Amazon Comprehend
Developer Guide

How It Works

Amazon Comprehend Medical uses a pre-trained model to examine and analyze a document or set of documents to gather insights about it. This model is continuously trained on a large body of text so that there is no need for you to provide training data. Amazon Comprehend Medical can examine and analyze documents in English.

With Amazon Comprehend Medical you can examine unstructured clinical text to detect:

  • Detect Entities —Examine unstructured clinical text to detect textual references to valuable medical information such as medical condition, treatment, tests and test results, medication, and PHI data.

    References to valuable medical information such as medical condition, treatment, tests and test results, medication (including dosage, frequency, method of administration, and so on), treatment, PHI data, and so on.

  • Textual references to protected health information (PHI) such as names and addresses.

You can use the DetectEntitiesV2 and DetectPHI operations to examine individual documents to detect valuable information in unstructured clinical text. These are synchronous operations. You send a document to the Amazon Comprehend Medical service and receive the results of the analysis in the response.

You can also use the StartEntitiesDetectionV2Job and StartPHIDetectionJob operations to start asynchronous jobs to analyze clinical text that is stored in an S3 bucket. The output of the detection jobs are written to an S3 bucket so that you can retrieve it when the job is complete.