Electronic Health Reporter

By Sachin Kalra, vice president of customer success, Infostretch.

Rapid advances in technology mean the chatbot market is now one of the fastest-growing segments in healthcare, with the market expected to be worth more than $314 million by 2023.

In some ways, this growth is not surprising.  Combined with the commercial benefits for healthcare providers, there is a genuine appetite for more advanced technologies to form part of patients’ healthcare.  A recent study in the US revealed that more than half of consumers would use an app for remote general consultation if given the option, while research in the UK found that apps would be used by 47 percent of patients to book appointments, and 42 percent to manage prescriptions.

Before long, it is likely to be very commonplace for prescriptions to be re-ordered through your smart speaker, for medical appointments to be made by Alexa, and for medical disclaimers and drug side effects viewed in augmented reality (AR) via Google Home. In fact, chatbot applications such as these already exist as proof of concept projects and even, in some cases, as deployed systems in the US. The providers that successfully deliver systems like these which make the lives of patients fundamentally easier will inevitably gain mind share and market share, as the good news of the improved service spreads.

Beyond these immediate applications, the potential of chatbot systems in healthcare is virtually endless, limited only by the imagination and needs of physicians and their patients.

The three types of chatbot

Whether they’re employed in healthcare, customer service or simply for general consumer use, there are three main types of chatbot.

• The first of these is the task-oriented chatbot, designed to deal with specific scenarios such as placing orders or scheduling appointments.
• The second type are information-oriented chatbots, which are more focused on the generative aspect of a conversation. Relying on AI and expert systems, they’ll offer answers as creatively as possible, avoiding repetition and attempting to keep the conversation interesting for the person they’re chatting with. In the context of telemedicine, for example, these chatbots could be used to explain side effects or to discuss concerns over drug interactions.
• The third type of bots are either open-domain, designed to retrieve information for questions such as what the weather will be like in a week’s time, or closed-domain. Also known as domain-specific, these chatbots operate with regard to a particular area of interest, aiming to give answers to narrow scenarios such as offering guidance through a museum by providing visitors with very specific types of information.

Most healthcare chatbot apps would typically fall into this latter, closed-domain category. Whatever the category of chatbot, so long as they provide users with an improved quality of experience, healthcare providers will be able to deliver a better service to more people at a lower cost.

Applying artificial intelligence

The performance of these chatbot apps – especially their ability to adapt as required – can be largely impacted by AI and machine learning technology, the application of which can enhance a number of areas.

By eliminating human bias from interactions, natural language processing can widen the topic of conversation, and increase the number of valid responses available to a chatbot. Of course, being able to answer a wider range of specific questions and provide more information will only make these apps more useful.

AI can also improve business performance for internal-facing bots which, in turn, will improve the customer experience for both practitioners and patients. Automating patient/admin interaction will enable more flexible scheduling options, for example, while the ability to more thoroughly convey information on side effects and conflicts from drug interactions will only improve patient outcomes.

What’s more, chatbots known as cognitive bots can use deep/machine learning to continually learn from their ongoing interactions, in order to provide more tailored responses to a patient’s needs. Accessing massive data sets and rapidly extracting insights from them is a task much better suited to AI versus humans who are limited by time.  Longer term, cognitive bots will deliver improved healthcare outcomes for more patients at a lower cost to the provider.

Adoption and appetite

The growth in the adoption of chatbot technology is likely to be organic. With each success that is achieved, a wider set of needs will be recognized and the technology developed further to address them. Its adopters will range from the largest healthcare innovators, where we would expect to see such innovation, to the smaller rural healthcare facilities who are set to benefit most from the resource and cost efficiencies it offers.

Within the last few years many of us have become accustomed to using Alexa, Siri and Google Home in our daily routines. As the healthcare industry continues to embrace chatbots, it won’t be long before we think nothing of asking them for medical advice, to carry out administrative tasks, or even to speak directly with our doctors. And given the rate at which this technology is evolving, who knows what the next few years might hold?