Effect of Old Treatment and Age on New Treatment, 35 Patients, Traditional Versus Regularized Regressions

The effect of old treatment and age on new treatment was assessed in a 35 patient crossover study, applying Traditional and regularized regression analysis. The worst result was obtained by the standardized Traditional regression with a poor correlation between predictor and outcome three times. The best results with 3 strong correlations were obtained by the Lasso regression. The Elastic Net produced 2 strong correlations, and 1 reasonably significant one. The unstandardized Traditional regression produced one significant p-value (0.048), the factor of which showed a poor correlation in the standardized Traditional regression.