Skip to main content

Table 3 Complete case and multiple imputed model compositions for the outcome measure persistent shoulder pain intensity

From: The search for stable prognostic models in multiple imputed data sets

Ā 

missing values

CCA

MI-5

Ā Ā 

rank

rank

persistent shoulder pain intensity*

76 (12.9%)

Ā Ā 

sporting injury

0 (0%)

1

1

concomitant lower back pain

0 (0%)

2

3

longer duration of complaints

1 (0.2%)

3

2

both shoulders afflicted

0 (0%)

4

4

inability to perform daily activities

8 (1.4%)

5

5

concomitant upper extremity pain

0 (0%)

6

6

sporting activities

0 (0%)

7

Ā 

higher physical workload

0 (0%)

8

Ā 
  1. CCAĀ Ā Ā Ā Ā Ā - complete case analysis
  2. MI -5Ā Ā Ā Ā Ā - multiple imputation using 5 imputation files
  3. rankĀ Ā Ā Ā Ā Ā - the order of appearance of predictors in the derived model arranged by their predictive ability (regression coefficient)
  4. %Ā Ā Ā Ā Ā Ā Ā Ā Ā Ā - inclusion frequency; the proportion of times that a variable with a univariable association with the outcome is retained in the automated backward selected models. When a variable was selected in each of the replications, the inclusion frequency was 100%
  5. *Ā Ā Ā Ā Ā Ā Ā Ā Ā Ā Ā - outcome measure