Episode 59 - Statistical Bias in Ophthalmic ResearchWelcome back to Basic Sciences in Ophthalmology.

Hi, I am Dr Thomas Whitelaw and in today's episode, we're moving away from the anatomy, physiology and optics of the eye and into another fundamental part of ophthalmic science: how we decide whether research evidence can be trusted.

Imagine reading a study reporting that a particular treatment dramatically reduces intraocular pressure, slows visual-field loss or improves visual acuity.

The statistics are significant.

The graphs look impressive.

And the conclusion appears convincing.

But there is a fundamental question we need to ask before accepting those findings:

Could bias have influenced the result?

Bias is particularly important because it is not simply random variation. It can systematically push a study towards a particular conclusion.

And even a beautifully calculated P value cannot rescue a study whose design has introduced serious bias.

The uploaded ophthalmology texts specifically discuss several important forms of bias, including selection bias, information bias, recall bias, confirmation bias, attrition bias, detection bias, lead-time bias and publication bias.

So today we'll look at three areas:

First: types of bias.

Second: designing unbiased studies.

And third: the impact of bias on study results.