Abstract:

Background:
The Internet has become a popular source of health-related information, but its content is not always reliable.

Objective:
To connect the collaborative learning of physiotherapy students with the doubts of inhaled therapy users through the training of a conversational artificial intelligence agent (chatbot).

Methods:
A focus group was conducted with people with respiratory problems who were users of inhaled therapy to collect their doubts related to aerosol therapy. Participants also completed the 10-item Test of Adherence to Inhalers (TAI), as well as a survey to collect their clinical and sociodemographic data. The questions raised in the focus group were analyzed qualitatively, resulting in a list of doubts that the students had to answer as a group to feed to a chatbot. The present study was carried out during the 2021-22 academic year as part of the 2nd year course "Specific Methods of Intervention in Physiotherapy: Respiratory System". Once the activity was completed, a Likert-type survey was given to the students for evaluation. The study was carried out in accordance with the required ethical standards and with the approval of the Research Ethics Committee.

Results:
The focus group included 2 men and 4 women (47.2±10.2 years); 1 of them suffered from bronchiectasis and 5 from asthma. On average they had been using inhaled therapy for 17.4±4.1 years. Their TAI scores reflected 16.7%/33.3% and 50% of good/intermediate and poor adherence, respectively. After focus group analysis, 50 questions were obtained; 40% were related to the management of inhaled therapy devices and 60% to more general aerosol therapy issues. Of the 174 students enrolled in the target subject, 11 participated and 8 of them completed the survey. The students rated their feeling of preparation for the activity as 9.2/10 and their satisfaction with it as 9.4/10. In addition, they highlighted the usefulness of their work and the interest generated by the knowledge of new therapeutic tools. The teachers reviewed the assignment and provided improvements to 100% of the answers given by the students; the main problems detected during supervision were the overly technical use of language and some misconceptions.

Conclusions:
Connecting students' learning with real social needs through an innovative tool seems to be motivating for them. However, despite their feeling of good preparation, the answers of the studied student profile could not be used to feed the chatbot directly.

Keywords:

Inhaled therapy, chatbot, learning.