Table 2.

Estimates of the Association between Inpatient Spending ($10,000s) and Thirty-Day Mortality of Medical Patients

	(1)-OLS	(2)-OLS	(3)-OLS	(4) – IV	(5) – IV	(6) – IV	(7) – IV
CHF patients	0.0034*	0.0025*	0.0032*	−0.0212*	−0.0255*	−0.0222*	−0.0302*
(mean of dep. var. = 0.08)	(0.0002)	(0.0002)	(0.0002)	(0.0037)	(0.0042)	(0.0034)	(0.0087)
AMI patients	−0.0016*	−0.0019*	−0.0022*	−0.0003	0.0084	−0.0086*	−0.0132*
(mean of dep. var. = 0.15)	(0.0001)	(0.0001)	(0.0001)	(0.0045)	(0.0055)	(0.0028)	(0.0043)
Stroke patients	0.0018*	−0.0001	−0.0004**	−0.0549*	−0.0468*	−0.0516*	−0.0399*
(mean of dep. var. = 0.17)	(0.0002)	(0.0002)	(0.0002)	(0.0084)	(0.0109)	(0.0072)	(0.0157)
GI-bleeding patients	0.0076*	0.0050*	0.0057*	−0.0116*	−0.0117*	−0.0137*	−0.0274*
(mean of dep. var. = 0.06)	(0.0004)	(0.0003)	(0.0003)	(0.0026)	(0.0042)	(0.0027)	(0.0087)
End-of-life instruments:
ICU days				Excluded	Excluded		Included
				Instrument	Instrument		Covariate
Non-ICU hospital days				Excluded	Included		Included
				Instrument	Covariate		Covariate
Inpatient spending (Dartmouth)						Excluded	Excluded
						Instrument	Instrument
Partial F-excluded instruments
CHF patients				20*	35*	66*	21*
AMI patients				13*	15*	35*	23*
Stroke patients				21*	32*	73*	13*
GI-bleeding patients				26*	33*	96*	24*
Patient characteristics	No	Yes	Yes	Yes	Yes	Yes	Yes
Hospital characteristics	No	No	Yes	Yes	Yes	Yes	Yes

Notes: See notes to table 1. In addition, for medical patients, we include among the hospital characteristics separate measures of the average total spending on three types of surgery patients who did not experience a complication: general, vascular, and orthopedic.