Measuring disease severity of PAIRS of eyes between treated and untreated groups

Dear forum.

Would someone please beable to provide me some advice on which statistical analysis to apply to an aspect of my PhD data...

I am grading the severity of eye disease x (from pictures of the fundus) between patients with disease x treated with a drug and those with disease x without treatment.

Disease doesn't present in equal severity between left and right eye in one patient.

My goal is to find the statistical difference in disease severity (If any) between treated patients and non treated patients.

Do I average the Left and right eye score per patient? The differences in severity between Left and Right can be quite large. I'm really not sure how to approach this. Any advice would be much appreciated!


TS Contributor
You could perform separate analyses for both eyes. Or perhaps some "mixed" ANOVA
including a within-subject factor.
How is severity of the disease measured (what scale was used), and how large is your
sample size?

With kind regards

Hi Karabiner, Thank you for replying!
I considered using just one eye, however disease doesn't seem to show any bias in severity for one eye over the other!
My sample size is 10 patients per group.
The severity is measured of a scale of 1 - 6 which takes into account severity of inflammation, infiltrate of cells, other clinical symptoms ...



TS Contributor
This seemingly goes beyond statistics, it is a matter of knowledge
about the disease. How is severity of this disease commonly
assessed? As an aggregate of severity measures from both eyes?
You could maybe use the median (probably not the mean) of both
measures. Or, is the measurement from the most affected eye
routinely considered as indicator for the severeity of the

Or maybe you could use both approaches and see whether
results are comparable.

With kind regards