Amazon takes a big leap in healthcare: using patient data for diagnosis

Called Comprehend Medical, Amazon will use machine learning to extract data from EHRs

The promise of big data in healthcare has been that technologies, such as artificial intelligence and machine learning, would make it easier for a doctor to diagnose a patient and personalize care to ultimately improve outcomes. Not that the technology would supplant the doctor, necessarily, but that it would look at a patient's medical records or their test results and pull data out faster so that a doctor could make a quicker and more accurate diagnosis. 

That sounds all well and good, but actually doing it is another story. There are plenty of companies looking to make this into a reality, but now one of the biggest tech companies is getting in on the action, and that might be a game changer for the space. 

Earlier this week, Amazon, said in a blog post that it's launched a service called Comprehend Medical. The announcement comes at the same time as the tech giant's AWS conference.

The new service will allow healthcare companies to take unstructured data, such as medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports, and quickly turn it into something actionable. 

"The majority of health and patient data is stored today as unstructured medical text, such as medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports. Identifying this information today is a manual and time consuming process, which either requires data entry by high skilled medical experts, or teams of developers writing custom code and rules to try and extract the information automatically," wrote Dr. Matt Wood, GM of Amazon's growing Artificial Intelligence and Machine Learning unit and Taha Kass-Hout, former U.S. FDA chief health informatics officer whom Amazon hired in the spring of 2018.

What Comprehend Medical can do is use the data to identify things like medical conditions, anatomic terms, medications, details of medical tests, treatments and procedures.

"Ultimately, this richness of information may be able to one day help consumers with managing their own health, including medication management, proactively scheduling care visits, or empowering them to make informed decisions about their health and eligibility."

The data will be available to entities such as health care providers, insurers, researchers, clinical trial investigators, health care IT, biotech, and pharmaceutical companies.

Amazon already revealed two institutes that are using Comprehend Medical, including Fred Hutchinson Cancer Research Center, through which Amazon says it was able to reduce the time it took to evaluate, extract and index clinical notes from hours to just seconds.  

For those concerned about privacy, Amazon says that Comprehend Medical is Health Insurance Portability and Accountability Act (HIPAA) compliant. It also maintains that it will highlight protected health information (PHI), and that no data will be stored on its servers.

Still, with all of the data scandals that have occurred in tech over the last year, it is likely that the company will eventually need to go into more detail in regards to how it is keeping patient data private to make people feel fully comfortable with there health data potentially being shared. 

Like many other big tech companies, most notably Google and Apple, Amazon has been making a bigger push into healthcare recently. 

The company, which first started making moves into the pharmaceutical space last year, announced in June that it acquired online pharmacy PillPack. Amazon also teamed up with  Berkshire Hathaway and JP Morgan Chase to create a new venture centered around improving employee healthcare.

Amazon's stock price ended at $1,673.57 a share on Thursday, up 74 percent from when it first announced its intention to get into healthcare last year, up 19 percent from when it announced the Berkshire/JP Morgan deal at the end of January and flat from where it was when it bought PillPack in June.

(Image source: geekwire.com)