Meta-analysis topic ideas in ophthalmology

Ophthalmology has some of the best-defined outcomes in medicine, such as best-corrected visual acuity, intraocular pressure and retinal thickness on optical coherence tomography, which makes pooling straightforward. It also has an unusual statistical issue: most patients contribute two eyes, and the two are not independent.

Topics tend to come from fast-moving areas such as anti-vascular endothelial growth factor therapy, myopia control in children, glaucoma surgery and imaging with artificial intelligence, where trials are frequent and a review dates quickly.

Find a ophthalmology topic that has not been done

SynthGap searches the published literature, checks each candidate against existing systematic reviews, and proposes topics with the primary studies behind them. Start on the Free plan: no card needed, and every run shows every topic it finds.

Where gaps tend to hide in ophthalmology

Myopia control

Atropine at different concentrations, orthokeratology, multifocal lenses and outdoor time are tested in trials with different follow-up and outcome definitions (axial length versus refraction). Network and dose-response syntheses depend on harmonising these.

Retinal disease treatments

Anti-VEGF agents, biosimilars, dosing intervals and treat-and-extend regimens are compared in many trials, some industry-sponsored, with outcomes at different time points.

Glaucoma and surgery

Minimally invasive glaucoma surgery, laser trabeculoplasty and interventions to improve adherence to eye drops have small trials and many observational studies.

Dry eye and ocular surface disease

Treatments are tested with symptom scores and clinical signs that correlate poorly, so reviews should keep symptoms and signs separate.

Screening and imaging with artificial intelligence

Diabetic retinopathy and glaucoma screening tools are tested in accuracy studies with different cameras, populations and reference graders.

Outcomes and effect measures that pool well in ophthalmology

Best-corrected visual acuity: mean difference in logMAR or in ETDRS letters (a gain of 15 letters is a common threshold). Convert Snellen consistently and say how counting fingers or hand movements was handled.
Intraocular pressure: mean difference in mmHg, with the time of day and the method of tonometry.
Central retinal thickness on optical coherence tomography: mean difference. Note the device, as thickness values from different machines are not interchangeable.
Axial length and spherical equivalent refraction in myopia: mean differences over a fixed follow-up, ideally reported separately.
Sensitivity and specificity of screening tools: hierarchical models, with the camera type, grader and reference standard stated.

Common traps in ophthalmology reviews

Counting both eyes of one patient as independent participants, which understates the uncertainty.
Pooling visual acuity measured on different charts without converting to one scale.
Treating different anti-VEGF drugs, doses and regimens as one intervention.
Short-term refractive changes in myopia trials that rebound after treatment stops.
Mixing diabetic macular oedema, age-related macular degeneration and vein occlusion in one analysis of retinal thickness.

Check before you commit

Before committing months to a topic, check that nobody has beaten you to it. Search PubMed with its systematic review filter, the Cochrane Library, the PROSPERO register (for reviews that are registered but not yet published) and Epistemonikos. Look at how recent the latest review is and whether new trials have appeared since its search date. Step-by-step guide.

Starting points to try in SynthGap

These are areas to explore, not claims that they are open. Some may already be well covered, and finding that out quickly is the point.

Low-dose atropine for myopia control in children: concentration and duration
Treat-and-extend versus fixed dosing of intravitreal injections
Interventions to improve adherence to glaucoma eye drops

Questions

How do I handle two eyes per patient?

Prefer analyses that use one eye per person or that account for the correlation between fellow eyes. If the primary studies do not report in a way that allows this, say so as a limitation and check in a sensitivity analysis whether it changes your results.

How do I check whether an ophthalmology meta-analysis already exists?

Search PubMed with the systematic review filter, the Cochrane Library (Cochrane Eyes and Vision has a long list of reviews), PROSPERO and Epistemonikos.

Is a diagnostic accuracy review feasible for screening with artificial intelligence?

Yes, but expect heterogeneity from devices and populations, and plan subgroups by camera type and by whether the algorithm was tested on a separate dataset.

Other specialties

This page is general methodological guidance, not medical advice, and it does not list specific unreviewed topics. SynthGap's suggestions are AI-generated from the published literature, so verify them before registering a protocol.