Ian Carmody

Introduction

When I began as a foundation doctor, I assumed ophthalmology would sit at the edge of my training. I expected the occasional red eye in A&E, but little more. Instead, ophthalmic problems kept appearing on the wards — often quietly, sometimes urgently — and each encounter reminded me that vision wasn’t a minor detail — it often sat at the heart of patient safety.

Falls, medication errors, delayed diagnoses, missed follow-ups: vision loss played a role in all of these. Yet it was rarely mentioned in clerking notes or discharge summaries. Ophthalmology taught me that patient safety isn’t just about vital signs and treatment charts. It can be as basic as whether someone can see well enough to eat, walk to the bathroom, or understand what’s being said to them.

Vision loss as a hidden safety risk

On the Medical Admissions Unit, I clerked a patient who had been admitted with “confusion.” Her notes described her as quiet and compliant. When I examined her more closely, she admitted she could barely see her food tray. Her glasses were at home, and without them she was lost in an unfamiliar environment. She wasn’t confused at all — she was living with significant visual loss and struggling to make sense of her surroundings.

I began to notice similar patterns elsewhere. Patients admitted after falls who could not see hazards clearly. Stroke patients whose hemianopia was overlooked until late in their admission. People with dementia whose agitation worsened because they couldn’t see faces or objects around them. Each time, the presenting problem was documented; the vision loss was not.

This is not a small gap. According to RNIB, more than two million people in the UK are currently living with sight loss, a number expected to double by 2050 (1). Studies show that visual impairment significantly increases the risk of falls, medication errors, malnutrition, and social isolation (2). Yet vision is rarely included in standard clerking templates or discharge summaries.

For me, the lesson was clear: patient safety isn’t just about observations and early warning scores. It often starts with the basics — can the patient see well enough to eat, move around safely, or follow instructions? Asking simple questions such as “Do you usually wear glasses?” or “Can you see your food?” became a quiet but powerful safety check, one that often uncovered risks others had missed.

Time-critical diagnoses

Ophthalmology also showed me how unforgiving the timelines can be.

I remember a man admitted with a headache who was treated initially for sinusitis. Only later, after formal visual field testing, did we discover bitemporal hemianopia and raised intracranial pressure. Another patient presented with a painful, red eye that had been dismissed as conjunctivitis; it was in fact acute angle-closure glaucoma, a condition where every hour counts.

These experiences were sobering. I realised quickly that in ophthalmology, delays could cost vision in a way I hadn’t seen elsewhere. In wet age-related macular degeneration, a missed anti-VEGF injection can mean irreversible central vision loss (3). In temporal arteritis, a delay in treatment risks bilateral blindness (4). Yet outside of ophthalmology, these conditions rarely carry the same recognition as “two-week wait” cancer referrals. For patients, the consequences of delay are just as life-changing — and in some cases more immediate.

What I learned was that patient safety in ophthalmology is about recognising when time matters. As a foundation doctor, I didn’t need to make the diagnosis myself, but I did need to recognise when the story didn’t fit a benign explanation. Escalating quickly — even when uncertain — became the safest thing I could do.

Diagnostic overshadowing

One of the most striking patterns I noticed was how vision problems were often hidden beneath other, more obvious diagnoses.

A gentleman admitted with multiple medications had been labelled as “non-compliant.” When I spoke with him, it became clear that the issue wasn’t unwillingness but inability: his eyesight was so poor that reading his prescription labels was almost impossible.

Another patient, living with Parkinson’s disease, was admitted following several falls. The clinical focus was understandably on mobility and tremor, but only later did we realise that his cataracts were so advanced he could barely see kerbs or steps. His risk of falling was not just neurological, but visual.

These were not failures of individual clinicians, but reminders of how easily vision can be overshadowed by dominant conditions. In complex patients, sight loss is rarely the headline problem, but it often sits in the background, making treatment harder to follow, recovery more dangerous, and daily life less independent (5).

For me, the safety lesson was simple: we need to look beyond the primary diagnosis. A patient’s vision may not be the reason they were admitted, but ignoring it can undermine the success of everything else we do.

Technology and risk

Ophthalmology is often at the forefront of new technology, and artificial intelligence is no exception. In diabetic retinopathy screening, AI is already being used to grade retinal images (6). Research is rapidly expanding into automated glaucoma detection, age-related macular degeneration classification, and triage of OCT scans (7).

But working on the wards reminded me that technology can only go so far. No algorithm can replace the simple act of asking whether a patient can see well enough to read their food menu or recognise a familiar face. AI may spot disease on a scan, but it cannot prevent a fall if a patient cannot see the edge of their bed.

Technology also carries its own risks. Automation bias — the tendency to trust a machine’s decision over our own clinical judgement — is real. If AI is used without oversight, there is a danger that safety nets are quietly removed. Ophthalmology could become a model for safe AI use in the NHS, but only if safeguards, transparency, and human validation are built in from the start.

For me, the balance was clear: technology may help us see more disease, but on the wards I found nothing replaced simply asking if a patient could see their surroundings.

Conclusion: Lessons for safety beyond ophthalmology

Ophthalmology changed the way I think about patient safety. It showed me that risks are not always written in the notes, but in how a patient moves around a ward or reaches for a meal. It taught me that some conditions demand escalation without delay, and that time really can mean the difference between sight and blindness. It revealed how vision problems are often hidden behind more obvious diagnoses, shaping outcomes in ways that are easy to miss. And it reminded me that technology, however advanced, cannot replace the awareness and vigilance of clinicians at the bedside.

The lessons are not only for ophthalmologists. For any doctor, nurse, or allied professional, the questions are simple: Can this patient see well enough to be safe? Do their symptoms fit a benign story? Might their vision be shaping their ability to recover?

I didn’t expect ophthalmology to shape my foundation years, but it has stayed with me. The safety lessons I learned through a different lens are ones I now carry into every specialty: noticing the small things others might miss, speaking up quickly when time is short, and keeping sight at the centre of safe care.

References

1. Royal National Institute of Blind People. The State of the Nation: Eye Health 2023. London: RNIB; 2023. Available from: https://www.rnib.org.uk
2. Crews JE, Campbell VA. Vision impairment and hearing loss among community-dwelling older Americans: implications for health and functioning. Am J Public Health. 2004;94(5):823-9.
3. NHS England. The NHS Long Term Plan. London: NHS England; 2019. Available from: https://www.longtermplan.nhs.uk
4. Hayreh SS, Zimmerman B. Management of giant cell arteritis. Ophthalmologica. 2003;217(4):239-59.
5. NatCen Social Research. Health Survey for England 2024: Adult health and wellbeing. London: NHS Digital; 2024.
6. Tufail A, Rudisill C, Egan C, et al. Automated diabetic retinopathy image analysis compared with manual grading for detecting referral-warranted disease. Eye (Lond). 2017;31(4):592-8.
7. Ting DSW, Pasquale LR, Peng L, et al. Artificial intelligence and deep learning in ophthalmology. Br J Ophthalmol. 2019;103(2):167-75.