Pandemic-Related Worries and Stressors and Use of Tobacco and Alcohol Among Postpartum Women in the United States Early in the COVID-19 Pandemic

Study Population

Data came from the COVID-19: Health in Pregnancy and Postpartum (CHIPP) study. CHIPP was a cross-sectional electronic survey designed to examine how the COVID-19 pandemic influenced pregnancy and postpartum health, health behaviors, and access to perinatal care in the United States.^26,27 The postpartum survey was conducted from May 21 through June 11, 2020. To recruit a diverse sample, CHIPP advertised the study through community health programs serving predominantly racial and ethnic minority or low-income women and through group Facebook accounts focused on people in racial and ethnic minority populations. The recruitment cap limited the target sample size to 50% African American women. The inclusion criteria for the survey were being a US resident, being aged 18 to 44 years, and having delivered a singleton baby in a US hospital at any time from March 2020 through the survey date. Among 865 survey participants, we excluded 4 participants for not meeting inclusion criteria and 69 participants for missing values for key variables, such as tobacco use and alcohol use, resulting in 792 participants in our analysis.

This study’s protocol was approved by the institutional review board at the University of South Carolina (exempt status), and all respondents gave informed consent to participate in the study.

Measures

The primary independent variables were pandemic-related worries and stressors. The CHIPP study modified the Cambridge Worry Scale ²⁸ for postpartum women and the pandemic setting. Ten questions assessed pandemic-related worries: (1) giving birth in the hospital setting; (2) changing birth plans; (3) their new baby; (4) friends and family; (5) stigma or discrimination from other people; (6) having enough basic supplies; (7) not getting baby supplies; (8) getting emotional or social support from family, friends, partners, a counselor, or someone else; (9) getting financial support from family, friends, partners, and organizations; and (10) other difficulties or challenges. Responses were measured on a 5-point Likert scale ranging from not at all (0) to all the time (4). The total score of the pandemic-related worries (range: 0-40) was the sum of responses to the 10 items on worries. The survey used the Pandemic Stress Index ²⁹ to measure pandemic-related stressors, which included the following 4 items: (1) participants getting tested for COVID-19, (2) losing income due to COVID-19, (3) family members testing positive for COVID-19, and (4) family deaths in or outside the household due to COVID-19 or related complications. We examined each stressor individually and as a composite measure by summing the 4 pandemic-related stressors (range: 0-4).

The primary dependent variables were tobacco and alcohol use. The term “tobacco use” refers to using combustible and noncombustible tobacco products and electronic nicotine delivery systems. Questions asked participants about using combustible tobacco cigarettes and other tobacco/nicotine products such as electronic cigarettes (e-cigarettes), hookahs, smokeless tobacco, and cigars, cigarillos, or little filtered cigars. We measured the postpartum use of combustible tobacco cigarettes in the following categories: never used, used throughout the pregnancy, or resumed smoking in 0 to 1 month, 2 to 3 months, or ≥4 months after childbirth. Additionally, the survey asked participants about their use of other tobacco products at survey time. We classified participants as negative for tobacco use if they did not use any tobacco products and positive for tobacco use if they reported any tobacco product use during the postpartum period.

We assessed the number of alcoholic drinks participants consumed in an average week since the new baby was born during the pandemic by using the following survey categories: no drinking and <1, 1 to 3, 4 to 6, 7 to 13, and ≥14 drinks per week. We considered use of ≥1 drink per week to be positive for alcohol use. We classified joint use of tobacco and alcohol by using the following categories: nonuse, tobacco use only, alcohol use only, and use of both tobacco and alcohol. We based these analyses on the consideration of alcohol drinking as a health behavior, rather than a health risk, according to CDC guidelines for alcohol drinking among breastfeeding postpartum women. ²⁵

Participants who reported they used tobacco and/or alcohol were also asked if they increased their postpartum use of these products during the pandemic. We used 2 categories to assess any changes in tobacco use among participants who used tobacco since the start of the pandemic: increased tobacco use and no increased tobacco use. We used the same classification for participants who consumed alcohol.

Besides the independent variables of pandemic-related worries and stressors and dependent variables of tobacco and alcohol use, the survey collected data on the following variables for assessment for potential confounding^30,31: maternal age, race and ethnicity, education, health insurance, parity, annual household income, chronic health conditions, completion of a postpartum checkup, and preterm birth.

Statistical Analysis

We used Pearson χ² tests of independence to determine if characteristics of participants differed significantly according to use and nonuse of tobacco and alcohol. We used multiple logistic regression models to measure the associations of pandemic-related worries and stressors with 2 outcomes for postpartum tobacco use: (1) tobacco use among all postpartum women and (2) increased use among tobacco users. We used the same approach for alcohol use. Furthermore, we used multinomial models to assess the association between the joint use of tobacco and alcohol. For each model, we simultaneously considered pandemic-related worries and stressors by adjusting for worries while studying the independent association of stressors with postpartum tobacco use, and vice versa. In these analyses, we assessed the associations for both individual items and total scores of worries and stressors when we evaluated worries or stressors as the primary independent variable, whereas we used only the total scores for the adjustment when adjusting for worries/stressors. We conducted sensitivity analyses that modeled pandemic-related worries and stressors separately to assess their independent association with tobacco and/or alcohol use. We estimated adjusted odds ratios (AORs) and 95% CIs. We used SAS version 9.4 (SAS Institute, Inc) to perform all analyses.

Strengths and Limitations

Our study had several strengths. For one, our study, to our knowledge, is the only study to report on pandemic-related worries and stressors in relation to tobacco and alcohol use among postpartum women, a population frequently overlooked in pandemic-related behavioral research, early in the COVID-19 pandemic, when COVID-19 caused large spikes in disease and death and led to widespread lockdowns. ³⁴ Our study was unique in that it examined tobacco use, alcohol use, and pandemic-related changes in use. Our findings provide evidence that external crises may be associated with changes in health behavior among postpartum women. The study population resided throughout the United States and was diverse, with a high proportion of African American women (45.5%).

Our study also had several limitations. First, tobacco and alcohol use were self-reported, but prior studies found that self-reported smoking history was valid and strongly correlated with actual smoking status. ³⁵ Second, the study adapted the Cambridge Worry Scale ²⁸ and the Pandemic Stress Index ²⁹ to assess pandemic-related worries and stressors among postpartum women, which were commonly used among the general population but had not been validated among postpartum women. Third, the cross-sectional design meant we could not definitively determine the temporal sequence between pandemic-related worries and stressors and postpartum tobacco and alcohol use. Fourth, our convenience sampling method yielded a sample primarily composed of women aged 25 to 44 years, with high proportions of African American and non-Hispanic White women, often associated with greater tobacco use than other groups of women,^36,37 and individuals with higher socioeconomic status, typically related to less use of tobacco compared with those with lower socioeconomic status. ³⁷ Fifth, the higher prevalence of tobacco use observed in our study, compared with other studies, may be due to pandemic-related stress^11-13 and differences in measurement compared.^36,37 Our study used online surveys of a sample population to examine any tobacco use during the postpartum period. Consequently, our prevalence estimates should be interpreted cautiously as specific to this sample. Sixth, we conducted the study early in the pandemic, when pandemic worries and stressors were very high, so the findings are most relevant to the early phase of the COVID-19 pandemic. Finally, we did not adjust for preexisting depression or anxiety. Preexisting anxiety-related disorders are associated with elevated pandemic-related worries and stressors. ³⁸ However, considering the low prevalence of respondents diagnosed with depression or anxiety during pregnancy in our sample (3.5%) and that survey participants may have been motivated by higher (or lower) worries or stressors, the influence of preexisting depression or anxiety on perceived worries and stressors is likely limited.