Knowledge of epidemiology and statistics for an ophthalmic practitioner

Ophthalmologists are frequently confronted with treatment options that claim to be better than those currently in use. Statistically significant P values are invariably provided by way of proof. For many ophthalmologists a simple look at this revered P value is enough evidence that a statistically significant result has indeed been obtained. Unfortunately, traditional interpretation of a study based on the P value at an arbitrary cut-off (P<0.05 or any other value) limits the ability to fully appreciate clinical implications. In this article the authors introduce the reader to and illustrate the use of "confidence intervals" as opposed to P values in examining the applicability of study results. Further, what is statistically significant may not necessarily be clinically significant; perhaps not enough for the practitioner to change from the currently preferred method of treatment. To resolve this, the authors have also used common ophthalmic examples to introduce the "number needed to treat", as a simple clinical approach for the practising ophthalmologist wishing to assess the clinical significance of treatment options.