Association of Rates of Smoking During Pregnancy With Corporate Tobacco Sales Policies

Maternal smoking during pregnancy (SDP) is a leading preventable cause of adverse birth outcomes in the United States, increasing rates of miscarriage, preterm birth, low birth weight, Sudden Infant Death Syndrome, and poor lung function.¹ Despite a 28% reduction in SDP rates from 2011 to 2016,² tremendous disparity remains among vulnerable populations. The SDP rate is higher among pregnant women who are younger, less educated, socioeconomically disadvantaged, and residents of rural areas, compared with other expectant mothers.^1-3

Tobacco retailer density (TRD) has been linked to high smoking rates, increased cigarette intake, low life expectancy, and high mortality.^4,5 The number of tobacco retailers in the southeast United States increased by approximately 8300 in late 2012 and early 2013, when Family Dollar and Dollar General, the 2 largest dollar-store chains in the country, started selling tobacco products to meet customer demand. The number of tobacco retailers in the southeast decreased in 2014 by approximately 2500 when CVS, the largest pharmacy chain, discontinued tobacco sales. The fluctuation in TRD has affected communities differently, as CVS pharmacies are more likely to be located in urban areas, whereas dollar stores are more likely to operate in poor and rural communities.⁶

In this study, we examined the association between change in SDP and corporate-policy change in TRD across 6 southeastern states, controlling for rurality, poverty, social vulnerability, and tobacco-control policies. All 6 states included in this study—Florida, Georgia, Mississippi, North Carolina, South Carolina, and Tennessee—declined to adopt the Affordable Care Act Medicaid expansion in 2014, limiting access to smoking cessation and contraception programs among these states’ low-income populations.

This retrospective study used open outcomes data aggregated at the county level, which are published and made available online by the state health departments of Florida, Georgia, Mississippi, North Carolina, South Carolina, and Tennessee. As such, no institutional review board approval and informed consent were required.

Annual county-level birth and self-reported SDP counts (from birth registries) between January 1, 2011, and December 31, 2016, were obtained from public-facing websites of state health departments. Addresses of CVS pharmacies and dollar stores in the region (as of January 1, 2018) were identified from corporate websites. County-level TRD change resulting from corporate policy was calculated as the number of dollar stores minus the number of CVS pharmacies per 10 000 adults. Percentage change in SDP was calculated for 2011 to 2012 and for 2015 to 2016 (before and after the corporate policy changes). We examined the association between percentage change in SDP and TRD change at the county level using a mixed-effects regression model. Control variables included socioeconomic vulnerability (Socioeconomic Status Index, 1 of the 4 components of the Centers for Disease Control and Prevention Social Vulnerability Index), rurality, existence of smoke-free law, and state.

With analysis of variance, we compared county-level percentage change in SDP across categories of rurality and socioeconomic vulnerability (below and above the southeast US median) and in states with high or low tobacco-control funding. All analyses were conducted with SAS, version 9.4 (IBM), and 2-sided P < .05 was used to indicate statistical significance.

Across the 6-state region, SDP decreased 15.6% between 2011 to 2012 and 2015 to 2016. The contrasting policy changes by the dollar stores and by CVS yielded an overall TRD increase in the 6 states of 1.0 additional store per 10 000 adults. The TRD change was 2.7 more stores per 10 000 adults in counties with high TRD increase and was 0.6 additional stores per 10 000 adults in counties with low TRD increase. Difference in percentage change in SDP between counties experiencing high TRD change and those experiencing low TRD change was statistically significant (–5.0 vs –19.0; P < .001) as well as across all rurality categories and socioeconomic vulnerability categories (Table 1). In the mixed-effects model, county-level TRD increase (t = 2.81; P = .005) as well as socioeconomic vulnerability (t = 1.89; P = .06) and rurality (t = –2.69; P = .007) were the most important indicators of the magnitude of SDP decrease, whereas tobacco-control measures did not seem associated with improvement in SDP during the study period (Table 2).

In this study, we identified an association of SDP with increased TRD after corporate sales decisions. The decision by CVS to discontinue selling tobacco was followed by a reduction of TRD in some counties, but the initiation of tobacco sales by dollar stores has increased access to tobacco products across the region. We examined several of the largest contributing factors to SDP and found that high TRD may be inhibiting progress in smoking reduction across the southeast, regardless of rurality, socioeconomic vulnerability, and ongoing tobacco-control efforts. Localities could look into licensing and zoning as lasting strategies to reduce TRD, and states should adequately fund smoking-cessation and health communications interventions with a mass reach.

Accepted for Publication: September 25, 2018.

Corresponding Author: Jaclyn Hall, PhD, Department of Health Outcomes and Biomedical Informatics, College of Medicine, University of Florida, PO Box 100177, Gainesville, FL 32610-0177 (jaclynha@ufl.edu).

Published Online: January 14, 2019. doi:10.1001/jamapediatrics.2018.4598

Author Contributions: Dr Hall had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hall, Guo, Maldonado-Molina, Thompson, Salloum.

Acquisition, analysis, or interpretation of data: Hall, Cho, Guo, Maldonado-Molina, Shenkman, Salloum.

Drafting of the manuscript: Hall, Cho, Salloum.

Critical revision of the manuscript for important intellectual content: Hall, Guo, Maldonado-Molina, Thompson, Shenkman, Salloum.

Statistical analysis: Hall, Cho, Guo.

Obtained funding: Shenkman.

Administrative, technical, or material support: Hall, Thompson, Shenkman.

Supervision: Hall, Guo, Shenkman, Salloum.

Conflict of Interest Disclosures: None reported.

Additional Contributions: Diana Tonnessen, MS, and Debra McDonald, MA, University of Florida, provided communications assistance. These individuals received no compensation for their contribution.