Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Mar 1.
Published in final edited form as: Prev Med. 2020 Jan 15;132:105994. doi: 10.1016/j.ypmed.2020.105994

Smoking Prevalence Among U.S. National Samples of Pregnant Women

Tyler D Nighbor 1,2, Sulamunn RM Coleman 1,2, Janice Y Bunn 4, Allison N Kurti 1,2,3, Ivori Zvorsky 1,3, Eva J Orr 1, Stephen T Higgins 1,2,3,*
PMCID: PMC7024639  NIHMSID: NIHMS1551535  PMID: 31952968

Abstract

Several data sources exist for estimating U.S. smoking prevalence among pregnant women, yet each differs in ways that have the potential to impact the estimates. In the present study we used the Population Assessment of Tobacco and Health (PATH), the National Survey on Drug use and Health (NSDUH), and the Pregnancy Risk Assessment Monitoring System (PRAMS), three common data sources, to evaluate the following questions about estimating U.S. smoking prevalence among pregnant women: To what extent are estimates impacted by differences in whether the samples include younger (< 18 years) or older (> 44 years) women, represent smoking in any trimester or only the 3rd, and use data from nationally representative or more selected national samples. Among the factors examined, inclusion of younger or older women does not appear to meaningfully alter prevalence estimates. Focusing on only the third trimester likely underestimates smoking prevalence, while the influence of basing estimates on selected national subgroups of women (i.e., only women who delivered live born infants) rather than nationally representative surveys has little discernible influence. Going forward, this research area would benefit from greater consistency in explicitly discussing the sampling methods used and how these various methods may have influenced the estimates reported.

Keywords: pregnancy, cigarette smoking, national sample, National Survey on Drug Use and Health, Population Assessment of Tobacco and Health, Pregnancy Risk Assessment Monitoring System

Introduction

Cigarette smoking during pregnancy is the leading preventable cause of poor pregnancy outcomes in the United States (U.S.) and other developed countries, increasing risk for numerous and sometimes catastrophic pregnancy complications as well as immediate and longer-term adverse health outcomes among exposed offspring (Cnattingius, 2004; Diez et al., 2010). These numerous and serious health risks make reducing smoking during pregnancy a U.S. public health priority (Centers for Disease Control and Prevention, CDC, https://www.healthypeople.gov). As such, monitoring progress in tobacco control and regulatory science efforts towards that goal necessitates having accurate estimates of U.S. smoking prevalence among pregnant women.

Multiple data sources exist for estimating U.S. smoking prevalence among pregnant women, which is perhaps not surprising considering the importance of the topic to public health. Estimates may be based on nationally representative surveys including, for example, the Population Assessment of Tobacco and Health (PATH), an annual, longitudinal-cohort survey of the U.S. non-institutionalized population aged ≥ 12 years (e.g., Kurti et al., 2017) and the National Survey on Drug use and Health (NSDUH), an annual, multi-year cross-sectional survey comprised of a U.S. national sample aged ≥ 12 years (e.g., Alshaarawy & Anthony, 2015). Another commonly used data source on this topic is the Pregnancy Risk Assessment Monitoring System (PRAMS), an ongoing surveillance project coordinated through the CDC and state health departments. PRAMS provides annual state-specific smoking prevalence estimates pre-pregnancy, during the 3rd trimester of pregnancy, and postpartum for 47 states, New York City, Puerto Rico, Washington, D.C., and Great Plains Tribal Chairmen’s Health Board (e.g., Shulman et al., 2018; Tong et al., 2013b). PRAMS data is obtained by each participating site through surveying stratified samples of women who delivered live infants approximately 2–6 months following delivery (see Shulman et al., 2018, for a review of PRAMS procedures).

Although not an exhaustive list, each of these data sources on U.S. smoking prevalence among pregnant women differs in ways that have the potential to substantially impact the estimates. For example, PATH and NSDUH surveys use different upper age limits for querying women on pregnancy status (49 and 44 years, respectively), and it is not uncommon for studies using those surveys to focus exclusively on adults, excluding women < 18 years from the estimates (e.g., Do et al., 2018; Kratz & Vaughn, 2012; Kurti et al., 2017). Additionally, estimates based on PRAMS are specific to the third trimester, include only women who delivered live infants, and are stratified at the state rather than national level.

The potential influence of different data collection methods on estimates of smoking prevalence during pregnancy have been reported previously (e.g., Dietz et al., 2010; Howland et al., 2015; Tong et al., 2013a). Dietz and colleagues (2010), for example, using multiple years of data (1999–2006) from the National Health and Nutrition Examination Survey, demonstrated that basing estimates exclusively on maternal self-report underestimated smoking prevalence among pregnant women by approximately 23%. Tong and colleagues compared prevalence estimates for 2008 in eight states using only data from the U.S. 2003 revised birth certificate (restricted to 3rd trimester so that it aligned with PRAMS), only PRAMS, or combined birth certificate and PRAMS. Prevalence estimates were 11.3% (95% CI: 10.3–12.4%), 14.0% (12.9–15.2%), and 15.2% (14.1–16.4%) for birth certificate only, PRAMS only, and combined birth certificate/PRAMS, respectively, demonstrating that reliance on birth certificates only significantly underestimated prevalence of smoking in the 3rd trimester.

To our knowledge, there are no reports in the literature explicitly examining whether and to what extent the common data-source and between-study differences regarding maternal age, trimester, and use of nationally representative surveys versus more select national samples discussed above impact prevalence estimates of smoking during pregnancy. Thus, the aim of the current study is to use data from PATH, NSDUH, and PRAMS to begin examining that knowledge gap.

Method

Data Sources

Table 1 summarizes key features of each data source. Data from the longitudinal PATH study were obtained from the restricted use files of the first, second, and third waves (N = 45,971). Briefly, PATH data are collected using in-person interviews and collect information on tobacco-use patterns, risk perceptions and attitudes towards current and newly emerging tobacco products, tobacco initiation, cessation, relapse behaviors, and health outcomes (see Hyland et al., 2017, for a detailed description of survey design and procedures). Data from Wave 1 (W1) were collected between September 2013 and December 2014 using address-based, area-probability sampling. Wave 2 (W2) data were collected between October 2014 and October 2015, and Wave 3 (W3) data were collected between October 2015 and October 2016. This report is limited to pregnant women 12–49 years who completed W1, W2, or W3 of the PATH study (n = 430, 388, and 327, respectively). Questions on pregnancy were asked only of women < 50 years old. For comparisons corresponding to inclusion of women from differing age ranges on the estimates of smoking prevalence, all three waves were collapsed into a single dataset, with women pregnant in multiple years being counted only during their first pregnancy in this data set. Weighting procedures adjusted for varying selection probabilities and differential non-response rates using a combination of census data and person-level data collected during the household screening interview, while appropriately accounting for the complex study design. The overall weighted response rate for W1 for adults was 74.0%, with a weighted retention rate of 83.1% at W2, and 78.4% at W3.

Table 1.

Key methodological features of the Population Assessment of Tobacco and Health (PATH), National Survey on Drug Use and Health (NSDUH), and Pregnancy Risk Assessment Monitoring System (PRAMS) surveys.

PATH NSDUH PRAMS
Research design Longitudinal Cross-sectional Cross-sectional
Representativeness U.S. general population U.S. general population State level
Data collection method In-person interview In-person interview Mailed questionnaire
Age restrictions on women queried on pregnancy status 12–49 12–44 None
Trimester(s) of pregnancy covered All All Third trimester only
Open in a new tab

Data from NSDUH were obtained from the 2014, 2015, and 2016 public-use files (N=169,314, Pregnant n = 758, 791, and 732, respectively), which approximate in time Waves 1–3 of PATH. NSDUH data are collected through in-person interviews and include assessments of tobacco and other substance use and related health issues (see Center for Behavioral Health Statistics & Quality, 2015, for a detailed description of survey design and procedures). The women included in this report were those 12–44 years. Questions on pregnancy were asked only of women < 45 years old. Participant weights were included with the survey data to obtain results representative of the U.S. population by correcting for selection probabilities, non-response, and post-stratification.

PRAMS data from 2014 and 2015 (n = 33,204 and 40,750, respectively) were obtained online (https://www.cdc.gov/prams/index.htm) while data from 2016 were obtained through written request to the CDC (CDC, 2019) (N = 35,441). PRAMS reporting areas select a stratified monthly sample of 100–300 resident mothers who delivered live births during the surveillance year and employ mixed-mode data collection including self-administered mailed questionnaires, and follow-up phone calls using interviewer-administered questionnaires for non-responders to the mailed version. Data are linked to reporting area birth-certificate data and weighted for sample design, nonresponse, and non-coverage (see Shulman et al. 2018, for a detailed description of survey design and procedures).

Study Measures

Pregnancy status.

All PATH female respondents < 50 years of age were queried regarding pregnancy status. W1 participants were asked if they had ever been pregnant, while W2 and W3 participants were queried about pregnancy in the previous 12 months. Across all waves, those who responded affirmatively to the pregnancy question received a second question asking if they were currently pregnant. Those answering in the affirmative on that second question were queried further regarding numbers of weeks pregnant. Those indicating they were pregnant at the time they completed the survey were treated as pregnant in the present study.

All NSDUH female respondents 12–44 years of age were asked to answer a yes/no question on whether they were pregnant, with those affirming pregnancy being queried further regarding trimester. All those affirming pregnancy at the time they completed the survey were treated as pregnant in the present study.

Finally, PRAMS participants were selected from stratified samples of 100–300 women per month who had live-birth deliveries in the surveillance year in each participating health department based on local birth-certificate records.

Current-smoking status.

Three different definitions of current smoking were used in PATH. Adults were identified as current established or current experimental users, with the former including respondents who reported smoking ≥ 100 cigarettes lifetime and smoking every day or some days at the time of survey completion, while the latter were those who reported smoking < 100 cigarettes lifetime and smoking every day or some days at the time of survey completion. Youth were identified as current smokers if they reported smoking at least one cigarette in the previous 30 days independent of lifetime number of cigarettes. Current smoking status in NSDUH was defined across all respondents as a report of smoking ≥ 100 cigarettes lifetime and use of at least one cigarette in the past 30 days. Participants were categorized as smokers in PRAMS if they reported any smoking during the last 3 months of pregnancy. All who met these various smoking-status criteria within the different surveys were categorized as current smokers in the present study.

Statistical Analyses

Descriptive analyses for PATH and NSDUH were conducted using PROC SURVEYFREQ in SAS 9.4 (SAS Institute, Cary, NC). Frequencies and percentages were generated across all respondents and were weighted to account for the complex sampling scheme, sampling probability, and differential non-response. Variance estimation was conducted as a variant of balanced repeated replication (Fay’s method) using a predetermined value ε set to 0.3, as recommended for the PATH study (Judkins, 1990; McCarthy, 1969), while variance estimation for NSDUH was based on the Taylor series linearization approach as recommended for this survey (Center for Behavioral Health Statistics and Quality, 2018). Age in PATH is defined as a continuous variable while NSDUH public use files provide age in categories. To address the first aim, PATH ages were recoded to correspond to the NSDUH age categories. Analysis of PRAMS data were limited to third-trimester prevalence estimates provided in the literature (years 2014 & 2015) or obtained by written request to the CDC (2016), with the analytical strategy used in the estimates described in detail in Shulman et al. (2018). As the purpose of this report was descriptive in nature, and the differing nature of the three surveys precludes combining them into a single analytic dataset, inferences regarding differences were based on means and associated confidence intervals. Logistic regression analyses examined within survey differences in smoking prevalence across trimesters in both the PATH and NSDUH datasets, using the PROC SURVEYLOGISTIC procedure, with variance estimation procedures as described above.

Results

Inclusion of Younger and Older Women in the Estimates

Comparisons addressing this aim were based on PATH and NSDUH surveys collapsed across all three years. We saw no evidence that including or excluding women below 18 years or above 44 years had a meaningful influence on overall prevalence estimates. Results using estimates from PATH were nearly identical independent of whether they were based on women between 12–44 (12.9%, 95% CI: 11.0–14.9%), 18–44 (13.2%, 11.1–15.3%), or 12–49 (13.2%, 11.1–15.3%) years (Figure 1, left panel). Similarly, estimates from NSDUH were nearly identical independent of whether they were based on women 12–44 (12.0%, 10.3–13.8%) or 18–44 (12.0%, 10.2–13.8%) years (Figure 1, right panel). Women above 44 years were not queried about pregnancy status in NSDUH, precluding the 12–49 years comparison in that survey.

Figure 1.

Figure 1.

Open in a new tab

Smoking prevalence among pregnant women, separated by the ages of women included in the estimate, collapsed across three survey years for Population Assessment of Tobacco and Health (PATH) and the National Survey on Drug Use and Health (NSDUH) studies. Bars represent 95% confidence intervals.

It is likely that we failed to observe an influence of age on prevalence estimates because so few women outside of the 18–44 age range were pregnant in PATH or NSDUH (Table 2). Only 4% (3.1–4.9%) and 1.9% (1.3–2.5%) of pregnant women in PATH and NSDUH, respectively, were < 18 years of age. In PATH, only 1% (0–2.4%) of pregnant women were >44 years of age.

Table 2.

Weighted percentages and 95% confidence intervals of ages of pregnant women from the Population Assessment of Health and Tobacco (PATH) study (2014–2016) and the National Survey on Drug Use and Health (NSDUH) study (2014–2016).1

Pregnant Women
PATH NSDUH
Weighted % 95% CI Weighted % 95% CI
Age
<18 4.0% 3.1–4.9% 1.9% 1.3–2.5%
18–25 37.7% 33.6–41.7% 33.4% 31.3–35.5%
26–34 48.5% 43.8–53.2% 50.2% 46.8–53.6%
35–44 9.8% 5.6–12.1% 14.6% 11.9–17.2%
44–49 1.0% 0–2.4% - -
Open in a new tab
1

Percentages were weighted to be representative of the U.S. population by correcting for selection probabilities, non-response, and post-stratification.

Influence of Trimester and Sample

Information on first and second trimester pregnancies were not available from PRAMS. Therefore, within-survey comparisons addressing this aim were based exclusively on PATH and NSDUH surveys collapsed across the three survey years. Decreasing trends in smoking prevalence were discernible across trimesters in both surveys (Figure 2). These differences were statistically significant in NSDUH (F (2, 50) = 7.57, p = .001) although not PATH (F (2,100) = .41, p = .662). NSDUH sample size was approximately double the sample size in PATH (n = 2,281 vs. 1,145), which almost surely increased the likelihood of discerning significant differences in between-trimester point estimates in the former.

Figure 2.

Figure 2.

Open in a new tab

Smoking prevalence among pregnant women, separated by trimester (T), across three comparable survey years for the Population Assessment of Tobacco and Health (PATH) and the National Survey on Drug Use and Health (NSDUH). Bars represent 95% confidence intervals.

Comparisons of prevalence estimates across the three surveys suggests that basing estimates exclusively on the third trimester underestimates overall smoking prevalence during pregnancy. Point prevalence estimates and associated confidence intervals from PATH and NSDUH, which include women across the three trimesters of pregnancy, overlapped in each of the three years examined: 2014 (PATH: 14.3%, 11.0–17.5%; NSDUH: 12.3%, 9.3–15.3%), 2015 (PATH: 12.4%, 8.7–16.1%; NSDUH: 13.6%, 10.4–16.9%), and 2016 (PATH: 13.8%, 9.8–17.7%; NSDUH: 10.2%, 7.2–13.2%). Estimates from PRAMS, which focus exclusively on smoking in the third trimester, were generally lower than those from both PATH and NSDUH in 2014 (9.9%, 9.4–10.5%), 2015 (8.8%, 8.4–9.3%), and 2016 (7.7%, 7.2–8.1%) (Figure 3), with PRAMS estimates only overlapping with PATH for 1 of 3 timepoints (2015), and with estimates from NSDUH for 2 of 3 (2014 and 2016).

Figure 3.

Figure 3.

Open in a new tab

Smoking prevalence overall among pregnant women across three comparable survey years for the Population Assessment of Tobacco and Health (PATH), the National Survey on Drug Use and Health (NSDUH), and the Pregnancy Risk Assessment Monitoring System (PRAMS) studies. Bars represent 95% confidence intervals.

When estimates from PATH and NSDUH were recalculated so that, like PRAMS, they were based exclusively on women in the third trimester, the point estimates and associated confidence intervals overlapped across the three surveys in each of the three years examined: 2014 (PATH: 13.4%, 8.1–18.8%; NSDUH: 10.9%, 5.7–16.0%; PRAMS: 9.9%, 9.4–10.5%), 2015 (PATH:11.1%, 5.6–16.5%; NSDUH: 9.0%, 4.1–13.9%; PRAMS: 8.8%, 8.4–9.3%), and 2016 (PATH: 13.2%, 6.9–19.6%; NSDUH: 5.7%, 2.6–8.8%; PRAMS: 7.7%, 7.2–8.1%) (Figure 4). This ability to reduce the between-survey differences discernible in Figure 3 by restricting the estimates across surveys to smoking in the third trimester as shown in Figure 4 underscores the influence of trimester on overall prevalence estimates. However, it also suggests that any influence of PRAMS estimates being based only on women who delivered live infants and samples stratified at the state rather than national level was negligible.

Figure 4.

Figure 4.

Open in a new tab

Smoking prevalence among pregnant women for the third trimester only across three comparable survey years for Population Assessment of Tobacco and Health (PATH), the National Survey on Drug Use and Health (NSDUH), and the Pregnancy Risk Assessment Monitoring System (PRAMS) studies. Bars represent 95% confidence intervals.

Discussion

The purpose of the present study was to examine whether excluding women under 18 or over 44 years of age, focusing exclusively on smoking in the third trimester, or basing assessments on nationally representative surveys or more selected national samples comprised exclusively of women who had live birth deliveries had a meaningful impact on U.S. prevalence estimates of smoking during pregnancy. Regarding the matter of maternal age, we saw no evidence in the PATH or NSDUH surveys that excluding women less than 18 or greater than 44 years meaningfully changed the overall prevalence estimates. The reason for this negligible impact is likely attributable to only a relatively small proportion of all U.S. pregnancies occurring among women in those two age brackets (e.g., Martin et al., 2018).

Regarding whether trimester matters, the present study provides evidence within and across surveys that basing estimates exclusively on smoking in the third trimester underestimates overall smoking prevalence. Other investigators have also reported significant decreases in U.S. smoking among pregnant women by trimester (e.g., Higgins et al., 2017) as well as month of pregnancy (Alshaarawy & Anthony, 2015). In trying to estimate how much prevalence is underestimated, comparisons of the PATH and NSDUH point estimates to PRAMS suggests a relatively robust influence of ~27–35%; changes within the NSDUH survey suggested a similarly robust influence of ~30%. The only observation that suggested a modest influence was the within-survey change in PATH of only about 7%, although, as was noted above, the accuracy of the point estimates on differences by trimester may have been limited by small sample sizes. However, even modest underestimates may undermine the precision in tracking progress in tobacco control and regulatory efforts to reduce smoking during pregnancy and risks understating the need for greater efforts to reduce smoking during pregnancy

Regarding the influence of basing national estimates only on women who delivered live births and the use of samples stratified at the state rather than national level, we did not discern a meaningful influence. That is, when smoking point-prevalence estimates and associated confidence intervals were all based on the 3rd trimester in PATH, NSDUH, and PRAMS, no meaningful differences across surveys were observed. As noted above, that result also suggests that other differences between PATH and NSDUH compared to PRAMS, such as querying pregnant women on their smoking status at the time of the interview in PATH and NSDUH versus retrospective reports among recently postpartum women on their smoking status during pregnancy in PRAMS, appears to have at most a negligible effect on the overall estimates.

Important to underscore is that all the estimates reported in the present study were based exclusively on self-report, and thus, already underestimate smoking among pregnant women by almost 25% (Dietz et al., 2010) in addition to the influence of the other factors discussed. Additionally, the design, data collection methods, and definitions differed across each data source, with PATH and NSDUH using longitudinal and cross-sectional research designs, respectively, and PRAMS relying on retrospective reporting on pregnancy at 2–6 months postpartum. Doing so can introduce limitations such as the one discussed above regarding the greater power in the NSDUH than PATH surveys for comparing smoking rates by trimester as cross-sectional surveys are designed to obtain a snapshot in time regarding differences in population subgroups and typically including larger Ns than longitudinal studies that are designed to facilitate investigation of questions about changes within population subgroups followed over time typically involving smaller sample sizes. Additionally, the definition of current smoking status varied across the different surveys used in the present study. Such variability is sure to introduce noise when comparing between the different surveys. Finally, sample sizes in PATH were relatively small, particularly when estimates were restricted to trimester, which could have impacted the precision of the PATH point estimates, and thusly, the robustness of within-survey changes in PATH.

Despite these limitations, the present study provides evidence-based guidance on how prevalence estimates of smoking among U.S. pregnant women are impacted by differences in whether the samples include younger (< 18 years) or older (> 44 years) women, represent smoking in any trimester or only the 3rd, and the use of data from nationally representative or more selected national samples. The current findings we believe will help to improve the precision in tracking progress in tobacco control and regulatory efforts to reduce smoking during pregnancy and increase awareness among investigators and consumers of this literature about factors that can influence prevalence estimates, and have the potential to be practically helpful to investigators and policy makers working in this important area of public health. The purpose of this report was not to recommend any one particular approach to estimating smoking prevalence among pregnant women, as each of the surveys discussed and others not examined have multiple valid reasons for their practices. What would improve the quality of the research in this area going forward, however, is if investigators were more conscientious about explicitly stating how the method used to estimate smoking during pregnancy may have influenced the validity of the estimate reported.

Highlights.

  • Data sources on U.S. smoking among pregnant women differ and may impact estimates

  • Trimester of pregnancy robustly impacts prevalence estimates

  • Age of women included has not meaningfully impacted prevalence estimates

  • National survey and PRAMS 3rd trimester estimates are comparable

  • Future research should acknowledge how varying methods may influence estimates

Acknowledgments:

The authors acknowledge the PRAMS Working Group (Alabama: Tammie Yelldell, MPH; Alaska: Kathy Perham Hester, MS, MPH; Arizonia: Enid Quintana-Torres, MPH Arkansas: Letitia de Graft-Johnson, DrPH, MHSA; Colorado: Ashley Juhl, MSPH; Connecticut: Jennifer Morin, MPH; Delaware: George Yocher, MS; Florida: Tara Hylton, MPH; Georgia: Florence A. Kanu, MPH; Hawaii: Matt Shim, PhD, MPH; Illinois: Julie Doetsch, MA; Indiana; Brittany Reynolds, MP; Iowa: Jennifer Pham; Kentucky: Tracey D. Jewell, MPH; Louisiana: Rosaria Trichilo, MPH; Maine: Tom Patenaude, MPH; Maryland: Laurie Kettinger, MS; Massachusetts: Hafsatou Diop, MD, MPH; Michigan: Peterson Haak; Minnesota: Mira Grice Sheff, PhD, MS; Mississippi: Brenda Hughes, MPPA; Missouri: Venkata Garikapaty, PhD; Montana: Emily Healy, MS; Nebraska: Jessica Seberger; New Hampshire: David J. Laflamme, PhD, MPH; New Jersey: Sharon Smith Cooley, MPH; New Mexico: Sarah Schrock, MPH; New York State: Anne Radigan; New York City: Lauren Birnie, MPH; North Carolina: Kathleen Jones-Vessey, MS; North Dakota; Grace Njau, MPH; Oklahoma: Ayesha Lampkins, MPH, CHES; Oregon: Cate Wilcox, MPH; Pennsylvania: Sara Thuma, MPH; Puerto Rico: Wanda Hernandez, MPH; Rhode Island: Karine Tolentino Monteiro, MPH; South Carolina: Harley T. Davis, PhD, MPSH; South Dakota: Maggie Minett; Texas: Tanya Guthrie, PhD; Tennessee: Ramona Lainhart, PhD; Utah: Nicole Stone, MPH; Vermont: Peggy Brozicevic; Virginia: Kenesha Smith, PhD, MSPH; Washington: Linda Lohdefinck; West Virginia: Melissa Baker, MA; Wisconsin—Fiona Weeks, MSPH; and Wyoming: Lorie Chesnut, PhD); and Centers for Disease Control and Prevention PRAMS Team, Applied Sciences Branch, Division of Reproductive Health.

Funding: This project was supported in part by Tobacco Centers of Regulatory Science (TCORS) award U54DA036114 from the National Institute on Drug Abuse (NIDA) and Food and Drug Administration (FDA), Institutional Training Grant award T32DA07242 from NIDA, Centers of Biomedical Research Excellence P20GM103644 award from the National Institute on General Medical Sciences Abuse, and Research award R01HD075669 from the National Institute of Child Health and Human Development (NICHD) and Centers for Disease Control and Prevention (CDC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Food and Drug Administration.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflicts of interest: None to declare.

References

  1. Alshaarawy O & Anthony JC (2015). Month-wise estimates of tobacco smoking during pregnancy for the United States, 2002–2009. Maternnal Child Health Journal, 19(5), 1010–1015. doi: 10.1007/s10995-014-1599-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Cnattingius S (2004). The epidemiology of smoking during pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tobacco Research, 6, 125–40. doi: 10.1080/14622200410001669187 [DOI] [PubMed] [Google Scholar]
  3. Center for Behavioral Health Statistics and Quality. National Survey on Drug Use and Health: Methodological summary and definitions. Substance Abuse and Mental Health Services Administration; 2015, Rockville, MD. [Google Scholar]
  4. Center for Behavioral Health Statistics and Quality. (2018). 2016 National Survey on Drug Use and Health: Methodological Resource Book (Section 13, Statistical Inference Report). Substance Abuse and Mental Health Services Administration, Rockville, MD [Google Scholar]
  5. Centers for Disease Control and Prevention (2019). Prevalence of Selected Maternal and Child Health Indicators for all PRAMS sites, Pregnancy Risk Assessment Monitoring System (PRAMS), 2012–2015. Retrieved from: https://www.cdc.gov/prams/pramstat/pdfs/mchindicators/PRAMS-All-Sites-2012-2015-508.pdf
  6. Dietz PM, England LJ, Shapiro-Mendoza CK, Tong VT, Farr SL, & Callaghan WM (2010). Infant morbidity and mortality attributable to prenatal smoking in the U.S. American Journal of Preventive Medicine, 39(1), 45–52. doi: 10.1016/j.amepre.2010.03.009 [DOI] [PubMed] [Google Scholar]
  7. Do EK, Green TL, Prom-Wormley EC, & Fuemmeler BF (2018). Social determinants of smoke exposure during pregnancy: Findings from waves 1 & 2 of the Population Assessment of Tobacco and Health (PATH) Study. Preventive Medicine Reports, 12, 312–320. doi: 10.116/j.pmedr.2018.10.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Howland RE, Mulready-Ward C, Madsen AM, Sackoff J, Nyland-Funke M, Bombard JM, & Tong VT (2015). Reliability of reported maternal smoking: Comparing the birth certificate to maternal worksheets and prenatal and hospital medical records, New York City and Vermont, 2009. Maternal and Child Health Journal, 19(9), 1916–1924. doi: 10.1007/s10995-015-1722-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Higgins ST, Redner R, Arger CA, Kurti AN, Priest JS, & Bunn JY (2017). Use of higher-nicotine/tar-yield (regular full-flavor) cigarettes is associated with nicotine dependence and smoking during pregnancy among U.S. women. Preventive Medicine, 104, 57–62. doi: 10.1016/j.ypmed.2017.07.02 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hyland A, Ambrose BK, Conway KP, Borek N, Lambert E, Carusi C, … Compton WM (2016). Design and methods of the Population Assessment of Tobacco and Health (PATH) Study. Tobacco Control, 26(4), 371–378. doi: 10.1136/tobaccocontrol-2016-052934 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Judkins DR (1990). Fay’s method for variance estimation. Journal of Official Statistics,6, 223–229. [Google Scholar]
  12. Kratz L, & Vaughan E (2012). Mental health problems, legal involvement, and smoking during pregnancy. Substance Use & Misuse, 47(6), 718–725. doi: 10.3109/10826084.2012.664238 [DOI] [PubMed] [Google Scholar]
  13. Kurti AN, Redner R, Lopez AA, Keith DR, Villanti AC, Stanton CA, … Higgins ST (2017). Tobacco and nicotine delivery product use in a national sample of pregnant women. Preventive Medicine, 104, 50–56. doi: 10.1016/j.ypmed.2017.07.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, & Drake P (2018). Births: Final Data for 2017. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics. National Vital Statistics System, 67(8), 1–50. [PubMed] [Google Scholar]
  15. McCarthy PJ (1969). Pseudo-replication: Half samples. Review of the International Statistical Institute, 37, 239–264. doi: 10.2307/1402116 [DOI] [Google Scholar]
  16. Shulman H, D’Angelo D, Harrison L, Smith R, & Warner L (2018). The Pregnancy Risk Assessment Monitoring System (PRAMS): Overview of design and methodology. American Journal of Public Health, 108(10), 1305–1313. doi: 10.2105/ajph.2018.304563 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Tong VT, Dietz PM, Farr SL, D’Angelo DV, & England LJ (2013a). Estimates of smoking before and during pregnancy, and smoking cessation during pregnancy: Comparing two population-based data sources. Public Health Reports, 128(3), 179–188. doi: 10.1177/003335491312800308 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Tong VT, Dietz PM, Morrow B, D’Angelo DV, Farr SL, Rockhill KM, & England LJ (2013b). Trends in smoking before, during, and after pregnancy — Pregnancy Risk Assessment Monitoring System, United States, 40 Sites, 2000–2010. Morbidity and Mortality Weekly Report: Surveillance Summaries., 62(6), 1–19. [PubMed] [Google Scholar]

RESOURCES