A simulation study of loss to follow up: The impact on time to event and prevention.
e14097 Background: Patient attrition during study follow up is a concern in all clinical trials, although its impact on study results has rarely been assessed. In oncology, in particular, where studies are lengthier and may be extended into longitudinal studies, there is an increased likelihood of loss to follow up (LTFU) (Gill et al., 2018). This creates a heightened need to understand how it affects the trial’s validity. The loss of data from patients who have been LTFU can reduce a study’s precision and power. This imprecision not only impacts the results of the current study but can also affect future research as well as future patient treatment options. Studies have found that participant characteristics differ in individuals LTFU as compared to those who remain in follow up (Childs et al., 2011; Geng et al., 2008; Hochheimer et al., 2016). This further emphasizes how attrition can skew study results and their interpretation and supports the need to minimize patient attrition during follow up in order to reduce bias and generate robust study estimates. Methods: This study assessed the impact of LTFU rates on the study estimates through simulations using SAS software. While all endpoints can be affected by LTFU, this study assessed time-to-event endpoints. Exponential distribution was assumed with varying rates of LTFU. In addition, the work covered suggestions for reducing LTFU. Results: Even for low rates of LTFU, biases are introduced in time-to event endpoints. Conclusions: Researchers should make every effort to minimize the extent of LTFU in the design and of and conduct of their trials.