Skip to main content

Table 3 Reporting of methods for MI under MAR

From: A review of the use of controlled multiple imputation in randomised controlled trials with missing outcome data

  n
(N = 110)
%
Type of primary outcome
 Binary 52 47%
 Continuous 52 47%
 Count 4 4%
 Time-to-event 2 1%
Method of MI
 Not stated 47 43%
 Multiple Imputation using Chained Equations (MICE/FCS) 35 32%
  Specified Imputation Model Type(s) within MICE/FCS 12 11%
MCMC MI/algorithm/method 8 7%
 Regression based MI 7 6%
  Specified imputation Model Type within regression based MI 4 4%
PMM 5 5%
MVN imputation 4 4%
MVN imputation (non-monotone missing patterns) and regression MI model (monotone patterns) 1 1%
MICE (non-monotone missing patterns) and regression MI model (monotone patterns) 1 1%
MCMC (non-monotone missing patterns) and PMM (monotone patterns) 1 1%
Propensity score MI 1 1%
Specified variables in imputation model 52 47%
 Imputation model incl. All variables in analysis model only 6 5%
 Imputation model incl. All variables in analysis model + auxiliary variables 37 34%
 Imputation model did not include all variables in analysis modela 9 8%
Did not specify variables in imputation model 58 53%
Imputation model incl. All variables in analysis model + auxiliary variables 8 7%
Reported the number of multiple imputations 70 64%
No. of imputations
 5 9 8%
 10 8 7%
 11–20 25 23%
 21–50 17 15%
 100 8 7%
 200 1 1%
 1000 2 2%
 Not stated 40 36%
Specific procedure/command(s) (software) for implementing MI 26 24%
 IVEware software 1 1%
 MICE (R) 3 3%
 Proc MI (SAS) 3 3%
 Proc MI and Proc MIANALYZE (SAS) 5 5%
 Proc MIANALYZE (SAS) 2 2%
 Realcom Impute 1 1%
 Ice (Stata) 2 2%
 MICE (Stata) 1 1%
 Mi impute (Stata) 5 %
 MI impute and mi estimate (Stata) 1 1%
 Missing data module in SPSS 24b 2 2%
 Not stated 84 76%
Rubin’s rules used for inference
 Yesc 25 23%
 Nod 1 1%
 Not stated 84 76%
Analysis status
 Primary 2 13%
 Sensitivity 14 87%
Performed diagnostic check of imputations 1 1%
  1. a9 trials did not include all variables in the analysis model in the imputation model and included auxiliary variables. bOne trial specified that the Multiple Imputation-Automatic method was used. c Explicitly stated (n = 18) or inferable from specified software or reference (n = 7). dOne trial reported presented the overall 95% confidence using the mean of the values for the lower and upper confidence intervals. Percentages are rounded to 0 decimal places so may not sum exactly to 100%