Skip to main content

Table 1 Change in study results* after imputation of censored data

From: Reporting of loss to follow-up information in randomised controlled trials with time-to-event outcomes: a literature survey

   Imputation method
Original treatment effect N Equal case
n (%) 		Best case
n (%) 		Worst case
n (%)
Significant 24 8 (33) 7 (29) 9 (38)
Not significant 23 13 (57) 19 (83) 18 (78)
Total 47 21 (45) 26 (55) 27 (57)
  1. * After imputation of censored data the effect estimate changed direction and the corresponding p-value changed from significant to not significant or vice versa (α = 5%).
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us