Hygieia - A Gen AI Solution for Pharma

2 minute read
Content level: Intermediate

The problem statement concerns the misinterpretation of handwritten doctor prescriptions at a high level. While this is a well-known issue in healthcare, it extends beyond and is sometimes observed even in the insurance industry. Handwritten medical reports submitted by healthcare providers, claims submitted by policyholders (and in some cases, agents), and notes written by underwriters can all pose risks, lead to coverage disputes, and, in some instances, facilitate fraud.

Misinterpreting handwritten doctor prescriptions is a problem. For example, in the United States, it has been reported that sloppy handwriting by doctors has led to adverse events, with one report from the National Academies of Science Institute indicating that 7000 deaths occur in the US due to the sloppy handwriting of doctors. (https://www.nature.com/articles/s41598-022-07571-z).

Good news is, Gen AI has the potential to significantly improve it. After seeing models such as Claude 3 Sonnet, we have a solution to the problem. Here is what we can potentially do to solve. This would not only improve patient safety but also free up pharmacists' time to focus on providing better care to patients.

Here is the springboot sample application to interact with Anthropic Claude 3 foundation model using Amazon Bedrock API. You can upload an image and interact with Anthropic Claude 3, ask questions.

Enter image description here

Enter image description here


https://github.com/awsdocs/aws-doc-sdk-examples/blob/main/javav2/example_code/bedrock-runtime/src/main/java/com/example/bedrockruntime/InvokeModelWithResponseStream.java https://github.com/awsdocs/aws-doc-sdk-examples/blob/main/javav2/example_code/bedrock-runtime/src/test/java/ImageToTextAsyncTest.java https://docs.aws.amazon.com/sdk-for-java/latest/developer-guide/java_bedrock_code_examples.html

profile pictureAWS
published 20 days ago1055 views