Amazon launches Machine Learning-based service to analyse unstructured medical text

analyse unstructured medical text

Machine Learning-based service to analyse unstructured medical text

Amazon has launched Amazon Comprehend Medical, a machine-learning service to process the unstructured medical data of the patients.

Amazon has recently announced Amazon Comprehend Medical, which is a new HIPAA-eligible machine learning service that helps the developers in processing unorganized and unstructured medical text in clinical trial reports and doctors’ notes and recognize the information, for instance, patients’ diagnosis, prescriptions, medical treatments, medicinal doses, disease symptoms, and more.

Today identifying the patients’ health-related data involves physical work, and it is a time-consuming process. It either requires the highly proficient medical experts to do all the data entry work or a team of developers to write custom codes and rules to take out the information automatically. In both the cases, this undistinguishable heavy work takes away the time and efforts that otherwise could be used to improve patient outcomes through technology.

Comprehend Medical has been designed to help health care providers, researchers, insurers, and clinical experimental investigators along with healthcare IT, biotech, and pharmaceutical companies.

It will help in improving medical decision support, updated revenue cycle and maintain data privacy and protect health information requirements.

Amazon is mainly aiming to eventually use the tool to help patients to manage their own health issues, schedule their visits, and make some well-informed decisions in regards to their own healthcare.

Comprehend medical and cancer research

According to their blog post, Amazon has confirmed that it is already in a close working relationship with Seattle’s Fred Hutchinson Cancer Research Center to use the service for cancer care patients. Comprehend Medical reads the information from patients’ EHR and is helping the providers in identifying the patients for clinical trials who might be able to benefit from particular cancer treatments.

Thus far, the cancer treatment center has assessed the millions of clinical notes to collect and file the medical conditions, medications, and the required cancer therapeutic options to considerably lessen the time needed for processing the health record of each patient. After all, time is of the essence for the cancer patients.

According to the CIO of Fred Hutchinson Cancer Research, Matthew Trunnell, for both cancer patients and the researches, time is a big issue, and it is available in a limited capacity. He further added that Amazon Comprehend Medical will bring down the time burden from hours per record to seconds.

Amazon has also been getting its service previewed by Roche Diagnostics. Roche Diagnostics is a diagnostic division of Hoffmann-La Roche which makes equipment and components which are used for research and medical diagnostics.

As per Roche Diagnostics Information Solutions Director of Software Engineering Anish Kejariwal, Roche’s NAVIFY decision support portfolio offers the solutions that can speed up the research and allow personalized healthcare. With a large amount of unstructured data being produced in hospital systems every day, Roche is aiming to take the data and change it into valuable insights that are understandable and easily accessible. That is where Amazon Comprehend Medical is proving useful as it helps in quickly digging out and structuring the information from medical documents.

During the month of August, Amazon joined Microsoft, Google, IBM, Oracle, and Salesforce to show their commitment towards moving forward with healthcare interoperability and health data exchange using the cloud and Artificial Intelligence (AI).

Image credit:


Leave a reply

Your email address will not be published. Required fields are marked *


© 2018 Dr. Hempel Digital Health Network

Dr. Hempel Digital Health Network is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to


Log in with your credentials

Forgot your details?