Artificial Intelligence
Transforming a clinical library into a trusted advisor for clinicians
About
Our client is one of the largest end-to-end healthcare providers in America. With increasing patient caseload, particularly following COVID-19, there is a growing challenge of reduced time for patient interactions and rising clinician burnout. HCLTech has collaborated with the organization to create a GenAI-powered clinical advisor – one that is trusted by healthcare specialists.
The Challenge
The client had made significant investments in building a digital, clinical library – a vast trove of medical research documents, vetted academic papers and standard operating procedures (SOPs).
The objective was to assist its clinicians during patient interactions by providing them with cutting-edge medical research and standardize medical treatment. However, contrary to the expectation that this clinical library would serve as a one-stop reference for clinicians, it averaged only one visit per clinician per day.
Clinicians were using ChatGPT or similar GenAI enabled tools – a serious concern given the highly regulated nature of the industry, access to PII data (Personally Identifiable Information) and implications on treatment protocols.
- Ever-growing burden on clinicians
- Reduced patient interactions
- Increased clinician burnout
The Objective
A trusted, advanced technology solution that transforms the clinical library to an interactive clinical advisor.
The main goals were
- Enhance usability: Equip the existing clinical library with a search engine like interface for quick and timely access to critical information
- Improve efficiency: Summarize patient medical history and offer evidence-based recommendations
- Trusted: Ensure system does not hallucinate nor offer recommendations based on parametrical knowledge
- Compliant to regulations: Anonymize and secure PII data
The Solution
Improving the usability of the clinical library with GenAI.
HCLTech created a two-part solution to make healthcare more efficient.
- On-prem processing: Patient data and medical history are processed locally using SLMs (Small AI models) on specialized computers. This helps patients avoid repeating their medical history multiple times.
- Cloud-based processing: LLMs (Large AI models) in the cloud answer doctors' questions and give real-time advice.
This method follows HIPAA and GDPR rules to keep patient information safe and private.
We designed an advanced retrieval, augmentation, and generation (RAG) engine, which minimizes the risk of misinformation by accurately retrieving relevant medical data. The AI-powered clinical advisor comprises several additional features, including:
- Develop prompts that save specialist clinicians time and ensure precise responses
- Build trust and explainability through clickable citations bolstering confidence in treatment recommendations
- Allow clinicians to generate high quality personalized responses using preferred formats and instructional styles
HCLTech Clinical Advisor powered by NVIDIA
Built on NVIDIA AI Enterprise, this innovative platform harnesses the power of GPU-accelerated technologies and combines NVIDIA's advanced AI tools with intelligent workflows explicitly tailored for healthcare. It integrates technologies such as TensorRT-LLM and Triton Inference Server, ensuring high-performance inference for large language models (LLMs).
By leveraging NVIDIA's enterprise-grade AI stack, HCLTech Clinical Advisor achieves the responsiveness and reliability required in high-pressure healthcare environments.
The intelligent agent framework is powered by NVIDIA GPUs, which deliver results in real-time without compromising accuracy or safety.
By leveraging NVIDIA's powerful ecosystem of tools and technologies, we're helping clinicians work smarter—not harder—to deliver better patient outcomes.
The Impact
Saving time, cost and reputation.
The most profound impact of the GenAI-powered clinical advisor lies in enabling clinicians to spend more time with patients while handling an increased caseload, all without burning out.
- Give time back Save 4,000+ hours of clinicians collectively per day
- Cost savings: Reduced nearly $100 million in operational costs through more efficient and intuitive workflows
- Better patient experience with more face time, empathy, and personalized treatment recommendations
- Trusted advisor: Ensure clinicians are abreast of the latest in medical research and treatment protocols and benefit from valuable advise based on trusted data sources
