Table 4 Sensitivity analyses: comparison of odds ratio of 30-day mortality from different PS methods and covariate adjustment.

From: Propensity-score-matched evaluation of under-recognition of acute kidney injury and short-term outcomes

Under-recognition vs. Timely-recognition

Odds Ratio

95%CI

p-value

PS Matchinga (Crude)

1.04

0.84–1.29

0.711

PS Matchinga (covariate adjustmentb)

1.01

0.62–1.64

0.967

Logistic Regression (Crude)

0.36

0.28–0.45

<0.001

Logistic Regression (covariate adjustmentb)

0.89

0.60–1.32

0.575

Stratification 5 strata

0.93

0.68–1.28

0.551

PS as covariate

0.92

0.66–1.29

0.626

PS as covariate “doubly robust”b

0.90

0.60–1.35

0.602

IPTW

0.72

0.62–0.84

<0.001

IPTW “doubly robust”b

0.88

0.72–1.09

0.253

  1. aLogistic regression model using Generalized Estimated Equations; badjusted by 17 matched variables listed in Table 1. IPTW: Inverse Probability of Treatment Weighting; PS: propensity score.
Back to article page