Effects of fee-for-service, diagnosis-related-group, and mixed payment systems on physicians’ medical service behavior: experimental evidence

Table 8 REM regression of the loss of patient benefit

Independent variable	DRG			FFS
Independent variable	(1)	(2)	(3)	(4)	(5)	(6)
Payment: DRG/FFS (ref)
Mix-more-DRG(2)	−0.053***	−0.053***	−0.053***
Mix-more-DRG(2)	(0.009)	(0.009)	(0.009)
Mix-more-DRG(4)	−0.118***	− 0.118***	− 0.118***
Mix-more-DRG(4)	(0.017)	(0.017)	(0.017)
Mix-more-DRG(6)	−0.088***	− 0.088***	− 0.088***
Mix-more-DRG(6)	(0.011)	(0.011)	(0.011)
Mix-more-FFS(8)				−0.019**	−0.019**	−0.019**
Mix-more-FFS(8)				(0.006)	(0.006)	(0.006)
Mix-more-FFS(6)				−0.047***	−0.047***	−0.047***
Mix-more-FFS(6)				(0.012)	(0.012)	(0.012)
Type of illness: illness A (ref)
Illness B		−0.039***	−0.039***		−0.013**	−0.013**
Illness B		(0.004)	(0.004)		(0.005)	(0.005)
Illness C		−0.013**	−0.013**		0.025***	0.025***
Illness C		(0.004)	(0.004)		(0.007)	(0.007)
Severity of illness: Moderate (ref)
Intermediate		0.016*	0.016*		−0.057***	−0.057***
Intermediate		(0.007)	(0.007)		(0.009)	(0.009)
Severe		0.045***	0.045***		−0.100***	−0.100***
Severe		(0.013)	(0.013)		(0.013)	(0.013)
Age			−0.010			0.003
Age			(0.006)			(0.006)
Gender (female for ref)			0.029			−0.027
Gender (female for ref)			(0.020)			(0.020)
Education (undergraduats for ref)			0.004			−0.014
Education (undergraduats for ref)			(0.027)			(0.033)
Constant	0.168***	0.165***	0.377**	0.128***	0.176***	0.121
Constant	(0.012)	(0.010)	(0.127)	(0.012)	(0.018)	(0.136)
Observations	8100	8100	8100	5400	5400	5400
Subjects	90	90	90	60	60	60
R²	0.096	0.126	0.158	0.017	0.128	0.136

This table shows results from random effects model. The dependent variable is the loss of patient benefit. Panel A (columns 1 to 3) and Panel B (columns 4 to 6) show regression results for DRG-based schemes and FFS-based schemes, respectively. The reference category is DRG in panel A and FFS in panel B. ‘Mix-more-DRG(2)’, ‘Mix-more-DRG(4)’, ‘Mix-more-DRG(6)’, ‘Mix-more-FFS(8)’ and ‘Mix-more-FFS(6)’ are dummy variables for the mixed payment schemes. Additionally, we control for the type of illness and severity of illness with illness ‘A’ and severity of illness ‘moderate’ being the reference categories. The variable’ demographics comprise age, gender and education. Robust standard errors, in parentheses below the coefficients, are clustered at individual subject. *** p < 0.001, ** p < 0.01, * p < 0.05
DRG Diagnosis-Related-Group, FFS Fee-for-Service

ISSN: 1472-6963