Florence.chat : How to improve the adoption of Chatbots in Healthcare

Chatbot adoption in healthcare

Improved chatbot adoption in healthcare can be achieved by adding a set of key features that promote engagement and usability.

In general, long-term adoption of consumer healthcare technologies has been limited. On the one hand, a large portion of the population believes that they are healthy and that they don’t need any help or guidance regarding their wellbeing. This behavior is especially prevalent among the younger population.

On the other hand, consumer healthcare related technologies tend to be adopted mostly by healthier, wealthier and technology enthusiast subsegments of the population — i.e., segments that tend to benefit little from using them. In this way, once the newness and coolness factors have worn off, the ability of healthcare devices and apps to capture users’ attention and motivate them rapidly decays over time, leading to high dropout rates (e.g., 42% of users stop using their fitness trackers within six months).

To address this high churn, consumer healthcare developers and vendors have been experimenting with different sets of new features in their products. Furthermore, the more segment-specific these features are, the more effective they tend to be in becoming part of the users’ routine.

For example:

- younger generations tend to value more intense interactions that involve

1) a technological component

2) a high level of social connectivity

3) gamification

4) a competitive component

5) unpredictable rewards

- older generations tend to value interactions that:

1) are about improving quality of life

2) are personal

3) and that involve their family, friends or caregivers

Healthcare chatbot technology has been following a similar pattern of adoption to other consumer healthcare technologies. Being on an early stage of the adoption curve, its early adopters tend to be young, curious and impatient, and the technology itself still requires significant improvement — current chatbot-human interactions are not the same as human-human (and they may never be), often leading to user frustration.

In this way, to keep healthcare chatbot early adopters engaged, to guide them through their experience, and to minimize their frustration, segment-specific features, and aids (as listed above) are essential.

That’s why the menu and quick replies of Florence are different for every user, and we ask a few setup questions at the beginning (see image below). For example, it doesn’t make sense to show a healthy person the medication reminder. After that Florence learns more about the user based on his input. We are currently working on more segment-specific content.