Table 4.

PCS Model Corrected For Response Bias (N = 1605)

		95% Confidence Interval
Variables	Coefficient	Lower Limit	Upper Limit
Mean patient age (per 1 year increment)	-0.07	-0.11	-0.03
Male gender	-1.00	-4.05	2.04
Caucasian race	-0.67	-1.62	0.28
Hx of MI	-0.89	-2.39	0.62
Hx of Chronic heart failure	-2.95	-4.22	-1.69
Hx of Coronary artery bypass graft surgery	-3.16	-4.23	-2.09
Hx of Percutaneous coronary intervention	-1.24	-2.75	0.28
Substance or alcohol abuse	-0.17	-1.85	1.52
Current smoker	-0.48	-1.48	0.51
Dementia	-0.77	-2.89	1.34
COPD	-3.91	-4.91	-2.91
Stroke	-2.92	-4.14	-1.70
Depression	-4.68	-6.06	-3.31
Arthritis	-1.71	-2.75	-0.67
Diabetes mellitus	-2.33	-3.45	-1.20
Coronary revascularization during index admission	1.76	0.32	3.19
Cardiogenic shock	1.44	-3.72	6.60
Hypotensive episode	0.48	-1.50	2.46
Positive stress test	1.06	-0.38	2.50
Admitted to Tertiary VA	-0.78	-1.81	0.25
ST-segment elevation MI on ECG	2.41	0.97	3.85
Mean serum creatinine (per 1 mg/dl increment)	-0.66	-1.11	-0.22
Left bundle branch block on ECG	0.05	-2.30	2.39
Cardiac catheterization during index admission	0.60	-0.41	1.60
Discharge diagnosis Unstable Angina (vs. MI)	-0.62	-1.42	0.19
Inverse Mills Ratio *	-2.15	-5.66	1.37

* Inverse Mills Ratio: Variable derived from the Heckman Selection Equation; MI: myocardial infarction; Hx: History; COPD: chronic obstructive pulmonary disease; ECG: electrocardiogram