Table 1

Analysis methods

Method abbreviationAnalysis modelImputation method for missing dataComments
ANCOVA+LOCFANCOVALOCFA common sensitivity analysis
RC+OBSRCNoneNew proposed analysis
ANCOVA+LEANCOVALECurrent standard analysis
RC+FULLRCNot applicableIdeal scenario as benchmark, not representative of real trials. Results expected to be similar to ANCOVA+FULL
ANCOVA+FULLANCOVANot applicableIdeal scenario as benchmark, not representative of real trials. Results expected to be similar to RC+FULL
  • ANCOVA+LOCF: ANCOVA model using LOCF imputed data.

  • RC+OBS: RC model using observed dataset with missing data.

  • ANCOVA+LE: ANCOVA model using LE imputed data.

  • RC+FULL: RC model using full dataset without missing data.

  • ANCOVA+FULL: ANCOVA model using full dataset without missing data.

  • A summary of the steps used in each simulation is shown in online supplemental file 2.

  • ANCOVA, analyses of covariance; LE, linear extrapolation; LOCF, last observation carried forward; RC, random coefficient.