Say one continuous variable differentiates between disease and nondisease quite accurately, but as people progress in age, this variable becomes less accurate.

Is there a way to determine the accuracy of this variable on the continuous scale of age?

This is for example the case considering the BMI in predicting the cardio-metabolic health; an increased BMI is less reflective for the cardio-metabolic health as people progress in age.

I have no data at the current moment, but if strictly necessary according to the rules of the forum, I can try to simlate a dataset.