Dr Jamie Chew, ophthalmologist and co-founder of i-scribe.
In ophthalmology, small errors can have big consequences. A short consultation involves visual acuity measurements, intraocular pressure readings, imaging interpretations, medication adjustments, and surgical planning.
As Aayushi Khillan writes, capturing this accurately isn’t optional; it underpins patient safety and continuity of care.
With high patient volumes and frequently complex consultations, documentation is one of the heaviest parts of the job. For many ophthalmologists, note-taking time extends well beyond the consultation itself, adding to administrative burden and reducing time available for patient care.
On the frontline of the paperwork crisis: ambient artificial intelligence (AI) scribes listen to clinician-patient interactions and generate structured medical documents, offering a potential solution to one of the specialty’s most persistent challenges. But can these tools deliver the level of clinical accuracy required in a specialty as detail driven as ophthalmology?
Evaluating Accuracy in Real-World Ophthalmology Settings
To address this question, i-scribe, an AI-powered documentation platform, has taken a research-led approach to development. Rather than relying solely on technical benchmarks, the company has focused on evaluating performance in real clinical environments.
Findings from a recent study, presented at the 2025 Australian Society of Cataract and Refractive Surgeons (AUSCRS) meeting in Darwin, provide one of the first real-world evaluations of AI scribes in ophthalmology.1
The study analysed 26 recorded consultations, each conducted with patient consent. Recordings were played simultaneously across three identical systems: i-scribe and two alternative AI scribes, all using the same hardware.
A reference transcript was generated using a large language model, and the outputs from each system were compared against this benchmark. Researchers assessed the notes using a detailed framework covering eight key domains, including accuracy, completeness, clarity, clinical relevance, medico-legal quality, and the presence of unsupported or fabricated content.
The results showed a clear performance gap. i-scribe produced the highest overall quality scores, with a mean of 84.2 (± 8.2), while the competing systems averaged 71. Statistical analysis confirmed that this gap was significant (Friedman test p=0.0047).
As consultations became longer and more complex, competing systems showed a clear drop in performance. i-scribe, however, did not. Its outputs maintained consistently high-quality regardless of transcript length.
This is not a minor technical detail; it’s clinically critical. Ophthalmology often involves complex consultations, from surgical planning to multiple comorbidities. These are precisely when accuracy matters most. A system that degrades under pressure introduces risk; one that remains reliable supports safe, dependable care.
Focusing on Patients
Accuracy is just the starting point; clinicians also want to know how these technologies will change their day.
Traditionally, note-taking forces clinicians to split their attention, engaging with the patient while simultaneously typing and ensuring nothing is missed. This split attention can come at a cost, to both efficiency and to patient care.
By contrast, ambient AI systems operate in the background, capturing the consultation as it happens. Notes and letters are generated from the captured transcript, reducing the need for manual documentation during or after the appointment. The interaction becomes human again; more listening, more presence, and with better notes as a bonus.
Improving Patient Understanding Beyond the Consultation
Beyond time and rapport, there’s a deeper impact: Patient understanding of their conditions and treatment. Patients typically only retain a fraction of what’s discussed during a medical appointment. Complex terminology, anxiety, and information overload can all limit comprehension.2
AI-generated patient summaries help close this gap. While it produces detailed clinical documentation in medical language for doctors, i-scribe can also generate clear patient-friendly explanations of the consultation. This allows patients to revisit key information in their own time.
Enhancing Collaborative Care
Ophthalmology is inherently collaborative, depending on multiple providers and clear communication between them. AI-generated documentation can play a key role in this process.
Structured notes and letters ensure that all clinicians involved in a patient’s care have access to consistent, comprehensive information. This is especially important in outreach settings, where patients may return to local services after seeing a visiting specialist.
Clinician Built, Clinic Ready
i-scribe wasn’t built in a vacuum. It was built by clinicians who understand the realities of a busy ophthalmology practice.
Led by a founding team that includes practising ophthalmologists, the platform reflects firsthand experience of documentation pressure in high-volume clinical environments.
“i-scribe is fine-tuned for ophthalmology, with specialty-specific templates and training that recognise terms like CCT (central corneal thickness), VA (visual acuity), and IOP (intraocular pressure), while adapting language depending on whether notes are written for optometrists or GPs,” said co-founder and practising ophthalmologist Dr Jamie Chew.
The system is also designed to tailor its outputs depending on the audience, whether generating detailed clinical notes for specialists, referral letters for general practitioners, or simplified summaries for patients.
Early adoption data presented at AUSCRS 2025 suggested growing engagement within the profession since its official launch in early 2024. Today, more than one in three Australian ophthalmologists use i-scribe.
The Impact on Care Delivery
For all of this to work, accuracy remains fundamental.
In ophthalmology, AI scribes are emerging not just as documentation tools, but as enablers of better care delivery.
The AUSCRS data reinforces a critical point: accuracy and reliability are non-negotiable. Real-world experiences from outreach services point to broader benefit: improved communication, better patient understanding, and more efficient consultations.
This has real consequences for clinicians. Reducing administrative workload can help alleviate burnout and improve job satisfaction. Enhancing patient understanding can lead to better adherence and outcomes.
The real value isn’t just better notes; it’s better consultations. As further research emerges, particularly in diverse and underserved settings, these technologies are likely to play an increasingly important role in shaping the future of ophthalmic care.
An i-scribe Case Study: Transforming Outback Communication |
|---|
| In the vast distances of regional Western Australia, access to specialist eye care often comes with significant barriers. Patients may travel hundreds of kilometres to attend an outreach clinic, and for clinicians working in these high-demand settings, consultation time is precious.
At Lions Outback Vision outreach clinics, the introduction of i-scribe has reshaped the consultation experience. Rather than dividing their attention between the patient and the computer screen, clinicians find they can remain more present during appointments. “It helps us focus on building rapport, rather than documenting in real time,” explained Aboriginal Eye Health Coordinator Kerry Woods. The impact has been particularly noticeable in remote and culturally diverse communities across the Pilbara, Kimberley, and Goldfields, where communication challenges are common. Language differences, varying levels of literacy, and the sheer volume of information discussed during a consultation can make it difficult for patients to retain important details once they leave the room. Mr Woods saw the difference immediately during Vision Van clinics. Patients were no longer relying solely on memory. Instead, they left with a written AI summary of their appointment that clearly outlined what had been discussed and what would happen next. “They’re thorough but written in a way that patients can actually understand,” Mr Woods said of the i-scribe summaries. “It’s really valuable for patients to take something home and read it later so they can fully grasp what was explained to them.” For many patients, the summaries also became a practical tool for continuity of care. They could share the information with family members, carers, Aboriginal health workers, or other healthcare providers involved in their treatment journey. “Nine times out of 10, people won’t remember everything,” said Mr Woods. “Sometimes patients come out and don’t fully understand what the doctor has said. Having something written down makes a big difference.” In these remote healthcare settings, where patients often move between multiple providers and services, clear communication is critical. By reducing the documentation burden for clinicians and improving understanding for patients, i-scribe is helping to create more connected and informed care across rural Australia. |
The Lions Outback Vision Van.
Aayushi Khillan BBMed is a medical student, she is simultaneously pursuing a Doctor of Medicine and a Doctor of Philosophy. Ms Khillan is the research coordinator at i-scribe.
References
- Assessing the clinical quality of AI-generated medical notes. Presented at the 2025 Australian Society of Cataract and Refractive Surgeons (AUSCRS), Darwin, Northern Territory, 16-19 July 2025.
- Kessels RP. Patients’ memory for medical information. J R Soc Med. 2003 May;96(5):219-22. doi: 10.1177/014107680309600504.
