First off I want to give much needed credit to the authors and all the investigators that were involved in this worldwide trial of a potential new therapy for obesity - an ever increasing problem that needs more therapeutic options and research.
My question is concerning the use of "last observation carried forward" in the data analysis. Obviously every data imputation method has its limits, and with last observation carried forward the individuals that drop out early could skew the data to show less of an effect at one year (presuming they had little weight loss at the time of dropout) compared to if they lost significant weight and then dropped out late in the study.
I ask because it looks like 692 out of 2481 (~28%) did't complete the study, which isn't an insignificant number. Is there any indication of when most individuals dropped out of the liraglutide arm? Do you think the weight loss at the end of the study is biased on the low side because of 'early' drop outs?