. Author manuscript; available in PMC: 2023 Jan 1.

Published in final edited form as: Int J Drug Policy. 2021 Sep 26;99:103449. doi: 10.1016/j.drugpo.2021.103449

Table 4.

Estimated odds ratios (OR) for Rx overdose associated with medical cannabis law (MCL) status stratified by sampling frame, and test of differences in association by sampling frame, NHIS participants 2000–2011. Region specific estimates at bottom.

	OR (p-value) 2000–2005	OR (p-value) 2006–2011	Ratio of OR’s 2006–2011 vs. 2000–2005
MCL vs. No MCL
crude model (year only)^a	1.43 (1.06–1.93)	1.09 (0.90–1.32)	0.76 (0.54–1.09)
state-level adjusted model^b	1.21 (0.82–1.78)	0.83 (0.58–1.20)	0.70 (0.49–1.00)
individual-level adjusted model^c	1.41 (1.04–1.92)	1.03 (0.85–1.25)	0.76 (0.53–1.08)
adjusted at both levels^d	1.22 (0.83–1.79)	0.87 (0.60–1.25)	0.71 (0.49–1.01)

crude model adjusts for analytical year only (the calendar match of cases and controls)

state-level model adjusts for census region and presence of “proactive” prescription monitoring programs.

individual-level model adjusted for age, race/ethnicity, sex, marital status, education, and poverty status.

adjusted at both levels includes all individual and state-level adjustments