In an era in which access to mobile devices is more widespread than access to nearby healthcare, medical chatbots have become convenient and useful tools that users can employ to receive advice, education, and personalised recommendations. However, because of the complex and precise nature of the language in the health sciences, medical chatbots are essentially useless if they cannot communicate in a user’s native language.
Medical chatbots have been around much longer than one might expect. Around 50 years ago, the first operative chatbot ELIZA was created. The bot was designed to mimic a Rogerian psychologist, asking patients questions by rearranging the question asked.
Today, chatbots are a bit more advanced, helping patients receive a diagnosis by analysing common symptoms and deriving information from a large database of medical data. While medical chatbots are by no means a replacement for human doctors, they serve as self-diagnosis tools, a sort of slightly more advanced “Dr Google”.
Advantages of health bots
In recent years, a wide variety of medical chatbots have been created, from standard medical advice bot Your.MD to cancer support chatbot CSource. The creation of this type of AI involves a collaboration between doctors and software in order to ensure the responses are accurate and credible but also to make the service more accessible and “empathetic”. In fact, the AI is often trained to reassure the patient and take some information about his or her age and health history to better provide personalised health advice.
This proclivity towards personalised health advice is one of the main appeals of health chatbots.
Unlike a medical information website such as WebMD, a chatbot can help narrow down possible diagnoses based on a patient’s specific symptoms and medical history.
Healthcare chatbots can also prove extremely useful in countries where specific health concerns or medications are taboo. This is particularly the case in regions where people receive little to no sex education and information about sexually transmitted diseases is not readily available, and are not able to ask family members or friends about questions they may have.
Medical chatbots also provide general information to help users live a healthy life and to pay attention to the symptoms that their body is exhibiting. Most medical chatbots also help connect users with local health services and schedule appointments.
The reality of developing multilingual chatbots
Despite the fact that the development of medical chatbots is increasing, many of the largest healthcare bots only operate in one or two languages. This language is most often English.
Creating multilingual chatbots can be costly, work-intensive, complex and error-prone, and not easily scalable. It’s somewhat understandable that makers of medical chatbots choose to produce their products in only one language. Each alteration in a bot needs to be replicated in every single language the bot supports. Many companies choose to automate their multilingual options with a Google Translate-like algorithm, but this can be problematic.
It takes native, in-country language experts to train a bot to recognise natural language and region-specific dialect and to respond appropriately.
It is impossible to deny that the process of teaching the AI is costly and time-consuming at first. Quality polyglot bots, therefore, are few and far between and are even rarer in the healthcare sector. However, there are many benefits to putting in the time and effort to create capable multilingual chatbots.
Firstly, providing a medical chatbot with multilingual capabilities means that more people will be using the product. And expanding to more markets means that the AI will have more data to teach itself, and will learn through interaction with users.
Secondly, providing another language option can be an opportunity to redesign the user experience in specific markets to appeal to different regions in different ways. This focus on localisation will allow medical chatbot designers to create bots that will be more engaging to a certain culture or region, and therefore more effective.
According to Vic Yankoff, developer for the health chatbot GYANT, ‘the benefits of multilingual bots are enormous. They are hard to get right, but if you succeed, the number of new users and markets you tap into are astronomical.’
Restricted access to healthcare
Users are unlikely to use chatbots, especially health bots, that can’t communicate with them in their native tongue, so it essential that chatbot developers expand their services to include multilingual capabilities.
Many bot services that support the life sciences, such as IBM Watson Conversation Service and rasa NLU, include a maximum of two languages. However, there have been several healthcare organisations that have successfully expanded their multilingual chatbot services. One example of this is iCliniq.
iCliniq was created in 2012 by Dhruv Suyamprakasam in India as an online medical consultation startup designed to connect users to doctors across the country in order to receive medical advice. However, due to the high volume of traffic and medical inquiries on the site, Suyamprakasam decided that it would make sense to add a chatbot service in order to handle some of the more simple questions.
Rural areas in India, access to healthcare is restricted and many medical clinics are overworked and overcrowded. Suyamprakasam rationalised that many individuals living in the less urbanised parts of the country would have a particular need to consult with specialists, such as dermatologists and gynecologists.
Regions in which there is limited access to healthcare are favourable markets in which to expand medical chatbot services.
The iCliniq site allows users to seek medical advice with a chatbot or to pursue a second opinion from a medical professional who speaks the same language as they do. Today, iCliniq supports chatbot capabilities and medical consultation in 12 languages, including Hindi, Tamil, Telugu, and Bengali.
With several smaller healthcare companies expanding their chatbots into multilingual and multinational sectors and reaching larger audiences, it’s only a matter of time before some of the larger pharmaceutical and healthcare organisations realise the advantages of doing the same and chatbots become ubiquitous tools for patients, caregivers and doctors alike.