JMIR Publications

ABSTRACT

Background:

Given the increasing global demand for ophthalmologic care and the anticipated shortage of ophthalmology professionals, innovative solutions are essential for optimizing healthcare delivery. Digital health technologies offer promising opportunities to efficiently manage high patient volumes. Cataract surgery, with its established safety profile and routine postoperative care, is provides an ideal setting for these innovations. Structured clinical questions can effectively identify patients requiring further assessment, enabling clinicians to screen for complications through telephone consultations. Moreover, an artificial intelligence-based follow-up system could take this a step further by automating the process, reducing the need for clinician involvement while enabling more efficient postoperative screening for complications.

Objective:

To assess the clinical safety and effectiveness of an artificial intelligence-based follow-up call system (Dora-NL1) in identifying patients who require further assessment after cataract surgery in the Netherlands.

Methods:

Patients who underwent uncomplicated cataract surgery were eligible to participate. All patients received a Dora-NL1 follow-up call at 1 and 4 weeks postoperatively, in addition to routine postoperative cataract care. Dora-NL1 evaluated postoperative symptoms and made suggestions for care management decisions. The Dora-NL1 outcomes were compared with clinician assessments of recorded Dora-NL1 calls and regular care. User-acceptability was evaluated using the Telehealth Usability Questionnaire (TUQ).

Results:

A total of 105 patients were included in the analysis. Dora-NL1 demonstrated high accuracy compared to clinician-supervised calls, with symptom evaluation accuracy of 89% to 99% and care management decision accuracy of 83% to 88%. At week 1, the sensitivity and specificity compared to standard telephone consultations were 100% and 42%, with no clinical concerns missed. However, at post-operative week 4, compared to an in-person hospital visit, Dora-NL1 did not identify unexpected management changes in 4 patients (4.1%). All 4 of these patients had issues that were only detected after slit lamp examination, reflecting the system’s reliance on patient reported symptoms. Patients rated Dora-NL1 positively, with a mean TUQ score of 3/5, highlighting its simplicity, ease of use, and audibility.

Conclusions:

Dora-NL1 is a safe and effective screening tool for postoperative cataract surgery, offering a safe alternative to telephone consultations, but this cannot fully replace in-person examinations. Clinical Trial: Not applicable.