Looking for some signposting really. Thanks in advance

I'm trying to determine how useful a biomarker panel is.

I've recruited 15 participants who have provided a sample on 5 consecutive days the panel has been used on each of these samples and provided 90 measurements for each sample.

I've looked at coefficients of variation to describe the within subject variability of each measure.

I'd like to describe how similar each patients' panel looks across the 5 measurements. i.e. does the panel look similar across the 5 visits? when i do simple PCA plot, each participants' samples appear to cluster closely to one another but I'm not sure of the best way to describe that objectively.

I think it must be to do with the spatial Euclidean distance between visits per patient but not sure how to construct that kind of summary statistic. I think spatial Euclidean distance would only give me pair-wise comparisons. Would it be reasonable to measure the distance between each pair and take an average?

Even then, though, I suppose, I need some sort of reference value? i.e. what is a good "average spatial Euclidean distance". Maybe a measure of the total variability of all the samples?

I think I have got myself caught down a rabbit hole and there is a more simple and efficient approach to take. Grateful for any advice


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.