They were able to connect the HAALSA cohort back to the underlying census data in Agincourt, South Africa to identify people who would have been eligible to enroll had they survived to baseline. They were able to identify 9000+ premature deaths in the source population #SER2024

We can use different methods to addresss censoring from death during a study, from prebaseline mortality, but what can we do about pre-selection mortality? #SER2024

There are some identifiability assumptions for death. 1. Exchangability, 2. Positivity, 3. Consistency. #SER2024

What are competing events? Ones that preclude the outcome of interest SER2024 @epiresearch