Chatbots: A New Kind of Clinical Assistant?

November 21, 2017 Daniel Ocampo and Mauricio Giacomello - Mobile Engineers

Soon clinicians will be saying hello to their new personal assistant: the chatbot.

The latest extension of the "chat" feature in patient engagement software, the chatbot—designed to imitate human conversation using artificial intelligence—is a service that humans can interact with through a messaging interface. Essentially, it's a digital personal assistant that could give clinicians time to see more patients, more efficiently.

POST CONTINUES BELOW

Find out more about machine learning in healthcare by downloading our report. Click the link now!

Currently, the chat feature of many applications or software systems allows patients to communicate with their doctor through online messaging. The chatbot builds on this concept but is much more advanced, as it's able to converse with a patient, fundamentally changing the traditional doctor-patient relationship. The chatbot may save patients with minor health concerns from a visit to the doctor, as the chatbot is able to answer simple medical questions. This would allow clinicians more time to treat patients who need a consultation the most, and save those whose symptoms don’t necessitate an actual consultation from having to visit a doctor at all.  

The opportunity for future development of chatbots is plentiful—it's just a matter of companies defining how the bot can be most useful to clinicians. For primary care physicians who spend a lot of their time on administrative tasks, the chatbot could reduce these demands and allow the physicians to spend more time consulting patients. For instance, the bot could assist the physicians in appointments by listening and taking notes so the physician can conduct a more efficient and effective consultation. 

A further possibility is having the bot as a third “person” in the conversation, one that can pick up on something a clinician mentions, then search the database for pertinent information in real time. For example, while the clinician recommends a particular food to the patient to help with a health issue, the bot could search the database and identify that this patient is allergic to something in that food that the patient had forgotten about. This would be highly useful, as the clinician can then adjust their care advice to make it more accurate and patient specific. Previous information can also be added to the conversation, as the bot can quickly present things such as the patient’s last weight check or blood test result.

In order to make people feel comfortable engaging with a chatbot, it uses a combination of Natural Language Processing (NLP)—both to and from the bot—and machine learning. NLP processes and interprets data coming into a bot and converts the machine’s response into human dialogue.

The beauty of the chatbot is that, through this process, it will get smarter, as it uses machine learning to be able to respond to more complicated questions and recognize less common words or phrases. Consequently, chatbots in a healthcare environment would need to be used by many and have access to rich data sets in order for them to increase their knowledge of medical terms, symptoms, and treatments.

The purpose of the chatbot is not to replace the clinician—it's to serve as a resource to enhance patient engagement. In the previous scenario, the clinician would be there to correct the chatbot if it presents the wrong information, or if it isn’t relevant. Therefore, the clinician can teach the bot as it goes.

Not limited to text alone, when replying to messages the bot can provide options for the user to choose from, such as a selection of images in a carousel, videos, and links to other information.

Using machine learning to provide specific options can accelerate the conversation between clinician and patient by rapidly feeding the bot more information. With further development, the chatbot can transcend the traditional text-to-text format of chat applications, providing highly valuable assistance to clinicians.

***

Find out more about machine learning in healthcare by downloading our report. Click the link now!

Previous Article
Aggregated data or Federated data: Is one better than the other?
Aggregated data or Federated data: Is one better than the other?

As big data platforms become more popular, two models of data management—aggregated and federated—are emerg...

Next Article
Opening up a world of opportunity for healthcare through APIs
Opening up a world of opportunity for healthcare through APIs

Healthcare has experienced a significant shift toward value-based care and patient empowerment, creating an...

×

Subscribe to our Blog!

First Name
Last Name
Thank you!
Error - something went wrong!