?url=https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare) [envelope icon](mailto:?body=https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare) ## The NSQIP Challenge The **National Surgical Quality Improvement Program (NSQIP)** aims to gather patient data to enhance post-operative care and improve clinical outcomes. To participate, hospitals must identify qualifying patients based on specific procedures performed within set timeframes. Traditionally, healthcare staff or data abstractors spend hours manually reviewing patient notes to identify these details; a process that’s both time-consuming and prone to error. While SQL queries can expedite the search by matching specific CPT codes, billing processes aren’t always up-to-date, leaving clinical staff to fill in the gaps by manually reviewing notes. ## Introducing AutoChart At Cognome, we developed **AutoChart** to tackle this very challenge. Building on the concepts discussed in our earlier articles (where we explored issues like [**hallucinations**](https://cognome.com/blog/reducing-hallucinations-in-large-language-models-for-healthcare), [**RAG**](https://cognome.com/blog/large-language-models-are-transforming-healthcare), and domain-specific [**fine-tuning**](https://cognome.com/blog/fine-tuning-clincialbert)), AutoChart uses Language Models to: 1. **Extract Procedure Information**: A Language Model combs through patient notes to pinpoint which procedures were performed. This step replaces manual review and ensures that even if billing codes aren’t recorded yet, critical details aren’t missed. 2. **Assign CPT Codes via RAG**: Once the procedure name is identified by the Language Model, AutoChart relies on a **Retrieval-Augmented Generation** (RAG) approach. We maintain a vector store containing CPT codes, their descriptions and synonyms. The extracted procedure is used to query this knowledge base, returning the most relevant codes. We then pass these shortlisted codes to a **fine-tuned Language Model**, specialized in matching a procedure to the best CPT code(s). This synergy of extraction and RAG-driven code assignment not only saves time but also boosts accuracy compared to manual workflows. ### Why It Matters - **Speed and Efficiency**: Nurses and clinicians spend countless hours hunting for the right codes and patient attributes. By automating the identification and coding process, they can focus more on patient care and less on paperwork. - **Improved Consistency**: Human error and fatigue often lead to inaccuracies. With the right guardrails, our LLM-powered systems are designed to be consistent, ensuring higher quality of data. - **Scalability**: As more hospitals adopt NSQIP or similar programs, automating data extraction becomes essential for scaling quality improvement initiatives across multiple departments or facilities. - **Continuous Improvement:** Our system operates in tandem with healthcare staff, allowing them to review and correct any extraction errors or mismatched CPT codes as they finalize the cohort. Each correction, along with the rationale behind it, feeds back into the model, fine-tuning its accuracy over time. By keeping humans in the loop, we’re not replacing their expertise; we’re amplifying it—making data verification faster and more transparent while continuously refining the model’s performance. ## Beyond NSQIP While NSQIP is a perfect example of how Language Models can accelerate data abstraction for quality programs, the underlying framework is highly adaptable. From **clinical trial matching** to **PHI de-identification**, LLMs can revolutionize a variety of workflows that rely on accurate data extraction. By integrating advanced guardrails—like we described when tackling hallucinations—these systems can achieve both reliability and trustworthiness. If you'd like to learn more about working with Cognome, check our [collaboratives of Healthcare AI & ML Practitioners](https://cognome.com/collaboratives). [AIModel](https://cognome.com/blog/tag/aimodel), [AI](https://cognome.com/blog/tag/ai), [Machine Learning](https://cognome.com/blog/tag/machine-learning), [LLM](https://cognome.com/blog/tag/llm), [AIinHealthcare](https://cognome.com/blog/tag/aiinhealthcare), [AIHallucinations](https://cognome.com/blog/tag/aihallucinations), [AIExplainability](https://cognome.com/blog/tag/aiexplainability), [#RAG](https://cognome.com/blog/tag/rag), [NYSQIP](https://cognome.com/blog/tag/nysqip), [AutoChart](https://cognome.com/blog/tag/autochart), [ChartAbstraction](https://cognome.com/blog/tag/chartabstraction) ## Related posts [![](https://cognome.com/hs-fs/hubfs/AI-Generated%20Media/Images/The%20image%20depicts%20a%20bustling%20exhibition%20space%20at%20the%20HIMSS26%20conference%20in%20Las%20Vegas%20This%20should%20be%20a%20photo%20realistic%20rendition%20A%20large%20modern%20booth%20l.png?height=200&name=The%20image%20depicts%20a%20bustling%20exhibition%20space%20at%20the%20HIMSS26%20conference%20in%20Las%20Vegas%20This%20should%20be%20a%20photo%20realistic%20rendition%20A%20large%20modern%20booth%20l.png)](https://cognome.com/blog/cognome-and-intel-partnership) [Intel](https://cognome.com/blog/tag/intel), [AI](https://cognome.com/blog/tag/ai), [LLM](https://cognome.com/blog/tag/llm), [HIMSS](https://cognome.com/blog/tag/himss), [AIinHealthcare](https://cognome.com/blog/tag/aiinhealthcare), [AIExplainability](https://cognome.com/blog/tag/aiexplainability), [DataScience](https://cognome.com/blog/tag/datascience), [AutoChart](https://cognome.com/blog/tag/autochart) ## [Cognome and Intel Partner to Bring Clinically Validated AI to Healthcare at Scale](https://cognome.com/blog/cognome-and-intel-partnership) ![Picture of James Green](https://cognome.com/hs-fs/hubfs/Jame%20Green.png?width=50&name=Jame%20Green.png) [James Green](https://cognome.com/blog/author/james-green) Feb 23, 2026, 2:32:43 PM Come see us at the Hewlett Packard Enterprise booth 12138 at HIMSS26, March 9–12 in Las Vegas. We... [Read more](https://cognome.com/blog/cognome-and-intel-partnership) [![](https://cognome.com/hs-fs/hubfs/AI-Generated%20Media/Images/The%20image%20depicts%20a%20modern%20hospital%20environment%20bustling%20with%20activity.jpeg?height=200&name=The%20image%20depicts%20a%20modern%20hospital%20environment%20bustling%20with%20activity.jpeg)](https://cognome.com/blog/reducing-hallucinations-in-large-language-models-for-healthcare) [AIModel](https://cognome.com/blog/tag/aimodel), [AI](https://cognome.com/blog/tag/ai), [Framework](https://cognome.com/blog/tag/framework), [ClinicalBERT](https://cognome.com/blog/tag/clinicalbert), [LLM](https://cognome.com/blog/tag/llm), [chatGPT](https://cognome.com/blog/tag/chatgpt), [prompt](https://cognome.com/blog/tag/prompt), [Prompt Engineering](https://cognome.com/blog/tag/prompt-engineering), [AIinHealthcare](https://cognome.com/blog/tag/aiinhealthcare), [AIHallucinations](https://cognome.com/blog/tag/aihallucinations), [AIExplainability](https://cognome.com/blog/tag/aiexplainability), [AIEthics](https://cognome.com/blog/tag/aiethics), [#RAG](https://cognome.com/blog/tag/rag) ## [Reducing Hallucinations in Large Language Models for Healthcare](https://cognome.com/blog/reducing-hallucinations-in-large-language-models-for-healthcare) ![Picture of Christos Kritikos](https://app.hubspot.com/settings/avatar/5676d5e9806e554e5303a7856d374a69) [Christos Kritikos](https://cognome.com/blog/author/christos-kritikos) Jan 13, 2025, 4:47:17 PM Large Language Models (LLMs) are revolutionizing healthcare, offering unprecedented capabilities to... [Read more](https://cognome.com/blog/reducing-hallucinations-in-large-language-models-for-healthcare) [![Hospital setting where LLM automatically abstracts clinical data from patient medical charts](https://cognome.com/hs-fs/hubfs/Using%20Machine%20Learning%20to%20Automate%20Nurse%20Chart%20Abstraction.webp?height=200&name=Using%20Machine%20Learning%20to%20Automate%20Nurse%20Chart%20Abstraction.webp)](https://cognome.com/blog/using-large-language-models-to-automate-nurse-chart-abstraction) [AIModel](https://cognome.com/blog/tag/aimodel), [AI](https://cognome.com/blog/tag/ai), [LLM](https://cognome.com/blog/tag/llm), [LearningHealthSystem](https://cognome.com/blog/tag/learninghealthsystem), [AIinHealthcare](https://cognome.com/blog/tag/aiinhealthcare), [AutoChart](https://cognome.com/blog/tag/autochart), [ChartAbstraction](https://cognome.com/blog/tag/chartabstraction), [Operational Efficiency](https://cognome.com/blog/tag/operational-efficiency) ## [Using Large Language Models to Automate Nurse Chart Abstraction](https://cognome.com/blog/using-large-language-models-to-automate-nurse-chart-abstraction) ![Picture of Christos Kritikos](https://app.hubspot.com/settings/avatar/5676d5e9806e554e5303a7856d374a69) [Christos Kritikos](https://cognome.com/blog/author/christos-kritikos) Feb 26, 2025, 2:28:03 PM The healthcare industry is increasingly turning to technology to improve operational efficiency,... [Read more](https://cognome.com/blog/using-large-language-models-to-automate-nurse-chart-abstraction) - [AI Sniffer](https://cognome.com/ai-sniffer) - [OMOP Transforms](https://cognome.com/ai-ml-services) - [Self Service for Research](https://cognome.com/solutions-for-research) - [AutoChart](https://cognome.com/autochart) - [Privacy Policy](https://cognome.com/en-us/privacy-policy) **COGNOME** - *Smarter, Safer, Secure AI for Healthcare*.  Follow us on [Linked In](https://www.linkedin.com/company/cognome) 800 N King Street, Suite 304 1586, Wilmington DE 19801 USA. © 2018-2026 Cognome, Inc. All rights reserved ```json { "@context" : "https://schema.org", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "Adil Ahmed", "url" : "https://cognome.com/blog/author/adil-ahmed" }, "dateModified" : "2025-02-12T16:56:21.894Z", "datePublished" : "2025-01-28T16:03:13.000Z", "headline" : "Leveraging LLMs for Streamlined Data Extraction in Healthcare", "image" : [ "https://cognome.com/hubfs/AI-Generated%20Media/Images/The%20image%20features%20a%20nurse%20looking%20intently%20at%20a%20screen%20displaying%20a%20patients%20chart.jpeg" ], "mainEntityOfPage" : { "@id" : "https://cognome.com/blog/leveraging-language-models-for-streamlined-data-extraction-in-healthcare", "@type" : "WebPage" }, "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://cognome.com/hubfs/(Machine)%20Learning%20Healthcare%20(500%20x%20500%20px)%20(500%20x%20200%20px).png" }, "name" : "Cognome" } } ```