Factors influencing compliance in RRD patients with the face-down position via grounded theory approach

A Pilot Survey of Patient Perspectives on an Artificial Intelligence-Generated Presenter in a Patient Information Video about Face-Down Positioning after Vitreoretinal Surgery

INTRODUCTION

Video education is a commonly used patient education tool. However, the impact of integrating artificial intelligence (AI) into video education remains unexplored. This study aimed to examine the acceptability of an AI-generated presenter in a patient information video about face-down positioning after vitreoretinal surgery.

METHOD

We prospectively enrolled participants who were planned for vitreoretinal surgery in which postoperative face-down positioning was recommended at the Royal Adelaide Hospital between December 2022 and September 2023. Participants were preoperatively provided with an educational video presented by an AI-generated presenter, incorporated into a surveyredcap. A pre- and post-video questionnaire was administered electronically.

RESULTS

There were 15 participants included in the study. In the pre-video questionnaire, most participants rated their awareness of special equipment for positioning as "not aware" (33%) and "slightly aware" (33%). The median pre-video six-item Spielberger State-Trait Anxiety Inventory Score was 12 (interquartile range 12-15). In the post-video questionnaire, most participants rated the video's quality as "excellent" (73%) and would recommend it to others (73%). The majority of participants strongly agreed that they understood the AI presenter (60%), felt at ease with the presenter (60%), and trusted the presenter (60%). Four participants (22%) disagreed with the statement: "I was aware the presenter was computer generated."

CONCLUSIONS

Video-based education may provide information that patients find useful, particularly for physical maneuvers such as face-down positioning. The use of an AI-generated presenter was well-received by the majority of patients. Further research regarding the use of AI to develop educational video content is warranted.

Retinal detachment

Retinal detachment (RD) occurs when the neurosensory retina, the neurovascular tissue responsible for phototransduction, is separated from the underlying retinal pigment epithelium (RPE). Given the importance of the RPE for optimal retinal function, RD invariably leads to decreased vision. There are three main types of RD: rhegmatogenous, tractional and exudative (also termed serous) RD. In rhegmatogenous RD, one or more retinal breaks enable vitreous fluid to enter the subretinal space and separate the neurosensory retina from the RPE. In tractional RD, preretinal, intraretinal or subretinal membranes contract and exert tangential forces and elevate the retina from the underlying RPE. Finally, in exudative RD, an underlying inflammatory condition, vascular abnormality or the presence of a tumour causes exudative fluid to accumulate in the subretinal space, exceeding the osmotic pump function of the RPE. The surgical management of RD usually involves pars plana vitrectomy, scleral buckling or pneumatic retinopexy. The approach taken often depends on patient characteristics as well as on practitioner experience and clinical judgement. Advances in surgical technology and continued innovation have improved outcomes for many patients. However, even if retinal re-attachment is achieved, some patients still experience decreased vision or other visual symptoms, such as metamorphopsia, that diminish their quality of life. Continued research in the areas of neuroprotection and retinal biology as well as continued surgical innovation are necessary to enhance therapeutic options and outcomes for these patients.

Artificial intelligence-assisted management of retinal detachment from ultra-widefield fundus images based on weakly-supervised approach

Background

Retinal detachment (RD) is a common sight-threatening condition in the emergency department. Early postural intervention based on detachment regions can improve visual prognosis.

Methods

We developed a weakly supervised model with 24,208 ultra-widefield fundus images to localize and coarsely outline the anatomical RD regions. The customized preoperative postural guidance was generated for patients accordingly. The localization performance was then compared with the baseline model and an ophthalmologist according to the reference standard established by the retina experts.

Results

In the 48-partition lesion detection, our proposed model reached an 86.42% (95% confidence interval (CI): 85.81-87.01%) precision and an 83.27% (95%CI: 82.62-83.90%) recall with an average precision (PA) of 0.9132. In contrast, the baseline model achieved a 92.67% (95%CI: 92.11-93.19%) precision and limited recall of 68.07% (95%CI: 67.25-68.88%). Our holistic lesion localization performance was comparable to the ophthalmologist's 89.16% (95%CI: 88.75-89.55%) precision and 83.38% (95%CI: 82.91-83.84%) recall. As to the performance of four-zone anatomical localization, compared with the ground truth, the un-weighted Cohen's κ coefficients were 0.710(95%CI: 0.659-0.761) and 0.753(95%CI: 0.702-0.804) for the weakly-supervised model and the general ophthalmologist, respectively.

Conclusion

The proposed weakly-supervised deep learning model showed outstanding performance comparable to that of the general ophthalmologist in localizing and outlining the RD regions. Hopefully, it would greatly facilitate managing RD patients, especially for medical referral and patient education.