. Author manuscript; available in PMC: 2016 Nov 3.

Published in final edited form as: Crit Care Med. 2015 Oct;43(10):2141–2146. doi: 10.1097/CCM.0000000000001149

Table 2.

Sensitivity analyses

Type of Analysis	N	Adjusted OR (95% CI)	p-value
Primary propensity-matched analysis	38,788	1.08 (1.02–1.14)	0.004
GEE^a adjusted by propensity score	61,122	1.05 (0.99 –1.11)	0.1
Inverse probability of treatment weights	61,122	1.07 (1.00 – 1.13)	0.03
Stratified by propensity score deciles	38,788	1.06 (1.02 – 1.11)	0.001
Multivariable-adjusted GEE^a	61,122	1.00 (0.95 – 1.06)	0.88
Instrumental variable (per 10% incremental increase in hospital use of dopamine)	61,122	1.05 (1.04 – 1.07)	< 0.001

GEE: generalized estimating equation logistic regression analysis with hospital clusters