Review of smartphone digital phenotyping feasibility in 25 individuals with inherited retinal diseases over 12 months

Photo by Worshae / Unsplash

medRxiv Published April 17, 2026 Medically reviewed April 25, 2026 Study authors: Jones, L.; Higgins, B.; Devraj, K.; Crabb, D.; Thomas, P.; Moosajee, M. DOI ↗ By Dr. Lars van Dijk, PhD · Surgical, Procedural & Diagnostic

Key Takeaway

Note that feasibility results from this small IRD cohort are exploratory and not definitive for clinical practice.

This feasibility review assessed the use of a smartphone application for digital phenotyping in individuals with inherited retinal diseases (IRDs). The study was conducted at Moorfields Eye Hospital in the UK and involved 25 participants followed for 12 months. The intervention utilized passive data via HealthKit and SensorKit alongside active patient-reported outcome measures. Primary feasibility indices included enrolment, retention, and completeness of passive datastreams, while secondary outcomes explored associations between digital behavioural markers, visual function, and mental health.

During the 12-month follow-up, 25 participants were enrolled and 23 were retained, yielding a retention rate of 92%. Regarding data validity, 17 participants met the threshold for HealthKit data, while 16 participants met the threshold for SensorKit data. Median daily step count was recorded at 6,087 steps, with a walking speed of 1.18 m/s and a typing speed of 2.19 characters/s. The analysis identified negative correlations between age and typing speed, as well as between age and anxiety-related word use. Additionally, a negative correlation was observed between photopic peripheral visual difficulty and the use of anxiety- and down-related words.

The review reports that adverse events, serious adverse events, discontinuations, and tolerability were not reported. The authors explicitly state that associations are exploratory and that the study is a feasibility study rather than a randomized trial. Consequently, the findings should be interpreted with caution regarding causality and generalizability. Practice relevance and funding conflicts were not reported in the source material.

Study Details

EvidenceLevel 5

PublishedApr 2026

View Original Abstract ↓

This study evaluated the feasibility of collecting passive and active digital phenotyping data using the OverSight iOS application in individuals with inherited retinal diseases (IRDs), and explored associations between digital behavioural markers, visual function, and mental health. Participants with IRDs were recruited from Moorfields Eye Hospital (UK) and followed for 12 months. OverSight passively captures mobility data through HealthKit and typing-derived metrics through SensorKit. Participants completed patient-reported outcome measures (EQ-5D, NEI-VFQ-25, HADS, and MRDQ) within the app. Passive data included step count, walking speed, typing speed, total words typed, autocorrections, and sentiment word categories (anxiety, down, and health-related terms). Feasibility indices included enrolment, retention, and completeness of passive datastreams. Twenty-five participants were enrolled and 92% were retained at 12 months. Seventeen participants met the validity threshold for HealthKit data and 16 also met SensorKit thresholds. Median daily step count was 6,087, walking speed 1.18 m/s, and typing speed 2.19 characters/s. Age was negatively correlated with typing speed and anxiety-related word use, and photopic peripheral visual difficulty was negatively correlated with anxiety- and down-related word use. Digital phenotyping using OverSight was feasible over 12 months. Exploratory analysis suggest mobility, typing behaviour and sentiment markers may represent useful adjunctive indicators of functional vision and psychological outcomes in patients with IRDs.