AI In 2023 Healthcare: What Does Amazon’s Healthscribe Do?

Just about every major tech company is interested in implementing their artificial intelligence tools into the healthcare industry. Amazon is just the latest to do so, rolling out a new service called Healthscribe that aims to help healthcare professionals lighten their workload.

There are a number of things that Healthscribe can do for the healthcare industry, which we will go into below.

What Does Healthscribe Do?

Healthscribe performs a specific role, which is always good for A.I., because that means that the A.I. is likely very strong at a specialized task that it was specifically trained for.

That specific task is creating clinical notes for doctors based on conversations between healthcare professionals and patients.

The Process of Healthscribe

What Healthscribe does is take a recording of a doctor-patient conversation, transcribe that conversation, then create a summarizing list of clinical notes.

This holds significance because doctors usually generate these clinical notes themselves, often outside of their working hours, which might result in burnout.

So, implementing Healthscribe will save doctors’ time and energy.

Specific Tasks Performed by Healthscribe

This A.I. platform is adept at identifying who is the patient and doctor in the conversation, so that the transcripts usually come out with accurate assignments of who said what.

There is even a “skip small talk” option that allows you to simply keep the most essential information in the transcript.

The transcript will not be a single large chunk of text; instead, it will have segmented portions with distinct summaries for each segment. These segments will include information about what the assessment and plans for care are. Being able to quickly find the history of a patient’s illness and the doctor’s plan makes navigating the transcript much easier.

The clinical notes derived from the transcript are also noteworthy for “citing their sources,” meaning the notes will guide you to the section of the transcript that a specific note summarizes.

How Could Healthscribe Improve Healthcare?

The biggest positive to this technology is an increased efficiency in documentation. Instead of becoming entangled in the task of transforming a meeting into clinical notes, doctors can automate this process and allocate more time to interacting with patients.

Healthscribe vs. ChatGPT: Which Is Better?

If you had to go with one, choose Healthscribe, because ChatGPT was not specifically trained to create clinical notes.

It certainly can make those notes for you, but the problem is that it is going to be much more prone to error.

Healthscribe is stronger for this very reason. Its specialization to the task will ensure that you will get better results.

Also, the above-named features speaks to the comprehensive level of services that you get with Healthscribe. As of this writing, you cannot give ChatGPT a recording of a conversation and ask it to transcribe it. The optimal approach with ChatGPT would involve providing it with pre-written notes and then requesting a summary of those notes.

For this reason, a healthcare professional should certainly go with Healthscribe in this situation.

Are There Any Downsides to Using Healthscribe?

Possibly so. For one, data security is probably the biggest issue here.

Because there is no such thing as 100% safe data security solutions, healthcare professionals should certainly consider the possible issue of a data breach exposing all of this confidential patient-doctor information that is given to the A.I. platform.

It is for this reason that many major corporations that have valuable trade secrets are very strict about the use of services like ChatGPT in the workplace. Though Amazon will not use your Healthscribe data to train its A.I. systems on, it is still worth considering the risk of a data breach. On the other hand, Amazon will certainly have very strong data security measures in place, so you can be confident that the data is safe.

Precautions for Using Healthscribe

The solution to the data safety issue would be to find ways to “anonymize” the recordings, so that no one could tell who the patient is. From shifting the pitch of the recording up or down to disguising the voices to not using the patients’ name or other identifying information in the transcription and notes, there are workarounds here.

Though the workarounds could be challenging, losing that confidential data in a breach could certainly be costlier.

Another thing that doctors should understand is that although these notes are convenient, it is good to keep in mind that the process of creating clinical notes is often conducive to reflecting on a patient’s case, and figuring out how to improve it. For this reason, doctors should look at the convenience of A.I. generating the clinical notes as an opportunity to spend more time reflecting on the case itself.