JMIR Publications

ABSTRACT

Background:

The emergence of next-generation video see-through head-mounted displays (HMDs), such as the Apple Vision Pro (AVP), has generated considerable interest in the medical field. While preliminary studies highlight AVP’s potential, no controlled experimental study has rigorously assessed its usability for precision-based medical tasks requiring fine motor control and real-world perception.

Objective:

This study aims to evaluate the feasibility of AVP for medical precision tasks, where real-time visualization is critical.

Methods:

To assess AVP’s feasibility, we conducted an experimental user study with 20 healthcare professionals, who performed three different suturing techniques across three intervention conditions. Participants completed the same tasks using AVP, the Microsoft HoloLens 2 (MHL2), and a baseline (without an HMD). A within-subject design was employed, ensuring that each participant experienced all intervention groups. We utilized a mixed-methods research approach, incorporating both quantitative metrics including task completion time, suturing performance, system usability score, cognitive load, virtual reality sickness, and presence score, as well as qualitative insights gathered through interviews.

Results:

Participants took significantly longer to complete the entire task using AVP (570.0 ± 192.0 sec) compared to MHL2 (456.0 ± 120.0 sec, P <.001) and baseline (472.0 ± 143.0 sec, P <.001). The analysis on participants' average suture performance revealed no significant differences across interventions (P = .76). The total raw NASA_TLX score among participants was significantly higher for AVP (43.9 ± 15.9) compared to MHL2 (21.5 ± 13.8, P < .001) and baseline (19.1 ± 15.1, P < .001). The analysis of the presence questionnaire demonstrated a significantly higher presence score for MHL2 (115.0 ± 11.4) compared to AVP (93.7 ± 12.7, P < .001). The overall virtual reality sickness questionnaire score was significantly higher for AVP (66.9 ± 19.8) compared to MHL2 (41.1 ± 9.32, P < .001). Moreover, the calculated system usability score for MHL2 (72.7 ± 8.54) was significantly higher compared to AVP (50.3 ± 14.4, P < .001).

Conclusions:

In conclusion, AVP has potential for non-time-sensitive medical applications or those that emphasize virtual elements over real-world interaction. Its current usability limitations, particularly increased cognitive load and prolonged task execution times suggest that further optimizations are necessary before widespread clinical adoption is feasible.