Development and validation of a novel mobility test for IRDs, from reality to virtual reality

PurposeTo validate a novel mobility test (MOST, MObility Standardized Test) and performance outcomes in real (RL) and virtual (VR) environments to be used for interventional clinical studies in order to characterize vision impairment in rod-cone dystrophies, also known as retinitis pigmentosa (RP). DesignProspective, interventional, non-invasive, longitudinal study (test-retest). Participants89 participants in three experimental phases: 15 non visually impaired (controls) in Phase 1 (average age, 27.4 years; 66% women), 14 participants with RP in Phase 2 (average age, 45.2 years, 36% women), and 60 participants (30 RP; average age, 47.4; 44.6% women; and 30 controls, average age, 47.6 years; 45.4% women) in Phase 3. MethodsWe designed a mobility test (MOST) to be used in both VR and RL and ran three experimental studies to (1) validate the difficulty of the mobility courses, (2) determine the optimal number of light levels and training trials, and (3) validate the reproducibility (test-retest), reliability (VR/RL), sensitivity, and construct, and content validity of the test. A comprehensive ophthalmologic examination was performed in all subjects. Main outcomes measuresThe primary outcome is the performance score in the mobility test. The secondary outcomes include visual acuity, contrast sensitivity, dark adaptation thresholds, static and kinetic visual field parameters, and ellipsoid zone from optical coherence tomography. Correlation between the performance score in the mobility tests and visual function were assessed. ResultsResults revealed that the mobility courses developed exhibited statistically similar difficulty, and that five trials are sufficient to control for the learning effect in a session. MOST is highly reproducible (test-retest intra-class correlations > .98) and reliable (correlation VR/RL = .98). MOST achieved a discrimination between RP participants and controls (accuracy larger than 95% in all conditions) and between early and late stages of the disease (mean accuracy of 82.3%). The performance score is correlated with visual function parameter (.57 to .94). ConclusionMOST is a tool offering validated mobility test, a controlled learning effect, which demonstrates excellent reproducibility and high agreement between real and virtual conditions, as well as sensitivity and specificity to measure disease progression and therapeutic benefit in IRD.