Main image courtesy of AI Magazine.

One of the major challenges of today’s healthcare field is striking the right balance between taking care of patients, and charting the accurate input of data into their electronic health records (EHRs). Physicians find that oftentimes they’re spending so many hours ensuring the information is accurate, they are no longer able to be completely present with the patient during a consultation. Clinicians are getting burned out worrying about taking medical notes as the patient speaks and then getting them into the file after their appointments are done for the day.

It’s essential to have accurate charting for patients, as this ensures there are no gaps in their history and it reflects their health situation at any given time. What if there was a better solution that would free up doctors’ time and energy so they could spend more time with patients? Artificial intelligence, (AI) might just be the answer the healthcare field has been looking for. 

Today’s discussion is going to cover:

• Whether AI can accurately write medical notes
• The advantages and disadvantages of using AI for medical notes
• How to find the best AI scribe for medical notes

Can AI write medical notes?

And is it accurate enough for physicians to count on?

The short answer is yes, AI can assist with transcribing medical notes. There are a lot of healthcare facilities that have already started to consistently use this type of technology, and have certainly noticed the benefits for both patients and doctors. One of the larger studies to focus on the effectiveness of this technological tool was done in northern California by Kaiser Permanente. They rolled out AI medical scribes at 21 locations, which were used by over 3,000 doctors over a 10 week period. They found that not only was the AI easy to use, it produced accurate medical notes and saved physicians hours at the keyboard. That is one of the major pluses to employing AI to do the repetitive task of inputting data into the EHR is that the clinician no longer has to sit at the computer and do that after seeing patients. The data is transcribed in real time (or very close to it) as the doctor/patient conversation progresses.

At the end of the appointment, the doctor will still need to go over the transcription, as the AI is  not always completely accurate and there are a small percentage of errors. But the amount of time that doctors now have far outweighs needing to glance over information before it is sent to the patient’s EHR. 

How does an AI system transcribe notes?

What may also surprise some folks is how easy it is to take advantage of AI for medical notes. Most applications can be downloaded to a smartphone, and some even have desktop versions as well. The clinician asks for patient consent, and then turns on the application so that it can start listening. It’s easy to turn off and on as well. But how does the AI know how to transcribe notes and collect important data? The system relies on natural language processing (NLP) and speech-recognition process algorithms to recognize speech patterns, linguistic rules, and complex meanings.

Using machine learning models, the application can sift through enormous quantities of data and learn to recognize words and ideas. In addition to creating accurate text of the conversation for medical notes, AI can also extract important data that allows it recognize intent and sentiments. The data is also easy to organize and can be designed to fit into whatever format template the facility uses for EHRs, so that any additional notes or patient data will automatically go to the proper place. 

What are the benefits of using AI for patient documentation?

Why you should be interested in using AI for medical notes

• Frees up clinician time. One of the most important benefits of deciding to use an AI medical scribe is that it gives physicians more time in their day. Oftentimes doctors not only have to listen to their patients as they relate their concerns in real time, but they also must keep part of their brain focused on documenting what’s going on. This can obviously be distracting, and doctors find that they can’t be entirely focused on the patient while in an appointment. 

When using AI to transcribe the conversation, doctors are no longer worried about missing something or mis-remembering what the patient said hours later when they’re inputting medical notes. This extra time allows them to focus more on the patients that they do see, or provides more time for breaks to help prevent burnout, which is very real in the healthcare profession. 

• Fosters better patient/doctor relationships. It’s difficult for doctors to split their time between hearing a patient and taking notes on a computer, but it’s also frustrating for patients who may feel they do not have the full attention of the physician. Having the AI there to pick up the conversation and transcribe it allows both parties to be fully present and speak about any concerns or questions the patient may have. This leads to better patient satisfaction knowing that they have their issues addressed during their assessment or examination.

• Medical coding compliance. Although it may not be the first benefit that comes to mind, AI can also assist with ensuring medical coding is done correctly. Medical codes are essential to any healthcare facility, as they contain the diagnosis, medical services, and treatment, which are used to bill patients. The codes come from the information in the medical notes, and many times are inaccurate or wrong. There are a high percentage of medical bills that are coded wrong, as the documentation is either incomplete or difficult to read. 

AI can eliminate a lot of these billing errors since it transcribes the diagnosis and treatment into an electronic format (avoiding handwritten notes) and can learn coding patterns so errors are more likely to be flagged before the bill is sent to the patient. 

• Easy to use. AI medical scribes are also easy to use and clinicians do not have to undergo extensive training to use them effectively. Most just require a few webinars and handouts, and the application is available right on their phone. They’re designed to be user-friendly so physicians will want to continue to use them. 

• Communicates with EHRs. It’s essential that AI for medical notes interact seamlessly with EHR systems so that information can be uploaded and stored securely. Luckily most of the AI medical scribes can easily speak to EHRs via an API or other similar method. AI medical scribes do not store patient data, which means that you’ll want to make sure whatever software you choose as a medical scribe will be able to communicate with your EHR system in order to easily organize and securely store data. 

• Uses pattern detection to create insights. As we know, accurate medical notes are essential for improving care for patients. AI can detect patterns in a patient’s medical history or diagnosis, making it easier for it to fill in any gaps or flag where there is missing information. Better records lead to better care and treatments, which can help the patient improve or recover faster.

• Seamlessly add or change patient records and find data. Physicians won’t have to worry about out of date information again when they can easily confirm patient data in a conversation and have it updated accurately. Additional notes can be added to the patient file, and if the clinician is looking for specific information, AI medical scribes can access it quickly so the appointment can continue without interruption. 

Are there any downsides to using AI for medical notes?

While all of these benefits sound great, AI technology for medical notes is still rather new, and there are bound to be some issues with it. 

Needs to be reviewed by doctors

AI technology can only do so much, but it still needs to be supervised by a clinician to ensure that everything makes sense and that the conversation was accurately transcribed. Although the applications come prepared to be used immediately, the machine learning process is ongoing, and the more it is used and is exposed to language and patterns, the better it will get at transcribing and analyzing. In some cases, the AI may transcribe something that was misheard or it did not recognize some of the words used, which can lead to inaccurate information in the report. Even though a doctor will need to review the medical notes, it takes significantly less time than entering them into the computer themselves.

There can be inaccuracies

As we mentioned, because the technology is so new and it is still learning, there are bound to be some inaccuracies in the clinical notes. Whether it’s a misspelling of a name, or a sentence with no context, any physician that uses the application should plan on reviewing each transcription for accuracy.

Threats of a cyber attack

Healthcare facilities aren’t the only ones using AI, as many cyber criminals are also turning to offensive AI to infiltrate hospitals and health centers. Although AI medical scribes do not retain the sensitive patient data, they still need to communicate with EHR systems, which if not properly secured may fall threat to a data breach. 

What is the best AI scribe for physicians?

How do you know which AI system is best for you?

There are a variety of AI medical scribe applications available for healthcare facilities, and the one you choose depends on what your specific needs are. When you’re comparing applications you should keep what’s important to your facility in mind, and list out exactly what you’d like to see in the software and what your expectations are. 

What to look for in AI systems for medical notes:

• Compatible with your EHR system
• Is HIPAA compliant
• Easy to use
• Secure about patient data
• Ready to use immediately

Some of the popular AI systems for medical notes include:

• Dax Copilot/Nuance
• Tali AI
• Epic 
• Suki
• Freed

Most systems offer a free trial, or so many free notes per month, so you can really take some time to figure out which one you’d like to invest in. 

You can also check out our article ranking the best AI medical scribes!

Will you start to use AI for medical notes at your facility?

Employing AI to transcribe medical notes during a patient visit is a great way to ensure doctors can stay focused on evaluating the patient, listening to concerns, and answering any important questions. Not having to be stuck at a computer entering notes ensures clinicians are engaging patients more, and will no longer need to worry about inputting massive amounts of notes into the patient’s file at the end of the day.