Skip to main content

Table 1 Models with the best penalty value and the penalty within 3% of the optimum

From: Validation of prediction models based on lasso regression with multiply imputed data

Covariates   Best model parameters   3% tolerance model parameters
  Number Uncalibrated β Re-calibrated β Number Uncalibrated β Re-calibrated β
Intercept 10 1.038638 0.880085 10 1.30578 0.862149
Anxiety medication 10 -0.04399 -0.04534 1 -0.00066 -0.00075
Cardiovascular medication 10 -0.01752 -0.01774 0   
Pulmonary medication 10 -0.01454 -0.01486 0   
Use of long acting beta agonists 3 -0.00143 -0.00129 0   
Influenza vaccination 10 -0.04283 -0.04365 0   
Renal:chronic kidney disease 10 -0.06797 -0.06939 0   
Lung function 6 0.000162 0.00018 0   
HADS depression score 8 -0.00372 -0.00403 7 -0.00274 -0.00311
Physical activity score 10 0.02788 0.028764 6 0.005337 0.006068
Self-efficacy 2 3 0.01409 0.014715 3 0.006139 0.006978
Self-efficacy 3 1 0.000367 0.000384 0   
Sit to stand test 10 0.001774 0.001901 10 0.001518 0.001632
crqdyspnea 10 0.654104 0.682812 10 0.60923 0.690902
crqfatigue 10 0.109821 0.114304 10 0.089453 0.100929
crqmastery 6 0.010376 0.010272 1 0.001909 0.002171
crqemotional 3 0.00078 0.00089 2 0.00134 0.001523
Feeling thermometer change score 10 -0.04587 -0.04658 0   
HADS depression change score 10 0.039783 0.040061 0   
crqdyspnea change score 10 -0.1582 -0.16403 10 -0.04866 -0.05207
  1. The column “number” gives the number of times each variable was selected across the imputed data sets.
\