Page 87 - AIH-1-4

Artificial Intelligence in Health Robotics and Vivaldi AI for ALS assessment

The agreement analysis between human and robotic
ALSFRS-R assessment demonstrated promising results.
Indeed, the Bland–Altman plot and ICC revealed good
to excellent agreement between the robotic and human-
administered questionnaires for both total scores and
individual domains (bulbar, motor, and respiratory).
This indicates that the robotic device equipped with AI
technology can effectively administer the ALSFRS-R
Figure 4. Observational grid analysis of patient interactions with the questionnaire, providing comparable results to those
robotic device throughout the ALSFRS-R questionnaire administration
process. Emotional reactions and engagement levels are depicted across obtained by human operators. Furthermore, the Vivaldi
various stages (T1–T4) of the interaction. Peaks in emotional responses AI system based on dichotomous answers (“yes” or “no”)
coincide with specific questionnaire items, highlighting areas of confusion reduces answers ambiguity.
and the need for clarification. Notable shifts in patient engagement and
confidence levels are observed, shedding light on the dynamics of human- In addition, longitudinal analysis showed significant
robot interaction during medical assessments. declining trends in functional status over time, with no
Abbreviation: ALSFRS-R: Amyotrophic lateral sclerosis functional rating significant differences observed between human- and
scale-revised. robot-based administration methods. This suggests that
the robotic device can accurately track disease progression,
maintained their focus on providing answers (14.28%), making it a valuable tool for monitoring ALS patients over
while some displayed neutrality (14.28%), puzzlement time.
(10.71%), and confidence (10.71%). By the end of the
questionnaire (T4), patients demonstrated confidence in The study also evaluated the impact of robotic device
using the robot (17.85%), accompanied by amusement usage on patients’ anxiety and openness to experience.
(14.28%) and interest in the experience (10.71%), with The findings revealed a statistically significant decrease in
fewer instances of puzzlement compared to baseline state anxiety after robotic administration of the ALSFRS-R
evaluation. questionnaire, although anxiety levels remained below the
clinical threshold even before robotic administration. This
6. Discussion suggests that the use of robotic technology in healthcare
settings may contribute to reducing patient anxiety levels,
The urgent need for precise symptom monitoring in potentially improving the overall patient experience.
ALS due to its progressive and ultimately fatal nature
underscores the quest for more accurate assessment There were no significant correlations between patients’
tools. While the ALSFRS-R is widely regarded as reliable, openness to experience and the difference in ALSFRS-R
concerns regarding subjectivity in score attribution scores measured by human operators versus robotic
persist. Moreover, recent advancements in the realm devices, indicating that patients’ personality traits did not
of neuromuscular diseases have seen the emergence of influence the accuracy of robotic assessments.
robotics integrated with AI technology. Finally, the observational grid analysis provided
To address the challenges posed by subjective scoring valuable insights into patients’ emotional reactions during
and explore the potential of robotics in neuromuscular interactions with the robotic device. Patients initially
exhibited curiosity and interest, although some also
care, our study delved into the accuracy and agreement experienced puzzlement. However, as patients became
of ALSFRS-R scores obtained from a robotic operator more familiar with the robotic device, confidence in using
enhanced with AI algorithms, employing the binary tree the instrument increased, accompanied by reduced levels
method, compared to those from a human operator. In of puzzlement and increased interest and amusement. This
addition, we examined patients’ emotional states and suggests that with proper training and familiarization,
perceptions during interactions with the robotic system. patients can become more comfortable with robotic
By investigating the efficacy of AI-powered robotics technology, enhancing the acceptability and usability of
in symptom assessment, we aim to mitigate scoring such devices in clinical settings.
arbitrariness and enhance the precision of ALS monitoring.
Moreover, understanding patients’ experiences and 7. Conclusion
emotions during interactions with robotic technology The study highlights the potential of robotic technology
provides crucial insights into the feasibility and acceptance integrated with AI to enhance functional assessment and
of such innovations in clinical practice. care for ALS patients. The findings support the feasibility

Volume 1 Issue 4 (2024) 81 doi: 10.36922/aih.3732