Table 2.

Results of multinomial and zero-inflated negative binomial regression models predicting recency of cannabis vaping, or number of puffs during most recent cannabis vaping, for US adults who reported ever vaping cannabis in the Population Assessment of Tobacco and Health (PATH) Study, 2018-2019 (n=5,331)

	Outcome 1: Recency of Cannabis Vaping (Relative to “No Use”)			Outcome 2: Number of Puffs
	≥3 days ago RRR (95% CI)	1–2 days ago RRR (95% CI)	Today RRR (95% CI)	No Puffs (Logit Model) β (95% CI)	Number of Puffs (Count Model) β (95% CI)
Unadjusted model
Sexual Minority (heterosexual, ref.)	2.39*** (1.69–3.38)	2.50*** (1.67–3.74)	2.71*** (1.86–3.97)	−0.42* (−0.78 – −0.06)	0.76 (0.55 – 1.06)
Adjusted model
Sexual Minority (heterosexual, ref.)	1.85** (1.27–2.67)	2.00** (1.28–3.11)	2.36*** (1.55–3.60)	−0.36 (−1.07–0.34)	0.86 (0.66–1.13)

Notes. Sample of 5,331 participants without missing data on any variable in the analyses. Weighted results. Unadjusted models regress the outcome on the exposure variable (self-identification as a sexual minority); adjusted models add race, age, marital status, sex, education, income, and homeownership.

Abbreviations. RRR = relative-risk ratio; CI = confidence interval; ref. = reference category

^*p < .05.

^**p < .01.

^***p < .001