Author
Dr. Ahmed Hamdy Oreaba, FRCOphth, FRCS, MRCS
Where the model comes from
- Operations in the development dataset
- 961,208 cataract operations
- Patients
- 682,381
- Centres
- 136
- Surgeons
- 3,198
- Source
- The Royal College of Ophthalmologists' National Ophthalmology Database Audit - Date: September 2025
- Observed PCR rate in the modelling dataset
- 1.01%
- Reported area under the receiver-operating curve
- 70.0%
What the estimate is, and is not
- The model provides an estimated probability, not certainty.
- Important clinical variables may be absent from the model.
- The model does not replace assessment of the individual patient, surgeon competence, supervision or local governance.
- Posterior polar and traumatic cataracts were excluded from the development dataset where PCR was expected as part of the surgery.
- This prototype uses rounded publicly reported coefficients, so its figures differ very slightly from the source’s worked examples.
- Full-precision coefficients and regulatory validation would be required before routine clinical deployment.
Coefficient table implemented in this prototype
Model version: RCOphth-NOD-Report17-published-rounded-coefficients-2025-09-10
| Term | Applied | Coefficient |
|---|---|---|
| Intercept | — | -6.391 |
| Operating surgeon grade — consultant | Reference | 0 |
| Operating surgeon grade — career-grade non-consultant | Yes | +0.461 |
| Operating surgeon grade — more-experienced resident | Yes | +0.910 |
| Operating surgeon grade — less-experienced resident | Yes | +1.323 |
| Age | per year | +0.015 |
| Sex category — female | Reference | 0 |
| Sex category — male | Yes | +1.114 |
| Age–sex interaction (male only) | per year | −0.013 |
| Able to lie flat and cooperate — yes | Reference | 0 |
| Able to lie flat and cooperate — no | Yes | +0.171 |
| Systemic diabetes mellitus — no | Reference | 0 |
| Systemic diabetes mellitus — yes | Yes | +0.215 |
| First treated eye | Reference | 0 |
| Second treated eye | Yes | −0.101 |
| Previous intravitreal anti-VEGF — no | Reference | 0 |
| Previous intravitreal anti-VEGF — yes | Yes | +0.356 |
| Pupil size — small | Reference | 0 |
| Pupil size — medium | Yes | −0.424 |
| Pupil size — large | Yes | −0.434 |
| Pupil size — not recorded | Yes | −0.420 |
| Anterior chamber depth | per mm | +0.107 |
| Preoperative distance VA | per LogMAR unit | +0.412 |
| Amblyopia | Yes | +0.187 |
| Brunescent, white or mature cataract | Yes | +0.879 |
| Diabetic retinopathy | Yes | +0.162 |
| Glaucoma | Yes | +0.538 |
| High myopia | Yes | +0.333 |
| Previous vitrectomy | Yes | +0.216 |
| Pseudoexfoliation or phacodonesis | Yes | +1.243 |
The published coefficient table gives +0.107 per mm of anterior chamber depth, while the accompanying narrative discusses increased observed risk with shallow anterior chambers. This prototype follows the published coefficient table exactly and this discrepancy must be clarified with RCOphth NOD before any clinical release.
Privacy by design
Every calculation runs in your browser. There is no database, authentication, server function, analytics, cookie, local storage or external API, and no patient-identifiable field exists anywhere in the application. Nothing you enter leaves your device, and nothing is retained once you close or reset the page.
Source publications
Before release
Before public or commercial release, confirm permission to reproduce the model content and terminology and obtain written clarification of any coefficient or documentation uncertainty from RCOphth NOD.
Prototype based on published RCOphth NOD coefficients. Not CE or UKCA marked. Not for use as the sole basis for patient counselling, surgeon allocation or treatment decisions. Apply independent clinical judgement.
Independent prototype. Not affiliated with or endorsed by RCOphth, the National Ophthalmology Database, Medisoft or mediSIGHT.