I have a dataset where I want to predict an outcome (EDSS at last follow up, measured in a numeric scale which is from 0 to 10, including only .5 decimals). From each patient, I have exactly 2 measures of biomarkers from each patient. I would like to create a predictive model using the biomarkers to predict EDSS at last follow up.

I tried linear mixed models, but it throws an error because convergence is not achieved. This is the formula I am using

EDSS_lastfollow ~ nfl + (1|ID)

I have tried using a gaussian link function and also using the outcome as a ordinal variable, but I could not reach convergence. I suspect it is because I do not really have a repeated outcome, but repeated covariates. I would like to take into account the ID variability of the covariates to model it. Do you have any idea on how I could proceed to model it? Just a simple lm would be better?