Table 4 Handling of missing data
Description | n (%) |
|---|---|
Methods used for dealing with missing data (N = 70)a | |
Complete case analysis | 52 (74.3) |
Multiple imputation | 10 (14.3) |
Full information maximum likelihood | 3 (4.3) |
Inverse probability weighting | 2 (2.8) |
Single imputation | 2 (2.8) |
Pattern mixture model | 1 (1.4) |
Compared differences between individuals with and without incomplete data (N = 65)b | |
Yes | 17 (26.2) |
No | 48 (73.8) |
Performed sensitivity analysis to test robustness of results (N = 70)a | |
Yes | 7 (10.0) |
No | 60 (85.7) |
Unclear | 3 (4.3) |
For multiple imputation (N = 10)c | |
Indicated number of imputed datasets | |
Yes | 5 (50.0) |
No | 5 (50.0) |
Unclear | 0 (0.0) |
Reported variables included in imputation model | |
Yes | 4 (40.0) |
No | 5 (50.0) |
Unclear | 1 (1.0) |
Described handling of non-normal and categorical variables | |
Yes | 2 (20.0) |
No | 8 (80.0) |
Unclear | 0 (0.0) |
Evaluated multiple imputation analysis | |
Yes | 1 (100) |
No | 8 (80.0) |
Unclear | 1 (10.0) |