If the three measurements used different methods, I think to compare them the results should be adjusted before being compared. The key for this strategy is to find the adjustments, probably by looking at the results of three different measurement methods at the same time using the same sample. I hope this will help.

 I think the most important at this point is to know if in each method the SCr was calibrated or no against IDMS-traceable creatinine?, if the answer is no... so, you can apply the IDMS-traceable creatinine by multiplying Scr x 0.95 (CKD-EPI equation for IDMS-traceable creatinine). The second option is to re-measure the creatinine with the gold method.

You can read this two papers:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5410977/pdf/gfw318.pdf

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5237152/

The idea put by Tana looks reasonable. At least few samples are examined by the three methods simultaneously and see if the three methods yield similar reading from each sample. If this is the case than comparing your test results at the three points in time are possible and can be accepted for exploring change over time. 

I wonder why did you use three different methods? 

Thank you very much for your answers.

Three different assays were used because the study started in 2002 and with time as we reached 2014 the assays for measuring creatinine changed, the older methods were no longer available in the labs.

I did find some information regarding the different methods and I am looking forward to use the calibration against IDMS. However, in all papers I find a mean offset assigned to each method. will a mean value be okay to apply on any plasma creatinine value. I was thinking a percentage adjustment depending on the plasma creatinint levels would make more sense.

So now I am stuck in finding the best possible calibration factor for the methods.

Yes they were kit tests.....   I will try if I could recalculate using the documentation.

Saima