JMIR Publications

ABSTRACT

Background:

Obesity and overweight (O&O) is a serious health problem worldwide with multiple and connected causes. Simultaneously, Chatbots are becoming increasingly popular to interact with users in mHealth applications.

Objective:

This study reports the user-centered design and feasibility study of a Chatbot to collect linked data to support social network analysis of overweight and obesity causes in populations.

Methods:

This study was organized in 5 phases according to a user-centered approach: (I) Defining context, users and goals of the Chatbot; (II) User’s needs investigation. User´s preferences in ascetics collected based on opinions about fifty-two wireframes designed by one hundred and fifty students of the Design School. We gathered wireframes preferences through an open survey, it also included questions about sociodemographic, diet and activity habits, desired functionality for O&O applications and specific diseases; (III) Design and development; (IV) Usability evaluation (Task test and System Usability Scale (SUS) questionnaire); (V) Pilot study for feasibility and social network analysis.

Results:

Four hundred fifty-two people answered the survey. Participants were recruited by an email invitation to the university community (students, academy, and staff). Most participants would use a Chatbot for O&O assessment and believed that it would help to prevent it. They preferred O&O's assessment functionality, physical activity, and diets recommendations. Preferred colors for the Chatbot were green, blue and white. They scored wireframes based on colors, simplicity, and figures. Accordingly, we developed a Telegram's Chatbot with functionality structured in 6 sections: personal, diet, physical activity, social network, status and about. The user's status score was defined as a normalized sum (0 to 100) of diet score (frequency of eating 50 foods), physical activity scores, Body Mass Index, and social scores. We selected a character as Chatbot's name (Wakamola), also to personalize it and implement gamification. Sixty-one volunteers participated in the usability test. They were able to use the Chatbot without help, although only a few were familiar with Telegram. Usability improvement mainly required reducing the diet number of questions and steps to build the network. Eighty-five healthy volunteers participated in the pilot, 74 completed all sections: 54 females, 20 males, mean age 20,7 years, university students, mean BMI 21,4 kg/m2 (normal weight), 5 people underweight, 8 people overweight, no obesity cases, mean Wakastatus 78,21, 55 living areas; Social network: 74 nodes, 178 relations, 9 communities.

Conclusions:

The Telegram´s Chatbot Wakamola is a feasibility tool to collect data from a population about socio-demographics, diet patterns, physical activity, BMI, and specific diseases. Besides, the Chatbot allows connecting users in a social network to study O&O´s causes from an individual and social perspective. Clinical Trial: Universitat Politècnica de València, Ethical Code: P7_12_11_2018