Skip to main content

Table 5 Complete case bootstrap model selection results for the outcome measure persistent pain intensity

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

  most frequently selected models rank
Predictors 1 2 3 4 5 B CCA
longer duration of complaints X X X X X 1 3
concomitant lower back pain X X X X X 2 2
both shoulders afflicted X X - X X 3 4
concomitant upper extremity pain X X X - - 4 6
shoulder complaints in the past year X - X X - 5 -
sporting injury*       - 1
inability to perform daily activities*       - 5
sporting activities*       - 7
higher physical workload*       - 8
Count 120 96 58 47 37   
% 24.0 19.2 11.6 9.4 7.4   
  1. *             - predictors that appeared in ≥ 40% in the first step of the of the bootstrap model selection are not used in the second step in model selection
  2. rank        - the order of appearance of predictors in the derived models arranged by their predictive ability (regression coefficient estimates)
  3. B            - the complete case date based bootstrap selected model (i.e. the most frequently occurring combination of predictors in 500 replicate data sets of the second bootstrap model selection step)
  4. CCA        - the complete case data based model derived without additional bootstrap did not occur in the bootstrap model selection since some of the included predictors occurred ≥ 40% in the first selection step
  5. Count     - the number of times the model was selected in the 500 replicate data sets of the second bootstrap model selection step