Investigating the Usability of a Tablet-Based Interface for Management of Anxiety-Related Arousal in Children with Autism Spectrum Disorder

Background:  Comorbid anxiety is a significant concern in autism spectrum disorder (ASD) due to its high prevalence, negative impact on physical and psychological well-being, and interaction with the core deficits of ASD. Current methods for assessment and treatment of anxiety rely on the observation of behaviour or self-reports, which are often ineffective in children with ASD due to known deficits in communication and diminished introspective ability. To reduce these barriers, we have developed a device, called the Anxiety Meter, that translates a user’s heart rate into a visual display of arousal level. The Anxiety Meter interface consists of a tri-colour gradient (red, yellow, and green) with a moving bar to mark the user’s level of arousal in real time. 

Objectives: Evaluate the usability of the Anxiety Meter interface in a sample of children with ASD. 

Methods: A sample of children with ASD (n=9; age: 10.7 +/- 1.8 years; full-scale IQ: 101.4 +/- 11.2; 6 male) completed device training and one anxiety-inducing task (public speaking), preceded and followed by a baseline task (movie-watching).  For the anxiety-inducing task, participants were given 2 minutes to prepare for a 3-minute speech on a topic of their choice. Participants sat in front of a tablet computer with the Anxiety Meter placed at eye level and were asked to indicate when they recognized they were anxious through pressing a button. Usability was evaluated according the to the ISO 9241 Standard using measures of interface effectiveness, efficiency, and user satisfaction. Interface effectiveness was defined as the user’s ability to recognize arousal states. Efficiency was measured as the time between indication of arousal on the device and the user’s recognition of the arousal level. Satisfaction was evaluated based on a user-rated likeability score from 1 to 5.

Results: Users were detected anxiety states indicated by the device with sensitivity and specificity of 0.7±0.3 and 0.9±0.2, respectively (effectiveness). Average time needed for recognition of anxiety states across participant was 70.6±48.5 seconds (efficiency). The likeability of the application was rated as 4.1±0.9 (satisfaction). Based on these results, the main usability issue identified was the need for an auditory or tactile indicator to user attention during arousal states. Participants also suggested personalization of gradient colors. 

Conclusions: Our results suggest that the Anxiety Meter interface is usable for children with ASD, though there is a need for an auditory or tactile signal to gain user’s attention during arousal states. Future studies should investigate the clinical effectiveness of the device in improving awareness of arousal signs in this population.