- Timeline: 9th Semester, 2020
- Type/Setting: School Project
- Team: Stefan Bjerregaard, Simon Bak Kjærulff, Tummas Jóhan Sigvardsen
- Supervisor: Eleftherios Papachristos, Niels van Berkel
- Tools: Natural Language Interaction, Google Cloud Console
Abstract
Tobacco use is one of the leading causes of various diseases that can cause lethal damage to an individual. Although nicotine dependency is preventable, more than one in five adults is dependent on nicotine, making nations allocate many resources on healthcare and tobacco prevention measures. The primary techniques utilized for tobacco prevention by health care professionals are usually a combination of group therapy and nicotine replacement therapy. Mindfulness-Based Relapse Prevention (MBRP) is a branch of Cognitive Behavioral Therapy, which has shown great promise in coping with potentially triggering thoughts, feelings, and situations that arise from nicotine abstinence. With the ongoing global pandemic (Covid-19), the central group therapy sessions may prove difficult, if not impossible, to practice. We present HaRiS, a novel approach to MBRP through a personal Voice Assistant. We will perform a pilot study (N=8) utilizing our prototype, which employs natural language interaction for guiding the user through specialized smoking cessation exercises and guiding the users for self-reflection on their smoking habits.
Pilot Study
After building the prototype, we conducted a pilot study to identify potential improvements of HaRiS before deploying it for a comprehensive study later. After the pilot study, we interviewed the participants to gain a deeper understanding of their usage.
The interviews were conducted following a semi-structured approach. The prototype was deployed for us to understand how the user engages with HaRiS. Specifically, how they engage in natural conversation with the VA regarding the inputs and outputs, how they navigate the system without a supporting visual interface and how they engage with the content of the action.
Below we have listed the hypotheses, that we base our study around:
- H1: “Users will find it difficult when speaking to the system, as natural conversation can be hard to replicate.”
- H2: “Users will not find it necessary to use the system with a screen.”
- H3: “Users will engage naturally with the system.”
