. 2014 Oct 23;4:26. doi: 10.1186/s13561-014-0026-2

Table 2.

Identifying the algorithm

Dependent variable: lnADMINCOST
	Model A	Model B	Model C	Model D
Independent variables	$\hat{β_{k}}$ (SE)	$\hat{β_{k}}$ (SE)	$\hat{β_{k}}$ (SE)	$\hat{β_{k}}$ (SE)
SUBCUTANEOUS	-3.2858 (0.57)^***	-2.6161 (1.106)^**	-3.2073 (0.902)^***	—
INTRAMUSCULAR	-2.4097 (0.90)^**	-2.2784 (1.219)^*	-5.1856 (1.335)^***	—
DOSFREQ	-0.0046 (0.011)	-0.0025 (0.012)	0.4206 (0.18)^**	-0.1022 (0.179)
PRODUCTBUND	—	-0.7395 (1.129)	0.4133 (0.928)	-2.7416 (0.678)^***
INDICATN	—	-1.0527 (1.219)	-0.2084 (1.467)	-3.7195 (1.693)^**
DOSFREQ2	—	—	-0.00105 (0.0003)^***	-0.00075 (0.0004)^*
DOSFREQ.INDICATN	—	—	-0.3126 (0.172)^*	0.17 (0.183)
Intercept	8.6223 (0.45)^***	9.6888 (1.274)^***	7.1058 (1.521)^***	11.305 (1.685)^***
Regression statistics
R² (Adjusted R²)	0.7188 (0.6586)	0.7376 (0.627)	0.8881 (0.8098)	0.6957 (0.5689)
F-statistic (>c)	11.9313 (>3.1122)	6.7151 (>2.9961)	11.3369 (>3.0717)	5.4864 (>2.9961)
SSR (N)	17.8023 (18)	16.6714 (18)	7.0858 (18)	19.2688 (18)
SE( $ln ADM \hat{IN} COST$ )	1.1276	1.1787	0.8418	1.2672
F-ratio (>c)	3.781 (<3.478)	6.764 (>4.1028)	—	10.3161 (>4.1028)

Notes: ${\hat{β}}_{k}$ = regression coefficients; ***p < 0.01; **p < 0.05; *p < 0.10; SE = default standard error; F-statistic = F-test for overall significance of the regression; c = critical value for F-test; SSR = residual sum of squares; N = number of observations; F-ratio = F-test for comparison with model C; c = critical value for the F-test; SUBCUTANEOUS = dummy variable for subcutaneous delivery; INTRAMUSCULAR = dummy variable for intramuscular delivery; DOSFREQ = dosing frequency; PRODUCTBUND = dummy indicating whether the drug product is sold as a bundled product with some of the equipment and consumables used in drug administration; INDICATN = dummy for type of illness a product is clinically indicated for.