TABLE 3—

Regression Table for Interrupted Time Series Model With State-Level Fixed Effects for Number of Prescriptions: United States, 2011–2016

	Opioids, b (95% CI)	Buprenorphine, b (95% CI)	Methadone, b (95% CI)	Naltrexone, b (95% CI)
Post (1 = postexpansion; 0 = preexpansion)	7 376.54 (2 769.77, 11 983.3)	−105.74 (−484.92, 273.43)	131.42 (26.78, 236.06)	25.49 (−3.01, 54.00)
Post × expanded	−2 090.27 (−7 884.19, 3 703.65)	−107.95 (−366.44, 582.34)	−78.90 (−206.34, 48.53)	−18.75 (−54.21, 16.70)
Time (quarters)	216.55 (−193.69, 626.79)	43.47 (8.98, 77.95)	−9.51 (−18.83, −0.19)	0.59 (−1.93, 3.12)
Time × post	−664.25 (−1 331.90, 3.40)	−32.54 (−87.52, 22.43)	−4.28 (−19.46, 10.90)	0.63 (−3.48, 4.75)
Time × expanded	−154.61 (−643.18, 333.96)	28.84 (−12.45, 70.13)	−0.02 (−11.05, 11.08)	0.74 (−2.24, 3.73)
Time × expanded × post	372.33 (−463.47, 1 208.13)	96.49 (27.74, 165.24)	3.66 (−14.79, 22.11)	8.56 (3.47, 13.64)
(Constant)	12 967.47 (11 268.62, 14 666.32)	1 135.77 (995.69, 1 275.85)	277.47 (240.71, 314.24)	55.62 (45.23, 66.02)

Note. CI = confidence interval. All prescribing rates are number of prescriptions reimbursed by Medicaid per 100 000 Medicaid enrollees.