Skip to main content

Table 5 Supporting explanations about statistics used that was related to attrition bias

From: Assessments of attrition bias in Cochrane systematic reviews are highly inconsistent and thus hindering trial comparability

Statistical information

N (%)

ITT

826 (8)

No ITT

238 (2.3)

PP

88 (0.9)

ITT, LOCF

87 (0.8)

LOCF

67 (0.7)

ITT not reported

47 (0.5)

ITT, PP

34 (0.3)

Completer analysis

27 (0.2)

mITT

25 (0.2)

Sensitivity analysis

15 (0.1)

BOCF

12 (0.1)

ITT, BOCF

8 (0.08)

Analysis not described

6 (0.06)

Available case analysis

5 (0.05)

ITT, Completer analysis

5 (0.05)

LOCF, BOCF

5 (0.05)

ITT analysis may have been of value

4 (0.04)

ITT, PP, LOCF

4 (0.04)

ITT, LOCF, WOCF

4 (0.04)

LOCF, PP

4 (0.04)

Partial ITT

4 (0.04)

WOCF

3 (0.03)

Unclear whether LOCF was used

3 (0.03)

ITT inadequate

3 (0.03)

Some participants were excluded from analysis

3 (0.03)

No ITT, PP

3 (0.03)

BOCF, WOCF

2 (0.02)

ITT, LOCF, NRI

2 (0.02)

No LOCF

2 (0.02)

We have not been able to re-analyse the outcomes for all of the enrolled infants (ITT)

1 (0.01)

LOCF, Sensitivity analysis

1 (0.01)

ITT, PP, LOCF, Sensitivity analysis

1 (0.01)

The trial states that the analysis was performed on an ITT basis, but the data seems to have been analysed on-treatment

1 (0.01)

ITT analysis possible

1 (0.01)

ITT analysis conducted but unclear how missing data were dealt with

1 (0.01)

PP, FAS

1 (0.01)

It is likely that the principle of ITT analysis was violated

1 (0.01)

Statistical analysis used the APT

1 (0.01)

Missing outcome data imputed in analysis

1 (0.01)

True ITT analysis was difficult

1 (0.01)

Missing participants were omitted from the analysis

1 (0.01)

Although the study was set up to be analysed on ITT basis, the participants with missing outcomes were not included in the primary analysis

1 (0.01)

ITT done only for P value

1 (0.01)

Not strict ITT analysis

1 (0.01)

mITT, but unclear how missing data were dealt with

1 (0.01)

ITT, WOCF

1 (0.01)

mITT, LOCF

1 (0.01)

mITT, PP

1 (0.01)

Equal distribution among groups, ITT analysis not necessary

1 (0.01)

It was unclear if data analysis was PP or ITT

1 (0.01)

The results are presented as available case analysis rather than ITT. The authors present a sensitivity analysis

1 (0.01)

No information about whether an ITT analysis was undertaken and, if so, how missing data were imputed

1 (0.01)

This is an “as treated” as opposed to an ITT analysis

1 (0.01)

LOCF, BOCF, SOCF

1 (0.01)

ITT, PP, mITT

1 (0.01)

ITT, No sensitivity analysis

1 (0.01)

LOCF, Completer analysis

1 (0.01)

Large number of cross-overs made ITT impossible after the first phase

1 (0.01)

Unclear if ITT

1 (0.01)

ITT, PP, Sensitivity analysis

1 (0.01)

No ITT, Completer analysis

1 (0.01)

No mention of how missing data from participants who dropped out were dealt with, e.g. ITT analysis

1 (0.01)

ITT, Sensitivity analysis

1 (0.01)

No sensitivity analysis

1 (0.01)

LOCF, WOCF

1 (0.01)

  1. Abbreviations: ITT intention-to-treat analysis, PP per protocol analysis, LOCF last observation carried forward, mITT modified intention-to-treat analysis, BOCF baseline observation carried forward, WOCF worst observation carried forward, NRI non-responder imputation, FAS full analysis set, APT all patients treated, SOCF screening observation carried forward