First-order Markov model suitability for different time points

data analysis modeling strategy
,
1

I’m working with a longitudinal dataset of patients divided into two cohorts:

Cohort 1:

  • Tested at 21 days (baseline), 28, 35, and 42 days.

Cohort 2:

  • Tested at 35 days (baseline), 42, 49, and 56 days.

Is this suitable for a first-order Markov model even though patients were not tested at the exact time points and each cohort has a different baseline time point?

1 Like
2

If the baseline measurement itself captures the time effect of entering the study at different points for the two cohorts, then you can compute days from baseline and luckily both cohorts have the same values of these relative measurement times.

2 Likes