Using Artificial Intelligence–Generated Video To Transform Cataract Consent: A Randomised Controlled Trial With Mixed-Methods Evaluation

Cataract surgery is the most commonly performed operation in the NHS, with over 400,000 procedures annually. Despite low complication rates, consent-related complaints remain common. This study evaluated an AI-generated multilingual video consent intervention developed using Synthesia, a platform producing photorealistic avatar-based educational videos to deliver consistent surgical information. The intervention aimed to address communication inequities by comparing outcomes between English- and Bengali-speaking patients. We assessed its impact on understanding, experience, and engagement versus written leaflets, and explored cultural and socioeconomic influences on perceptions of digital consent.

Prospective randomised controlled trial conducted within a high-volume NHS cataract service at Moorfields Eye Hospital. A parallel qualitative study was embedded within routine clinical pathways to explore patient perspectives on digital consent technologies.

Financial Disclosure:

This study was supported by a grant from the Medical Protection Society (MPS) Foundation. The authors have no financial or proprietary interest in the materials or methods described.

Patients undergoing first-time cataract surgery were stratified by language (English/Bengali) and randomised within strata to receive video or written leaflet information. The primary outcome was information retention via quiz score. Secondary outcomes included anxiety change from baseline to day of surgery, satisfaction, engagement time, and 3-month cancellations. Non-parametric comparisons used Mann–Whitney U tests and categorical variables used chi-square tests. A nested qualitative study (n=10) involved 45-minute semi-structured interviews across five domains: material preferences, digital confidence, communication, and cultural influences. Interviews were audio-recorded, transcribed verbatim, and analysed using thematic analysis.

Video participants showed greater information retention than leaflet groups (p<0.001). Longer engagement time independently predicted higher quiz scores (p=0.001). Satisfaction was significantly higher with video (p=0.003). Anxiety increased from baseline to the day of surgery in all groups, with no differences by language or format, and cancellation rates were similar.

Qualitative interviews found leaflets were often perceived as dense and difficult to retain, while video improved engagement and reassurance, particularly when culturally representative.

Digital literacy, language barriers, environmental pressures, and differing preferences for decision-making (clinician-led versus autonomous) further influenced how consent was experienced.

AI-generated multilingual video consent improves measurable understanding and patient satisfaction in high-volume cataract services. Engagement time emerged as a modifiable factor associated with comprehension, indicating that digital platforms may allow consent exposure to be monitored and optimised. However, improved information retention did not reduce perioperative anxiety, indicating that consent extends beyond information alone. 

Qualitative findings emphasised the roles of interpersonal communication, cultural context, and decision-making preferences. Nevertheless digital tools can greatly improve understanding and communication, and, through cultural representation, support patient autonomy alongside patient-centred care.