Measurement and control of bias in patient reported outcomes using multidimensional item response theory

Table 4 Comparison of results from the Cox proportional hazards models with competing risk data

	Model using raw scores				Model not adjusted for ERS (Model 1)				Model adjusted for ERS (Model 2)
Predictor	Hazard Ratio	SE	p-value	95 % CI	Hazard Ratio	SE	p-value	95 % CI	Hazard Ratio	SE	p-value	95 % CI
Age	1.143	0.012	<0.001	[1.116,1.169]	1.141	0.008	<0.001	[1.125,1.162]	1.144	0.012	<0.001	[1.117,1.171]
Male	1.059	0.177	0.746	[0.749,1.498]	0.961	0.134	0.694	[0.734,1.229]	1.069	0.177	0.707	[0.755,1.513]
Education	1.040	0.022	0.067	[0.997,1.088]	1.042	0.022	0.064	[0.998,1.089]	0.956	0.132	0.734	[0.739,1.237]
Neuroticism	1.041	0.013	0.003	[1.014,1.069]	1.176	0.044	<0.001	[1.082,1.285]	1.215	0.044	<0.001	[1.115,1.324]

Notes. SE = Standard error; CI = Confidence interval; ERS = Extreme response style. Values that are statistically significant are indicated in bold

ISSN: 1471-2288