Intent-to-treat analysis for longitudinal clinical trials: coping with thechallenge of missing values

Drop-out is a common phenomenon in clinical trials of drug treatments invol ving longitudinal assessments for a fixed duration of follow-up. For these trials intent-to-treat (IT) analysis is usually preferred because time effe cts are seen in practice. The IT analysis mandates that all subjects random ized to a treatment arm should be included in the analysis. The purpose of the present paper is to acquaint both clinicians and statisticians with rec ent statistical methodological advances in handling drop-outs and their usa ge for IT analysis. We discuss a sensitivity analysis of 12-month outcome data to investigate t he efficacy of drug therapy from a longitudinal double-blind placebo-contro lled clinical trial in the maintenance therapy of geriatric major depressiv e illness. Outcome measures consist of monthly Hamilton depression scores. The sensitivity analysis includes endpoint analysis, last observation carri ed forward analysis, repeated measures models and imputation models. Imputa tion models are based on multiple imputations of missing responses deriving from an 'as-treated' model. The model used imputed doses from a plausible treatment scenario after drop-out and a 'propensity-adjusted' model where t he imputations for the drop-outs were obtained from the adhering subjects w ith the same probability to remain on study (propensity) given the observed trajectory prior to withdrawal. Issues related to bias and efficiency of t he estimates obtained by different analyses are discussed. We recommend a m ore widespread use of imputation models for the IT analysis. (C) 1999 Elsev ier Science Ltd. All rights reserved.