Disparities in cigarette smoking and the health of marginalized populations in the U.S.: a simulation analysis

Levy, Douglas E.; Lee, Stephanie S.; Qian, Yiqi; Shebl, Fatma M.; Goldberg, Sydney L.; Mulroy, Nora M.; Anderson, Nicola K.; Hyle, Emily P.; Becker, Jessica E.; Reddy, Krishna P.

doi:10.1186/s12889-025-22658-8

BMC Public Health

Table 1 Model input parameters

From: Disparities in cigarette smoking and the health of marginalized populations in the U.S.: a simulation analysis

Parameter	Value (Range in sensitivity analysis)				Source
PARAMETERS USED IN BOTH FIRST AND SECOND ANALYSES
	Low SES	High SES	SPD	Non-SPD
Examples of all-cause mortality (at age 60y) among people who never smoked, monthly probability (× 10^− 4)^a	Women: 9.0 (95% CI 8.4–9.6) Men: 12.7 (95% CI 11.7–13.9)	Women: 4.0 (95% CI 3.7–4.3) Men: 6.5 (95% CI 6.0-7.1)	Women: 10.5 (95% CI 8.6–12.8) Men: 16.3 (95% CI 12.3–21.6)	Women: 5.0 (95% CI 4.8–5.3) Men: 7.4 (95% CI 7.0-7.9)	NHIS^b
All-cause mortality hazard ratio, people who currently smoke vs. people who never smoked	Women: 2.4 (95% CI 2.3–2.5) Men: 2.1 (95% CI 2.0-2.3)				NHIS^b
PARAMETERS USED ONLY IN SECOND ANALYSIS
	Low SES	Higher SES	SPD	Non-SPD
Women/men, %	55/45	51/49	61/39	51/49	NHIS^c
Initial age, mean (SD), years	46.1 (18.2)	48.2 (18.4)	46.7 (11.1)	46.8 (11.7)	NHIS^c
Examples of initial prevalence of never/former/current smoking, by age, %					NHIS^c
Age 20–24y	Women: 79.8/6.4/13.8 Men: 80.2/9.2/10.6	Women: 89.3/2.8/7.9 Men: 81.6/9.4/9.0	Women: 75.1/6.6/18.2 Men: 53.4/22.7/23.9	Women: 85.7/5.1/9.1 Men: 82.4/6.3/11.3
Age 40–44y	Women: 65.4/14.9/19.8 Men: 51.0/17.5/31.5	Women: 68.3/20.0/11.7 Men: 56.9/28.3/14.8	Women: 42.7/22.0/35.3 Men: 32.3/26.8/40.9	Women: 71.6/16.4/11.9 Men: 59.9/23.2/16.9
Age 60–64y	Women: 50.3/17.8/31.9 Men: 44.1/24.0/31.9	Women: 61.4/27.3/11.3 Men: 49.0/35.7/15.3	Women: 47.1/31.6/21.3 Men: 36.0/26.1/37.9	Women: 61.6/26.4/12.0 Men: 53.7/31.1/15.2
Smoking initiation probability, monthly, by age and gender^d	0-0.007 (0.75x-1.25x)	0-0.006 (0.75x-1.25x)	0-0.008 (0.75x-1.25x)	0-0.006 (0.75x-1.25x)	NHIS and Jeon [31]
Smoking cessation probability, monthly, by age and gender^d	0-0.072 (0.75x-2.0x)	0-0.132 (0.75x-2.0x)	0-0.065 (0.75x-2.0x)	0-0.120 (0.75x-2.0x)	NHIS and Jeon [31]
Smoking relapse probability, monthly (t = months since cessation)^e	0.62e^− 0.33t (0.75x-1.25x)				[54,55,56,57]
Annual relative decrease in current smoking prevalence among 20-year-olds entering the cohort, %	-9.0	-12.1	-8.3	-10.5	NHIS^c
All-cause mortality hazard ratio, people who formerly smoked vs. people who never smoked, by age at smoking cessation					NHIS^b, Thun 2013 [29]
<40y^f	Women: 1 (assumption) Men: 1 (assumption)
40–44y	Women: 1.3 Men: 1.1
45–54y	Women: 1.4 Men: 1.3
55–64y	Women: 1.7 Men: 1.5
65-69y	Women: 1.7 Men: 1.5
≥70y	Women: 2.2 Men: 2.1

Abbreviations:
NHIS, National Health Interview Survey. SES, socioeconomic status. SPD, people with serious psychological distress. SD, standard deviation. CVD, cardiovascular disease. CI, confidence interval
Low SES was defined as having income < 100% of the federal poverty level, or having less than a high school education, or being a Medicaid beneficiary. SPD was defined as having a score of 13 or higher on the Kessler-6 scale
In the first analysis, all individuals in each modeled cohort are homogeneous by subpopulation (low SES, higher SES, SPD, or non-SPD), initial age (40 years), gender (women or men), and smoking status (current or never smoking)
In the second analysis, each modeled subpopulation cohort comprises individuals with a specific distribution of age, gender, and current/former/never smoking status, derived from NHIS
^aMonthly mortality probabilities for 60-year-old people are provided as an example here. The model includes age-specific monthly probabilities for a range of ages
^bThis was derived from NHIS 2004–2014
^cThis was derived from NHIS 2019 (Low SES and Higher SES) and pooled NHIS 2009–2014 (SPD and Non-SPD) data
^dFor the sensitivity analysis around smoking initiation and cessation probabilities, multipliers (e.g., multiply by 1.25) were applied to previously derived (from NHIS 1997–2009 data) initiation and cessation probabilities that were stratified by age and gender but not by subpopulation [21].
^eThis is based on relapse probabilities reported in smoking cessation intervention trials, focusing on placebo arms
^fWe assumed no excess mortality among people who stop smoking before age 40y compared with people who never smoke [22, 29].

Back to article page

ISSN: 1471-2458

Contact us

General enquiries: journalsubmissions@springernature.com