I have a data set consisting of two cohorts of patients treated for a certain group of conditions, one from 2018 and one from 2023. There are good reasons to believe there might be differences between the cohorts in various respects, and I would like to study different things such as whether the gender distribution differs from the cohorts, or if the proportions of patients who get a certain treatment differs.
The problem is that the conditions under study tend to be chronic, which means that a fair amount of the patients included in the 2018 cohort are also included in the 2023 cohort. I estimate that approximately 30% of the patients appear in both cohorts. This cannot be resolved by removing those patients because it might introduce bias.
So how can I analyse this data set with respect to my research questions?
