Anna Brouwer

Summary, discussion & future perspectives Guideline to determine when an ERG has improved If future research shows an association between a prolonged cone b-IT and a worse prognosis in uveitis, an improvement or a worsening of the ERG can have clinical consequences. In such a future, guidelines need to be created to define when a significant change on the ERGhas occurred. ISCEV standards currently describe clearly how to determine when an ERG is abnormal or not, 31 but do not specify how much an ERG parameter needs to change to be defined as improved, stable or worsened. In Chapter 7 we addressed this problem. In our uveitis cohort we measured a LA ERG twice to compare the differences between the DTL positioned along the LLP and in the FP. This experiment was in essence a repeated measures test. Therefore, we used the distribution of the difference in implicit time from the two LA ERGs to calculate 2 SDs to determine how much an implicit time needed to change to be most likely due to uveitis and not due to a measurement error or variability. It would have been preferable to have repeated measures results from our reference data. Unfortunately, we did not have this data, probably because it is quite time consuming to create new reference values for an ERG and measuring an ERG twice to be able to use it for monitoring purposes is double the amount of work. A repeated measures test would be the best to gain insights in the measurement error of the ERG. Therefore, it would be optimal that when new reference values are obtained, an ERG is measured twice at different time points. This repeated measures test would enable us to create cut-off values for change more precisely than using, for instance, the distribution of the reference values alone. However, when repeated measure data are lacking, √ 2 * 2SD (the repeatability formula) of the distribution of the reference values might be the best, but conservative, way to determine a significant change. But, whichever option is chosen, it is paramount that such a guideline should be created to be able to determine when an ERG has clinically significantly changed. Especially because many ophthalmologists believe that the ERG can be used for monitoring. Without guidelines for monitoring, created by a society such as ISCEV or another, it is impossible to compare scientific results to one-another. 169 8