Oliver Ng: Impact of exposure definition on exposure-outcome associations: An investigation on the association between assigning exposure periods and its outcome.
Master thesis final presentation
Time: Thu 2023-06-08 11.25
Location: Meeting room 9, floor 2, house 1, Albano
Respondent: Oliver Ng
Supervisor: Daniel Ahlberg
Abstract.
Determining the association of a treatments effect during time exposure is of great importance in regards to having a proper understanding of the effects, both harmful and beneficial in the population. However, the definition of what is exposed time during treatment varies depending on how the user defines these exposure periods. To answer the question on the impact of exposure definition we make use of the Cox proportional hazards model to quantify the difference and accuracies of three common definitions and in addition, for comparison with a known association, a simulated data set is used. Initial model with the real data set showed that the three definition show differing results while the Cox model showed two model returned equivalent results with emphasis on how the users handled certain parameters such as the gap between refills, the simulated data further support this and also shows that given extreme conditions or just non-ideal makes some of the definition return wildly differing results or models with no statistical power.