In most scleral lens practices, keratoconus (KC) is one of the top indications for scleral lens fitting. In addition to correcting corneal irregularity and visual improvement scleral lenses can delay or reduce the need for corneal transplantation in those who have corneal irregularity and who cannot tolerate other corrective options.

In fact, one study shows that individuals who have KC fit in scleral lenses achieved faster vision improvement and experienced fewer complications compared to those who have KC and who underwent penetrating keratoplasty.¹

Similarly, another study reveals that 40 out of 51 severe KC eyes were successfully fit with scleral lenses, reducing the need for keratoplasty by more than 50%.² Additionally, a third study shows a significantly lower likelihood of keratoplasty in scleral lens wearers.³

Scleral lenses should be considered before surgical intervention for individuals who have corneal irregularity and who cannot tolerate other corrective options successfully.

Many individuals who have KC want additional information, such as the success rates of scleral contact lenses, on their condition. As a result, some search on “Dr. Google,” while more recently others look to artificial intelligence for patient education.

A study on assessing the understandability, accuracy, actionability, and readability of chatbot responses to patient-written questions about KC compared to patient education provided from the American Academy of Ophthalmology (AAO) was carried out.⁴

Twenty patient-focused questions about KC from the AAO website were posed to ChatGPT, Bard, and Copilot. Two ophthalmologists independently evaluated the responses from the chatbots and the AAO website based on, understandability, accuracy, and actionability using the Structure of Observed Learning Outcome (SOLO) taxonomy, the Patient Education Materials Assessment Tool (PEMAT)-Understandability, and the PEMAT-Actionability, respectively. Readability was also evaluated by comparing the Flesch Reading Ease scores obtained from the website.

Bard scored significantly higher than the chatbots ChatGPT-3.5, Copilot, and the AAO website based on the SOLO taxonomy and the PEMAT-Understandability (p < 0.001 for each), with no significant differences observed among the other groups. Additionally, the study reveals that Bard and ChatGPT scored significantly higher than the AAO website (p = 0.001) in the same area. However, the AAO website scored considerably higher than Bard on the Flesch Reading Ease scale (p = 0.017), while no significant differences were found among the other groups.

This study shows that chatbots appear to have the potential to provide accurate, understandable, and actionable answers, making them a resource for educating KC patients under clinician supervision. Further, their use may help reduce the burden of patient education on the health care system, while enhancing patient awareness.

Reference(s):

1. DeLoss KS, Fatteh NH, Hood CT. Prosthetic Replacement of the Ocular Surface Ecosystem (PROSE) scleral device compared to keratoplasty for the treatment of corneal ectasia. Am J Ophthalmol. 2014 Nov;158:974-982.

2. Koppen C, Kreps EO, Anthonissen L, Van Hoey M, Dhubhghaill SN, Vermeulen L. Scleral Lenses Reduce the Need for Corneal Transplants in Severe Keratoconus. Am J Ophthalmol. 2018 Jan;185:43-47.

3. Ling JJ, Mian SI, Stein JD, Rahman M, Poliskey J, Woodward MA. Impact of Scleral Contact Lens Use on the Rate of Corneal Transplantation for Keratoconus. Cornea. 2021 Jan;40:39-42.

4. Özer Özcan Z, Doğan L, Yilmaz IE. Artificial Doctors: Performance of Chatbots as a Tool for Patient Education on Keratoconus. Eye Contact Lens. 2024 Dec 31.

Melissa Barnett, OD, FAAO, FSLS, is the director of optometry at the University of California, Davis. She has received remuneration from ABB, Abbvie, Alcon, Azura, Bausch + Lomb, BCLA, Bruder, Bruno Vision Care, Cloudbreak Pharma, CooperVision, Dompé, Epion, JJVC Vistakon, Lentechs, Ocusoft, Orasis, Science Based Health, STAPLE Program, Tarsus, Théa Pharma, and Visus Therapeutics.

 This editorial content was supported via unrestricted sponsorship.