Client
A non-profit organization dedicated to conducting cutting-edge research in neuroscience, focused on developing innovative treatments for neurological disorders.
Business Challenge
The client faced a critical challenge in diagnosing and treating neurological emergencies in a time-sensitive manner. Patients often arrived at hospitals without an immediate diagnosis, and the client needed a solution that could expedite diagnosis and initiate timely treatment.
Solution
Nous proposed a Virtual Neurologist application to assist emergency responders in diagnosing neurological emergencies, particularly strokes, in real time.
Key Features:
- Deployed a real-time stroke triage assistant in ambulances using AWS Lambda and Amazon Bedrock (Claude) to assess symptoms and recommend rtPA treatment.
- Used Amazon Transcribe, Amazon Textract, and Amazon Comprehend Medical to convert emergency calls and documents into structured data and extract key clinical findings.
- Ingested and processed audio, images, and inputs with Amazon SQS, Amazon Kinesis Data Streams, and AWS Glue, indexing patient records in OpenSearch for instant retrieval.
- Leveraged Amazon Bedrock’s Titan embeddings to encode medical data, enabling rapid, context-aware searches against past cases stored in OpenSearch.
- Coordinated triage logic and chatbot flows via AWS Step Functions, with Amazon DynamoDB persisting conversation histories.
- Hosted the frontend on Amazon S3 and accelerated delivery with Amazon CloudFront, while AWS WAF safeguarded the pilot environment against web threats.
Technologies
Business Benefits
- Achieved faster diagnosis and treatment, significantly reducing critical response times.
- Scaled patient care capacity to treat up to three times more cases.
- Enhanced diagnostic accuracy, lowering misdiagnosis rates and accelerating intervention.
- Reduced overall hospital costs through quicker interventions, fewer complications, and shorter stays.
Outcome
3x increase in patient care capacity by empowering emergency teams to diagnose and initiate stroke treatment faster, reducing misdiagnoses and hospital stay durations.
