TABLE 5.

Relative Difference in Change Over Time in Health Utilization Between SDM Groups^a

	Hospitalizationsb (95% CI)	P Valuec	ED Visitsb (95% CI)	P Valuec	Office-based Visitsd (95% CI)	P Valuec	Prescriptionsd (95% CI)	P Valuec
SDM increasing vs unchanged high	−4.4 (−8.7, −0.1)	.04	−10.3 (−18.6, −1.9)	.02	−1.3 (−2.3, −0.3)	.01	−0.8 (−1.8, 0.2)	.1
vs unchanged low	−5.1 (−10.8, 0.7)	.08	−2.9 (−14.5, 8.6)	.6	−2.2 (−4.2, −0.2)	.03	−0.8 (−1.8, 0.3)	.1
vs decreasing	−1.1 (−8.6, 6.3)	.8	−6.5 (−18.6, 5.6)	.3	−1.7 (−3.2, −0.3)	.02	−0.4 (−1.6, 0.8)	.5
SDM unchanged high vs unchanged low	0.6 (−4.4, 5.7)	.8	−7.3 (−17.8, 3.1)	.2	0.9 (−1.7, 2.8)	.4	−0.1 (−1.0, 1.0)	.9
vs decreasing	3.3 (−3.4, 10.0)	.3	3.7 (−7.5, 15.0)	.5	−0.4 (−1.7, 0.9)	.5	0.4 (−0.8, 1.5)	.5
SDM unchanged low vs decreasing	3.9 (−3.6, 11.5)	.3	−3.6 (−17.4, 10.2)	.6	0.5 (−1.7, 2.7)	.7	0.3 (−0.9, 1.5)	.6

^aSince the inclusion of covariates did not change the association of SDM with utilization, unadjusted results are shown.

^bRates are per 100 patients per year.

^cP values calculated from linear regression with robust variance estimation accounting for the MEPS complex survey design.

^dRates are per patient per year.