Covariate patterns that do not fit or have considerable influence on estimated coefficient.

<p>Notations: </p><p></p><p></p><p></p><p><mi>π</mi></p><mo>^</mo><p></p><p></p><p></p>, estimated probability of death; Δχ², change in Pearson Chi-square; ΔD, change in deviance; Δ<p></p><p></p><p></p><p><mi>β</mi></p><mo>^</mo><p></p><p></p><p></p>, Cook’s distance.<p></p><p>† The number of covariate pattern is equivalent to the number of observations as there are several continuous covariates. The number itself was meaningless in itself but simply reflected the sequence of enrollment in the dataset.</p><p>Covariate patterns that do not fit or have considerable influence on estimated coefficient.