Purchase this article with an account.
Allison M McKendrick, Astrid Zeman, Ilham Aden, Dilek Aktepe, Daisy Bhagat, Kieren Do, Huy D Nguyen, Andrew Turpin; Robot assistants for perimetry: Patient experience and performance. Invest. Ophthalmol. Vis. Sci. 2018;59(9):6027.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Standard visual field testing can be boring for both the patient and operator. Having an operator actively invigilate a test is ideal, but competing demands often result in minimal supervision. This study aimed to evaluate performance and subjective experience of perimetrically naïve subjects who received feedback during testing from: a human; a humanoid robot; a disembodied voice (computer speaker); and no feedback; . We tested the hypothesis that the presence of a humanoid robot conferred advantages over voice commands alone.
Twenty-two perimetrically naïve adults (aged 21-31 years) participated in a four visual field tests conducted using an Octopus 900 (Haag-Streit AG), controlled with the Open Perimetry Interface to enable automated feedback. All received initial introduction to perimetry from a human operator, and then participated in four tests with feedback conditions: 1) human; 2) humanoid robot (Figure 1: NAO Robot, Softbank Robotics, Japan); 3) computer speaker; 4) no feedback. The robot and computer speaker output were generated using the same “voice”, with minor differences in pitch but identical voice shaping. Feedback rules for the speaker and robot were identical. After each test, a survey was completed regarding test engagement using Likert scaling. At the end, an exit survey comparing all four conditions was completed. Perimetric performance was evaluated by comparison of mean sensivitivity (dB), fixation losses and false positive responses between feedback modes (repeated measures ANOVA).
There was no effect of feedback type on mean sensitivity, fixation losses or false positives (p<0.01). Participants strongly preferred having feedback relative to no feedback (p<0.001). Subjective ratings of the experience with the three assistant types differed (p<0.001). Post-hoc testing revealed no difference in overall rating of experience between the human and the robot (p=0.40); but both were preferred to the disembodied voice (human versus voice, p<0.01; robot versus voice, p<0.01).
People like human interaction during perimetry. Our data suggests that humanoid robots can be used to replace aspects of this interaction. Regular visual field assessment is a necessary part of glaucoma management. Making the experience more enjoyable for both patients and operators may improve compliance and attitude to perimetry, leading to improved clinical outcomes.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.
NAO Robot assistant
This PDF is available to Subscribers Only