2007
DOI: 10.1111/j.1360-0443.2007.01946.x
|View full text |Cite
|
Sign up to set email alerts
|

Analysis of binary outcomes with missing data: missing = smoking, last observation carried forward, and a little multiple imputation

Abstract: The significance of the group effect did vary as a function of the assumed relationship between missingness and smoking. The 'conservative' missing = smoking assumption suggested a beneficial group effect on smoking cessation, which was confirmed via a sensitivity analysis only if an extreme odds ratio of 5 between missingness and smoking was assumed. This type of sensitivity analysis is crucial in determining the role that missing data play in arriving at a study's conclusions.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

4
169
1
1

Year Published

2011
2011
2024
2024

Publication Types

Select...
9

Relationship

0
9

Authors

Journals

citations
Cited by 146 publications
(175 citation statements)
references
References 25 publications
4
169
1
1
Order By: Relevance
“…Third, although a key advantage of the mixed-effects modeling approach is that it can be applied when participants are not measured at the same number of time points, some fraction of participants might be missing because they were not remitted. We repeated the analyses presented in Tables 3 and 5 with missing outcomes recoded as nonremission (43) and found similar results, which suggests the robustness of the findings.…”
Section: Discussionsupporting
confidence: 56%
“…Third, although a key advantage of the mixed-effects modeling approach is that it can be applied when participants are not measured at the same number of time points, some fraction of participants might be missing because they were not remitted. We repeated the analyses presented in Tables 3 and 5 with missing outcomes recoded as nonremission (43) and found similar results, which suggests the robustness of the findings.…”
Section: Discussionsupporting
confidence: 56%
“…Despite the conservative nature of this assumption, the assumption that loss to follow up meant participants were still smoking only contradicted one finding from the primary analyses (i.e., those achieving a reduction of greater than 50% at week 16). It thus appears that this assumption has the potential to under-estimate the beneficial effects of the intervention [30][31][32] , and although this approach has been widely advocated, more research into dealing with missing data in smoking trials is justified.…”
Section: Discussionmentioning
confidence: 99%
“…multiple imputation of missing data), which may provide more reliable estimates of treatment effects [30][31][32] .…”
Section: Downloaded Frommentioning
confidence: 99%
“…For the primary outcome, to assess the influence of the assumption that missing data equals 'smoking' on the effect size, we used the Hedeker method to test various scenarios of the association between smoking and having missing data. 84 Other outcomes for smoking cessation were analysed in a similar way.…”
Section: Sample Sizementioning
confidence: 99%