Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 23.
Published in final edited form as: Prev Med. 2022 Jul 7;161:107139. doi: 10.1016/j.ypmed.2022.107139

Labor force status as a buffer against mortality risks associated with alcohol consumption: A study of adult U.S. women, 2001–2015

Muntasir Masum a,*, Johnelle Sparks b
PMCID: PMC9507174  NIHMSID: NIHMS1835340  PMID: 35809823

Abstract

The association between women’s labor force participation, their alcohol consumption patterns, and mortality risk is unclear. This study assessed all-cause mortality risk among women in the United States, considering their labor force status and alcohol drinking. This study used discrete-time hazard models to examine this association using 2001–2015 National Health Interview Survey-Linked Mortality Files (NHIS-LMF) data (n = 147,714) for women aged 25 to 65 with 5725 deaths in this sample. Complex survey-weighted adjustments and E-values calculations were used to limit quantitative and observational biases. Alcohol consumption and labor force status together lead to substantial mortality risks. There is a statistically significant mortality risk among unemployed women (HR 2.15, 95% CI 1.18–3.91) and women not in labor force (HR 2.38, 95% CI 1.87–3.01). In the stratified models, non-Hispanic blacks (HR 1.48, 95% CI 1.30–1.67) and Asians (HR 1.93, 95% CI 1.54–2.44) have heightened mortality risks borne out of employment. Women with higher psychological distress have a 26% higher risk of all-cause mortality when not in labor force. With the help of cross-sectional data, this study demonstrates that women not in labor force and unemployed women are more likely to be affected by their drinking habits, and their employment status is associated with lower mortality risk. Further research should be focused on cause-specific mortality, gender roles and norms, reasons for unemployment, and comorbidities using more recent data, causal modeling techniques, and an extended mortality follow-up period.

Keywords: Alcohol drinking, Labor force status, Women, All-cause mortality, Psychological distress, NHIS-LMF

1. Introduction

Historically, light to moderate alcohol use was suggested to have a protective effect on health, such as ischemic stroke (Sacco et al., 1999) and multiple cardiovascular outcomes (Ronksley et al., 2011); but recent evidence shows that any level of alcohol consumption is not beneficial for health (Fillmore et al., 2006; Griswold et al., 2018; Rehm et al., 2010). Nearly one million people died from alcohol-related causes between 1999 and 2017; the majority of whom were in the 45 to 74 age range, and a statistically significant increase in death rates was among adults ages 25 to 34 (White et al., 2020). The historical gap in drinking between males and females is converging due to decreasing structural sexism, changing social norms, and completion of college education (McKetta et al., 2022). This study aims to examine women’s consumption of alcohol in relation to their labor force status for several reasons. First, following the 2010s, all-cause mortality rates, regardless of age, gender, race or ethnicity, increased among working-age adults (ages 25–65). Due to the opioid crisis, obesity, and drug and alcohol abuse, there is a penalty for mortality among middle-aged people. There was an increase in alcohol-related deaths in all but 13 states for males and all but 5 states for females among working-age adults (Harris et al., 2021). The trend for men is evident, whereas the trend for women is less pronounced. Second, according to recent trends, middle-aged adults are drinking more alcohol, binge drinking more frequently, and experiencing alcohol-related harms more frequently, with women in their 30s and 40s leading the way. Consumption increases appear to continue to occur at a faster rate for women than for men, even in older adult years (60 and older) (Keyes, 2022; Keyes et al., 2019a).

Women in younger cohorts are more likely to engage in heavy episodic drinking than women in older cohorts (Grucza et al., 2018; Holmila and Raitasalo, 2005; Keyes, 2022; Keyes et al., 2011, 2019b; Mulia et al., 2017; Roche and Deehan, 2002; White et al., 2015, 2020). For example, in 2020, around 20% of women ages 18 to 44 reported binge drinking in the United States (Centers for Disease Control and Prevention, 2020a; Dare et al., 2020; Evans-Polce et al., 2020). In addition, women with similar alcohol drinking levels as men face higher levels of alcohol-related health problems as they age, such as increased risk of breast cancer, cardiovascular disease, and chronic liver disease (Andersson et al., 2012; Erol and Karpyak, 2015; Jackson et al., 2015; Wagner and Anthony, 2007; White et al., 2020; Wilsnack et al., 2009).

Social and gender roles might influence women’s drinking habits due to expectations of household and parental responsibilities when women are subjected to greater social surveillance (Ahlström et al., 2001; Roos et al., 2006; Wilsnack et al., 2005). Conversely, as women make up the majority of the workforce and have responsibilities outside of the home, the stress associated with disproportionate responsibilities at work and conflicting role demands in managing home/work/life may result in increased alcohol drinking (Gutin and Hummer, 2020). Past research establishes that drinking levels, drinking-related problems, and other substance and drug use are positively but weakly associated with perceived job stress (Plant, 2008; Wilsnack et al., 2009).

Alcohol consumption levels differ depending on women’s labor force participation. Unemployment is frequently linked to dangerous drinking and vice versa (Atkinson et al., 2000; Cunradi et al., 2014; Jørgensen et al., 2019; Mustonen et al., 1994; Popovici and French, 2013). A few studies found that women prefer mild alcoholic beverages to drink at home with a friend or companion and are in control of their sobriety (Brooks et al., 2017; Graff et al., 2009). Contrary to this research, a positive association can also be seen between employed women and alcohol drinking and between longer work hours and high-risk drinking levels. Furthermore, alcohol consumption is often seen as a tension-reduction mechanism in popular culture (Erol and Karpyak, 2015; Hunt et al., 2015; Marchand, 2008; Wilsnack and Wilsnack, 1995). Research at this point is limited in examining women generally, working women in particular, and their drinking habits relative to their labor force status and mortality risk. This study addresses this gap by documenting the labor force status, alcohol drinking behavior, and all-cause mortality risks among women in the United States. Accordingly, we hypothesized that labor force participation would be associated with increased alcohol drinking among women ages 25 to 65, but labor force participation would serve as an effect modifier with a buffering effect resulting in lower mortality risks.

2. Methods

2.1. Data and sample

We utilized public-use data from the 2001–2013 National Health Interview Survey-Linked Mortality Files (NHIS-LMF), followed through December 31, 2015, derived from the IPUMS NHIS database (Blewett et al., 2019). The NHIS is a nationally representative cross-sectional survey that employs multistage probability sampling of the civilian noninstitutionalized U.S. population (National Center for Health Statistics, 2014). We included only adult women for this study and restricted the sample to ages 25 to 65 years (n = 163,721) to allow completion of education and current workforce participation. We excluded women ages 18 to 24 because the risk of mortality is low and excluded older women who may be out of the labor force. Finally, we excluded women with missing information, resulting in a final analytic sample of 147,714 women experiencing 5725 deaths. We used publicly available, de-identified NHIS data and did not involve human subjects research as defined at 45 CFR 46.102; hence did not require approval by the Institutional Review Board.

2.2. Outcome: All-cause mortality risk

All variables were measured at the same time (2001–2013), and mortality status was ascertained for two additional years, follow-up through December 31, 2015. The use of the NHIS combined with the follow-up mortality file provides a unique population-based data set to address health risks on mortality.

The primary outcome is all-cause mortality derived from the final mortality status determined by NCHS based on probabilistic matches of survey participants’ NHIS records to National Death Index (NDI) records (Blewett et al., 2019). Follow-up time was calculated as the number of years from the original interview date to either the date of death or the date of censoring, December 31, 2015. In this study, the follow-up period is up to 14 years.

2.3. Primary exposures

Self-reported labor force status is measured consistently through the study period. It reports employment status among working adults consisting of employed and unemployed individuals, while individuals who are neither employed nor unemployed are not in the labor force (U.S. Bureau of Labor Statistics, 2020). Therefore, we categorized respondents as employed (working with or without pay), unemployed (part of the labor force but unemployed), and not in the labor force.

Respondent’s self-reported current alcohol consumption status is a complete retrospective measure of current alcohol drinking status for the study period. It was categorized into six mutually exclusive categories: lifetime abstainer - fewer than 12 drinks during their lifetime; former drinker - at least 12 drinks in their lifetime, but none during the previous year; current light drinker [reference] – at least 12 drinks in the past year but three drinks or fewer per week, on average over the past year; current moderate drinker – more than three drinks but no more than seven drinks per week for women, on average over the past year; current heavy drinker – more than seven drinks per week for women, on average over the past year; and alcohol level unknown – unknown levels of alcohol consumption (Centers for Disease Control and Prevention, 2020b).

2.4. Covariates

Self-reported demographic, socioeconomic, behavioral, and anthropometric measures were included as covariates for the study period, 2001–2013. Demographic variables included age (25–29 [reference], 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–65), race/ethnicity (non-Hispanic White [reference], non-Hispanic black, Hispanic, non-Hispanic Alaskan Native or American Indian, non-Hispanic Asian, and non-Hispanic other), and marital status (Married [reference], widowed, divorced, separated, never married, unknown). Socioeconomic variables included educational attainment (High school graduate or equivalent [reference], less than high school, some college, college graduate or more, unknown). Behavioral and anthropometric variables included smoking status (Never smoked [reference], current every day smoker, current some day smoker, former smoker, unknown) and body-mass index (BMI) status, coded as (18.5 ≤ BMI < 25 = Normal weight [reference], BMI < 18.5 = Underweight, 25 ≤ BMI < 30 = Over Weight, BMI ≥ 30 = Obese, and unknown). To address the impact of stress on mortality, we used Kessler’s nonspecific psychological distress scale (K6), which was assessed using six questions: “During the Past 30 days, how often did you feel (1) so sad that nothing could cheer you up?, (2) nervous?, (3) restless or fidgety?, (4) hopeless?, (5) that everything was an effort?, (6) worthless?” Respondents had the following options: (1) all of the time; (2) most of the time; (3) some of the time; (4) a little of the time; (5) none of the time. After scoring responses 1 to 4 in respective order and 0 for “none of the time,” the total possible score ranged from 0 to 24. A score of 13 or greater on the K6 scale was defined as high psychological distress, described elsewhere (Kessler et al., 2003; Pratt, 2009). Cronbach’s alpha or internal consistency for the 6 items in the K6 scale is 0.88.

2.5. Analytic plan

In this study, we test the main association and then control for known covariates from theoretical models and tested in the literature. While confounders may be present, we controlled for all appropriate health behaviors, and physical and mental health correlates that could influence both alcohol consumption and labor force participation. Since no questions in the survey ask respondents to note which process might lead to another, we had to make use of existing literature to guide our model building strategy.

We do not have any issues of assignment randomization as complex survey design adjusts for who is selected. We used survey-weighted statistical analysis to adjust for clustering, stratification, and multistage sampling of NHIS. We examined labor force stratified bivariate associations across variables using chi-square (χ2) tests. Our multivariable analysis used complementary log-log discrete-time hazard models to estimate the all-cause and labor-force stratified all-cause mortality hazard among women as the outcome, controlling for all covariates. We estimated discrete-time survival models taking complex survey design into account. It is appropriate to use discrete-time survival models because the NHIS public data only include the year and not the exact date of death. This study utilizes person-years as the unit of observation, and the dependent variable indicates whether the respondent died that year. The models use a series of dummy variables for each year to implement a semi-parametric time model that makes the discrete-time estimates comparable to a Cox proportional hazards model (Allison, 2014; Daw, 2017; Jenkins, 1995; Singer and Willett, 1993). To examine the sensitivity of our findings to observational bias, we calculated a bias analysis using E-values and evaluated the minimal degree of association that an unmeasured confounder would need to have with both labor force status and mortality while controlling for all measured covariates to explain away the observed association (Bovell-Ammon et al., 2021; Mathur et al., 2018; VanderWeele and Ding, 2017; Watts et al., 2022). Analyses were performed using R software (version 4.1.2), and results are presented as hazard ratios (HR) with 95% confidence intervals (95% CI).

3. Results

3.1. Bivariate descriptive statistics

As seen in Table 1, women not in the labor force tend to be older at 47 years of age compared to employed (age 43) and unemployed women (age 41). The percentage of women not in labor force stays stable until age 55 and then sharply increases (55%) in the oldest age group (60–65). The average age at death for unemployed, employed, and women not in labor force are 45, 49, and 54, respectively. Women are most likely to be employed between 40 and 49 (approximately 74%); unemployment is highest at the youngest ages (5.8%). Women without a high school education are at a disadvantage for current employment (42.5%), while three-fourths of women with a college degree or more are employed (77.9%). Non-Hispanic Black (7.2%) women are overrepresented in the unemployed group, and Hispanic (35.2%) and non-Hispanic Alaskan Native or American Indian (35.8%) women are overrepresented in the not in the labor force category. Non-Hispanic White (69.3%) women have the highest employment rates than others.

Table 1.

Descriptive Statistics by Labor Force Status among Women Aged 25–65 Years, NHIS 2001–2015 (n=147,714)

Labor force status (weighted %)
Not in labor Force, n = 41,8911 Employed, n = 98,8701 Unemployed, n = 6,9531 p-value2
Age (years) 46.9 (12.3) 43.2 (10.7) 41.3 (10.8) <0.001
Age at death (years) 53.8 (9.6) 49.3 (9.9) 45.4 (10.7) <0.001
Age groups (years) <0.001
 25–29 24.2 70.0 5.8
 30–34 25.7 69.1 5.2
 35–39 24.5 70.6 4.9
 40–44 22.0 74.0 4.0
 45–49 21.6 74.4 4.0
 50–54 24.8 71.6 3.7
 55–59 34.3 62.5 3.2
 60–65 55.0 42.9 2.1
Educational attainment <0.001
 High school graduate or equivalent 32.7 63.0 4.3
 Less than high school 51.3 42.5 6.2
 Some college 25.5 70.0 4.5
 College graduate or more 19.4 77.9 2.8
 Unknown 33.9 58.5 7.5
Race and ethnicity <0.001
 Non-Hispanic white 27.5 69.3 3.2
 Hispanic 35.2 59.0 5.7
 Non-Hispanic Alaskan Native or American Indian 35.8 59.1 5.1
 Non-Hispanic Asian 27.8 67.8 4.4
 Non-Hispanic black 26.1 66.7 7.2
 Non-Hispanic other 27.0 66.3 6.7
Marital status <0.001
 Married 31.3 65.8 2.9
 Widowed 43.9 52.1 4.0
 Divorced 22.0 72.7 5.3
 Separated 28.2 63.7 8.1
 Never married 19.8 73.2 7.0
 Unknown 23.4 69.0 7.6
Kessler psychological distress index (0–24) <0.001
 High 57.3 34.0 8.6
 Low 27.1 69.0 3.9
Body-mass index <0.001
 Normal weight 25.8 70.6 3.7
 Underweight 27.4 68.9 3.7
 Overweight 27.7 68.2 4.2
 Obese 31.6 63.7 4.7
 Unknown 32.2 63.4 4.3
Smoking status <0.001
 Never smoker 27.2 69.2 3.6
 Current every day smoker 32.0 61.6 6.4
 Current some day smoker 26.5 67.3 6.2
 Former smoker 29.7 67.1 3.2
 Unknown 26.1 67.5 6.4
Alcohol drinking status <0.001
 Lifetime abstainer 38.4 57.4 4.2
 Former drinker 40.1 56.0 3.9
 Current light drinker 22.3 73.6 4.1
 Current moderate drinker 20.1 76.3 3.6
 Current heavy drinker 23.0 72.1 5.0
 Unknown 28.1 66.1 5.7
Open in a new tab
1

Mean (SD); %

2

Wilcoxon rank-sum test for complex survey samples; chi-squared test with Rao & Scott’s second-order correction

Never married and divorced women (73.2% each) have higher employment, and widowed are mostly not in labor (43.9%). Unemployment (8.1%) is highest for women who are separated. 57.3% of women with high psychological distress are not in the labor force. Another 8.6% of these women are unemployed. A graded association is observed between current employment and obesity status, so that a lower percentage of women are in the labor force as their obesity status increases. Similar patterns in the three labor force statuses are noted by smoking, although current smokers have the lowest percentage of employed women. Women who are current light drinkers (73.6%), moderate drinkers (76.3%), and heavy drinkers (72.1%) are more likely to be employed; conversely, lifetime abstainer (38.4%) and former drinker (40.1%) women are less likely to be in the labor force. Unemployment is the highest among current heavy drinkers (5%) and women whose alcohol level is unknown (5.7%). Weighted distributions of women in each drinking and labor force status category and the percentage of women that died over the period across these work/drinking categories are shown in supplementary table A. Women that were employed and light drinkers made up the largest share of respondents and were most likely to die over the follow-up period (18%), while light drinker not in the labor force or unemployed were less likely to die, 14% and 1%, respectively.

3.2. Multivariable adjusted results

Supplementary Table B presents the main effect of labor force participation on all-cause mortality adjusted for all the included covariates. Employed women have 51% (HR 0.49, 95% CI 0.46–0.53) and unemployed women have 40% (HR 0.60, 95% CI 0.51–0.72) lower mortality risks than women not in the labor force, respectively, keeping all the covariates constant, and these associations are statistically significant. The bias analysis indicates that unmeasured confounders would have to be very large in order to nullify the effect of being in the labor force on mortality compared to not being in the labor force, with E-values equal to 3.5 on all-cause mortality for exposure of employment and 2.72 on all-cause mortality for exposure of unemployment.

Table 2 presents the female labor force stratified adjusted hazard ratios of all-cause mortality estimated from the discrete-time hazard models. Age groups have a gradient relationship with all-cause mortality risk across the stratified labor force statuses, with older age groups having the highest risk for mortality. Lower college education increases mortality risks, and higher college education lowers the risk. For example, employed women with a college degree or more have a 33% lower mortality risk than women with a high school degree (HR 0.67, 95% CI 0.58–0.78). As expected, lower education shows a 27% higher all-cause mortality risk than employed women who completed at least 12 years of education (HR 1.27, 95% CI 1.08–1.49). Similar risks are also observed among women who are not in the labor force. All-cause mortality hazards are highest among employed Hispanic (HR 1.33, 95% CI 1.15–1.55), non-Hispanic Asian (HR 1.93, 95% CI 1.54–2.44), and non-Hispanic Black (HR 1.48, 95% CI 1.30–1.67) women compared to non-Hispanic white women, which hints at the impact of working conditions on all-cause mortality hazards for minority women. The Kessler psychological distress affects the hazard of all-cause mortality among those not in the labor force. The higher the distress score (baseline ≥13), the greater the risk of mortality (HR 1.26, 95% CI 1.11–1.43) for women not in labor force. The risks are non-statistically significant among employed and unemployed women in the labor force stratified models. Being married works as a buffer against mortality risk across different labor force statuses. Obese employed women have heightened all-cause mortality risks. A notable finding is that underweight women have higher mortality risks in the not in the labor force and employed categories than normal-weight individuals (HR 2.12, 95% CI 1.61–2.79) and (HR 1.78, 95% CI 1.28–2.49), respectively, and these associations are statistically significant. Unknown BMI is also a statistically significant mortality risk among employed women who are not in the labor force. Smoking presents serious health risks for women and current every day smokers have a statistically significant higher all-cause mortality risk than women who never smoked (HR 2.37, 95% CI 2.09–2.68), as seen in the model output in Table 2.

Table 2.

Discrete-time hazard models predicting all-cause mortality by labor force status, NHIS 2001–2015.

Not in Labor Force (n = 41,891) Employed (n = 98,870) Unemployed (n = 6,953)
Characteristic HR (95% CI)12 HR (95% CI)12 HR (95% CI)12
Age groups (ref: 25–29)
 25–29 — — —
 30–34 2.04 (1.43, 2.90) 1.15 (0.86, 1.55) 1.27 (0.61, 2.66)
 35–39 2.77 (1.97, 3.89) 1.75 (1.33, 2.30) 2.21 (1.13, 4.29)
 40–44 2.96 (2.12, 4.14) 2.45 (1.89, 3.18) 2.09 (0.99, 4.42)
 45–49 5.03 (3.65, 6.93) 3.14 (2.43, 4.04) 2.43 (1.15, 5.17)
 50–54 7.29 (5.34, 9.95) 4.29 (3.32, 5.55) 3.38 (1.63, 6.97)
 55–59 8.58 (6.29, 11.7) 5.87 (4.53, 7.61) 3.06 (1.40, 6.70)
 60–65 10.3 (7.63, 14.0) 8.06 (6.19, 10.5) 5.97 (2.93, 12.2)
Education (ref: High school graduate or equivalent)
 High school graduate or equivalent — — —
 Less than high school 1.20 (1.08, 1.34) 1.27 (1.08, 1.49) 1.84 (1.11, 3.07)
 Some college 0.91 (0.81, 1.02) 0.87 (0.77, 0.99) 1.26 (0.81, 1.96)
 College graduate or more 0.64 (0.54, 0.76) 0.67 (0.58, 0.78) 0.94 (0.50, 1.79)
 Unknown 0.76 (0.45, 1.31) 0.43 (0.19, 0.96) 2.45 (0.56, 10.8)
Race and Ethnicity (ref: Non-Hispanic White)
 Non-Hispanic white — — —
 Hispanic 0.88 (0.77, 1.00) 1.33 (1.15, 1.55) 0.75 (0.48, 1.17)
 Non-Hispanic Alaskan Native or American Indian 1.24 (0.86, 1.79) 1.08 (0.60, 1.95) 0.32 (0.04, 2.30)
 Non-Hispanic Asian 1.27 (0.94, 1.71) 1.93 (1.54, 2.44) 0.74 (0.23, 2.40)
 Non-Hispanic black 1.13 (1.01, 1.27) 1.48 (1.30, 1.67) 1.37 (0.96, 1.97)
 Non-Hispanic other 0.63 (0.25, 1.56) 0.62 (0.20, 1.93) 1.57 (0.21, 11.7)
Marital status (ref: Married)
 Married — — —
 Widowed 1.44 (1.25, 1.67) 1.18 (0.98, 1.42) 1.69 (0.83, 3.43)
 Divorced 1.34 (1.19, 1.51) 1.24 (1.10, 1.40) 0.93 (0.59, 1.47)
 Separated 1.18 (0.99, 1.42) 1.37 (1.12, 1.69) 1.50 (0.84, 2.69)
 Never married 1.49 (1.28, 1.74) 1.40 (1.21, 1.63) 0.89 (0.57, 1.41)
 Unknown 1.96 (1.12, 3.44) 1.07 (0.41, 2.80) 0.00 (0.00, 0.00)
Kessler psychological distress >=13
 Low — — —
 High 1.26 (1.11, 1.43) 1.20 (0.92, 1.55) 1.35 (0.75, 2.44)
Body mass index status (ref: Normal weight)
 Normal weight — — —
 Underweight 2.12 (1.61, 2.79) 1.78 (1.28, 2.49) 1.48 (0.23, 9.42)
 Overweight 0.94 (0.83, 1.06) 0.95 (0.83, 1.08) 0.82 (0.50, 1.35)
 Obese 1.08 (0.95, 1.23) 1.20 (1.06, 1.36) 0.92 (0.60, 1.40)
 Unknown 1.35 (1.17, 1.56) 1.37 (1.17, 1.61) 0.87 (0.48, 1.56)
Smoking status (ref: Never smoked)
 Never smoker — — —
 Current every day smoker 2.34 (2.08, 2.63) 2.37 (2.09, 2.68) 2.25 (1.53, 3.31)
 Current some day smoker 2.02 (1.62, 2.52) 1.65 (1.27, 2.14) 2.10 (1.17, 3.77)
 Former smoker 1.39 (1.23, 1.58) 1.33 (1.17, 1.52) 1.16 (0.63, 2.14)
 Unknown 1.10 (0.52, 2.35) 1.30 (0.65, 2.58) 1.91 (0.20, 17.8)
Alcohol consumption (ref: Current light drinker)
 Current light drinker — — —
 Lifetime abstainer 1.55 (1.36, 1.76) 1.51 (1.33, 1.71) 1.66 (1.09, 2.53)
 Former drinker 1.71 (1.51, 1.94) 1.43 (1.26, 1.63) 1.34 (0.81, 2.19)
 Current moderate drinker 0.92 (0.73, 1.16) 1.00 (0.81, 1.22) 1.16 (0.61, 2.20)
 Current heavy drinker 1.16 (0.92, 1.48) 1.13 (0.90, 1.42) 1.74 (0.84, 3.60)
 Unknown 1.18 (0.83, 1.68) 0.98 (0.65, 1.45) 0.90 (0.29, 2.72)
Open in a new tab
1

*p<0.05; **p<0.01; ***p<0.001

2

HR = Hazard Ratio, CI = Confidence Interval

Interesting patterns emerge when examining alcohol consumption coefficients across labor force status. Women identifying as former drinkers have the highest all-cause mortality risk across the labor force statuses compared to current light drinkers. Much like the descriptive patterns, lifetime abstainers who are not in labor force have a 55% increased risk of dying over the follow-up period than current light drinkers. There is no difference in all-cause mortality risks between current moderate/heavy drinkers and current light drinkers.

To test our key research question more directly, the complete matrix of 18 interactions (3 categories for labor force participation and 6 categories for the alcohol consumption variables) was tested based on employment status and alcohol consumption patterns with employed women that are current light drinkers serving as the reference group (Fig. 1, Supp. Table C). After holding all other variables in the model constant, 11 of the 17 interaction categories show a statistically significant difference from the reference category. For employed women, all-cause mortality risks are higher among those who identify as lifetime abstainers (HR 1.62, 95% CI 1.43–1.83) and former drinkers (HR 1.45, 95% CI 1.28–1.65). Women not in the labor force and former drinkers have the highest risk of all-cause mortality over the period (HR 3.43, 95% CI 3.05–3.85). In contrast, current moderate and heavy drinkers have similar risks to employed women that are current light drinkers. Women not in the labor force have an increased risk of mortality across all alcohol categories than currently employed light drinkers. Lastly, unemployed women and lifetime abstainers, former drinkers, and current heavy drinkers have higher risks of all-cause mortality than light drinkers who are employed.

Fig. 1.

Fig. 1.

Open in a new tab

Adjusted hazard ratios for all-cause mortality risk among women (ages 25–65): Interaction of labor force status and alcohol consumption behavior.

4. Discussion

This analysis aimed to investigate whether women’s all-cause mortality risk varies across different levels of alcohol drinking, assuming that employed women are more likely to drink for several reasons (access, stress, male work culture), and drinking increases all-cause mortality risks (Wilsnack et al., 2009). Our results indicate a notable impact of alcohol drinking on all-cause mortality risks for adult women in the United States. Lifetime abstainers and former drinkers show an elevated risk of all-cause mortality compared to other alcohol consumption groups; however, they are usually concentrated in women in the oldest age categories and disproportionately among Hispanics and non-Hispanic blacks, partly explaining the cumulative mortality risk (Rogers et al., 2013).

It is hard to distinguish a causal direction between not being active in the labor force and alcohol patterns, but these women are at the highest risk of death. A noteworthy finding is that controlling for other factors, mortality risks among employed woman aged 25–65 compared to women who are not in the labor force is lower. This is supported by a consistent finding in the literature that women who work have better health than those unemployed and not in the labor force. Furthermore, being unemployed and using alcohol might indicate relatively poor health functioning, which indicates reverse causation in such scenarios. Women who are unhealthy do not enter the workforce (and end up dying), while healthy women enter the workforce and are able to survive (i.e., secure housing, food, etc.) (Avendano and Berkman, 2014; Lundin et al., 2010).

Numerous studies have established the educational gradient in all-cause mortality, and education works through different mechanisms to influence health and mortality (Glymour et al., 2014). As women complete their education in their early twenties and enter the labor market, they have natural age protection on health and mortality outcomes, even if women in earlier age groups are exposed to higher levels of alcohol drinking. Higher educational attainment increases the chances of survivability among adult women across different labor force statuses in the United States (Krueger et al., 2015; McPherson et al., 2004), likely because the women have an easier time securing and keeping jobs or marrying men with higher levels of education and better jobs/incomes that may afford their wives to maintain a home and raise children.

Even though non-Hispanic black women are equally employed as non-Hispanic white women, they have a statistically significant heightened risk of mortality than non-Hispanic whites and Hispanics. Past studies have also concluded that non-Hispanic black women with similar drinking habits are more susceptible to poor health than non-Hispanic white and Hispanic women (Jackson et al., 2015; Ransome et al., 2017). Part of the story is likely tied to social support mechanisms that can help when these women go through periods of unemployment or as a potential buffer against heavy alcohol consumption.

Being married has known health benefits, and earlier studies have demonstrated that marital status is a protective factor for health and mortality (Evans-Polce et al., 2020; Liu et al., 2020). The fully adjusted hazards ratios show that any marital status other than the reference category, ‘married,’ has a heightened risk of all-cause mortality among women ages 25 to 65. Similar hazards are seen among women who are not in the labor force. However, being employed and unemployed does not impact women’s mortality risk in a statistically significant way. Marital status can work through complex pathways; the Kessler Psychological Distress mechanism can be one of them (Pratt, 2009). Past studies have concluded that individuals with a higher psychological distress score (≥ 13) are less likely to have higher educational attainment or be married. This study finds evidence that women have a higher risk of all-cause mortality when not in the labor force and experience high psychological distress. Again, it is difficult to determine the temporal order of these processes and associated outcomes in the absence of causality.

Smoking is strongly correlated to other risky health behaviors, such as alcohol consumption, and is associated with cumulative health risks attributable to smoking over the life course (Nandi et al., 2014). Our results provide strong evidence of every day smoking’s impact on women’s mortality across labor force statuses. The rate of every day smoking is higher among employed women; similarly, the all-cause mortality risk is highest among women employed in the labor force and current every day smokers. Compared to women who never smoked, every day smokers and someday smokers have around 2 to 2.5 times higher risk of mortality, controlling for other covariates (Carter et al., 2015).

Participation in the labor force, either employed or unemployed, serves as a potential buffer against increased mortality risks for women compared to women not in the labor force. Stated differently, employed or unemployed women have lower death hazards in the follow-up period (Heikkilä et al., 2012). While the coefficient for current moderate drinkers remains non-statistically significant, the direction of influence changes for employed and not in labor force women in the stratified models compared to the total sample. This could indicate that moderate alcohol consumption influences increased health risks, most likely through other mechanisms and potentially reverse causation; this association warrants further research for these two groups of women. The all-cause mortality hazard for heavy drinkers and unemployed women has one of the most statistically significant hazards observed for the drinking status variable across the labor force stratified models. Lastly, former drinkers have a lower hazard ratio for death over the follow-up period if they are currently employed compared to women that are not in the labor force or unemployed (Azagba and Sharaf, 2011; Garcy and Vågerö, 2012).

The analyses in this study adjusted for many confounding factors, but residual confounding cannot be completely eliminated. People with disabilities and comorbidities are included in the analytic sample and distributed in different labor force categories. Lower mortality risks among employed women may occur partly as a function of selection biases, such as the self-selection of healthy participants into the labor force. Theory and prior empirical research guided variable selection and model construction to minimize selection bias. One of the strengths of this study is that the study sample is large, particularly since multiple survey years were combined in the final analytic sample, and NHIS tracks health behavior and mortality records over time. There are likely unobserved heterogeneities; however, we address the core variables necessary to test our hypothesis (Gregg et al., 2018; Nie et al., 2020; Ohlsson et al., 2021; Rosenbaum et al., 1998).

Our findings highlight the need for future research to consider potential underlying causes of the associations observed in this analysis, including gender roles and norms, workplace peer pressure, reasons for unemployment, duration of unemployment, and the broader social context. Future research should also focus on comorbidities, occupational categories, hours of work, insurance status, work and employment status over the past 12 months, and working and living conditions. To better understand how other health behaviors may influence mortality risks, it would be helpful to explore the health and mortality risks of lifetime abstainers and former drinkers separately. Furthermore, a longer mortality follow-up period would include more deaths, resulting in more precise estimates.

4.1. Limitations

We have used the most recent publicly available data linking health behaviors to mortality records, and all the variables included in the analysis are time updated. However, data up to 2015 were analyzed because the public use NHIS is linked to mortality records up to December 31, 2015. NHIS is an ongoing national health survey, and more recent public data with linked mortality records will be available in the future, which will allow a longer follow-up period to analyze these associations, which would be especially helpful for younger cohorts of women. In this study, we adjusted for many confounding factors, but residual confounding cannot be eliminated entirely. Lower mortality risks among employed women may occur partly as a function of selection biases, such as the self-selection of healthy participants into the labor force. Although people with disabilities and comorbidities are not excluded from the sample, it may only increase the degree of selection bias if people who report poor health are removed from the study. We conducted stratified analysis by including individuals in specific categories of labor force participation to avoid selection and misclassification biases. The results were consistent. However, these methods might not be adequate to deal with the issue (Knott et al., 2015; Xi et al., 2017).

This study’s limitations also include not looking at cause-specific mortality, controlling for occupational categories, personal lifestyle choices, and considering additional random effects such as cohort, which would have allowed for an age-period-cohort analysis. Furthermore, we did not have data on alcohol misuse (i.e., AUDIT), and there are limitations to looking solely at alcohol use frequency. Finally, the alcohol consumption status was derived from survey responses and may therefore be subject to recall bias. Nevertheless, the primary strength of this work is that it looks at working-age women, their labor force participation status, and alcohol use and assesses their mortality risks, controlling for known correlated health risks such as smoking and obesity.

5. Conclusion

Using a nationally representative large sample of U.S. working-age adult women, our study findings indicate that women’s labor force status is associated with their drinking habits and is prospectively associated with various all-cause mortality hazards; however, more research is needed to disentangle these processes’ temporal order. The risks associated with drinking are greater for women not in labor force and unemployed women than employed women who drink lightly. However, we do not have information to determine if labor force participation, or not, has a direct influence on alcohol consumption behaviors. In spite of higher levels of alcohol consumption, employment in the labor force serves as a potential buffer against all-cause mortality. There is potential for selection bias and misclassification bias, and women’s socioeconomic status should be taken into account when estimating their alcohol consumption-related mortality risks so as to minimize the effects of comorbidities and unmeasured confounders. Identifying factors that put women at a disadvantage for problematic alcohol consumption and unstable employment is essential for researchers and practitioners. This study is particularly useful when considering socially disruptive events, such as COVID-19, which lead to changes in working conditions and alcohol consumption.

Supplementary Material

Supplementary tables

Acknowledgments

The authors would like to thank Dr. Corey Sparks for his guidance with the model specification of discrete-time survival models. Preliminary study findings were previously presented at the 2021 annual meeting of Population Association of America (PAA), virtual format.

Role of funding sources

The work described was supported by T32 DA017629 from the National Institute on Drug Abuse (NIDA). The content is the sole responsibility of the authors and does not necessarily represent the official position of NIDA or the National Institutes of Health.

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ypmed.2022.107139.

References

  1. Ahlström S, Bloomfield K, Knibbe R, 2001. Gender differences in drinking patterns in nine European countries: Descriptive findings. Subst. Abus. 22, 69–85. 10.1080/08897070109511446. [DOI] [PubMed] [Google Scholar]
  2. Allison P, 2014. Event History and Survival Analysis. SAGE Publications, Inc., 2455 Teller Road, Thousand Oaks California 91320 United States: 10.4135/9781452270029. [DOI] [Google Scholar]
  3. Andersson C, Eklund M, Sundh V, Thundal KL, Spak F, 2012. Women’s patterns of everyday occupations and alcohol consumption. Scand. J. Occup. Ther. 19, 225–238. 10.3109/11038128.2010.527013. [DOI] [PubMed] [Google Scholar]
  4. Atkinson JS, Montoya ID, Trevino RA, Richard AJ, 2000. Labor force participation in a sample of substance users. Am. J. Drug Alcohol Abuse 26, 355–367. 10.1081/ADA-100100249. [DOI] [PubMed] [Google Scholar]
  5. Avendano M, Berkman LF, 2014. Labor markets, employment policies, and health. In: Berkman LF, Kawachi I, Glymour MM (Eds.), Social Epidemiology, 2 ed. Oxford University Press, New York, pp. 182–233. 10.1093/med/9780195377903.001.0001. [DOI] [Google Scholar]
  6. Azagba S, Sharaf MF, 2011. The effect of job stress on smoking and alcohol consumption. Heal. Econ. Rev. 1, 1–14. 10.1186/2191-1991-1-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Blewett LA, Drew JAR, King ML, Williams KCW, 2019. IPUMS Health Surveys: National Health Interview Survey, Version 6.4 [dataset] [WWW Document]. Minneapolis, MN: IPUMS, 2019. 10.18128/D070.V6.4. [DOI] [Google Scholar]
  8. Bovell-Ammon BJ, Xuan Z, Paasche-Orlow MK, LaRochelle MR, 2021. Association of incarceration with mortality by race from a National Longitudinal Cohort Study. JAMA Netw. Open 4, e2133083. 10.1001/jamanetworkopen.2021.33083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brooks AT, Magaña Lòpez M, Ranucci A, Krumlauf M, Wallen GR, 2017. A qualitative exploration of social support during treatment for severe alcohol use disorder and recovery. Addict. Behav. Rep. 6, 76–82. 10.1016/j.abrep.2017.08.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carter BD, Abnet CC, Feskanich D, Freedman ND, Hartge P, Lewis CE, Ockene JK, Prentice L, Speizer FE, Thun MJ, Jacobs EJ, 2015. Smoking and mortality — beyond established causes. N. Engl. J. Med. 372, 631–640. 10.1056/NEJMsa1407211. [DOI] [PubMed] [Google Scholar]
  11. Centers for Disease Control and Prevention, 2020a. Chronic Disease Indicators (CDI) [WWW Document]. URL. https://www.cdc.gov/cdi/index.html (accessed 12.6.21).
  12. Centers for Disease Control and Prevention, 2020b. Glossary - Alcohol [WWW Document]. National Health Interview Survey. URL. https://www.cdc.gov/nchs/nhis/alcohol/alcohol_glossary.htm (accessed 10.8.20). [Google Scholar]
  13. Cunradi CB, Ames GM, Xiao H, 2014. Binge drinking, smoking, and marijuana use: the role of women’s labor force participation. J. Work. Behav. Health 29, 210–223. 10.1080/15555240.2014.928600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dare J, Wilkinson C, Traumer L, Kusk KH, McDermott ML, Uridge L, Grønkjær M, 2020. “Women of my age tend to drink”: the social construction of alcohol use by Australian and Danish women aged 50–70 years. Sociol. Health Illness 42, 35–49. 10.1111/1467-9566.12991. [DOI] [PubMed] [Google Scholar]
  15. Daw J, 2017. Contribution of four comorbid conditions to racial/ethnic disparities in mortality risk. Am J. Prevent. Med. Social Determin. Health Approach Health Dispar. Res. 52, S95–S102. 10.1016/j.amepre.2016.07.036. [DOI] [PubMed] [Google Scholar]
  16. Erol A, Karpyak VM, 2015. Sex and gender-related differences in alcohol use and its consequences: contemporary knowledge and future research considerations. Drug Alcohol Depend. 156, 1–13. 10.1016/j.drugalcdep.2015.08.023. [DOI] [PubMed] [Google Scholar]
  17. Evans-Polce RJ, Jang BJ, Maggs JL, Patrick ME, 2020. Gender and age differences in the associations between family social roles and excessive alcohol use. Soc. Sci. Med. 244, 112664 10.1016/j.socscimed.2019.112664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fillmore KM, Kerr WC, Stockwell T, Chikritzhs T, Bostrom A, 2006. Moderate alcohol use and reduced mortality risk: Systematic error in prospective studies. Addict. Res. Theory. 10.1080/16066350500497983. [DOI] [PubMed] [Google Scholar]
  19. Garcy AM, Vågerö D, 2012. Social Science & Medicine The length of unemployment predicts mortality, differently in men and women, and by cause of death : a six year mortality follow-up of the Swedish 1992 e 1996 recession. Soc. Sci. Med. 74, 1911–1920. 10.1016/j.socscimed.2012.01.034. [DOI] [PubMed] [Google Scholar]
  20. Glymour MM, Avendano M, Kawachi I, 2014. Socioeconomic status and health. In: Berkman LF, Kawachi I, Glymour MM. (Eds.), Social Epidemiology, 2 ed. Oxford University Press, New York, pp. 17–62. 10.1093/med/9780195377903.001.0001. [DOI] [Google Scholar]
  21. Graff FS, Morgan TJ, Epstein EE, McCrady BS, Cook SM, Jensen NK, Kelly S, 2009. Engagement and retention in outpatient alcoholism treatment for women. Am. J. Addict. 18, 277–288. 10.1080/10550490902925540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gregg EW, Cheng YJ, Srinivasan M, Lin J, Geiss LS, Albright AL, Imperatore G, 2018. Trends in cause-specific mortality among adults with and without diagnosed diabetes in the USA: an epidemiological analysis of linked national survey and vital statistics data. Lancet 391, 2430–2440. 10.1016/S0140-6736(18)30314-3. [DOI] [PubMed] [Google Scholar]
  23. Griswold MG, Fullman N, Gakidou E, 2018. Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 6736, 31310–31312. 10.1016/S0140-6736(18)31310-2. [DOI] [Google Scholar]
  24. Grucza RA, Sher KJ, Kerr WC, Krauss MJ, Lui CK, McDowell YE, Hartz S, Virdi G, Bierut LJ, 2018. Trends in adult alcohol use and binge drinking in the Early 21st-century United States: a meta-analysis of 6 national survey series. Alcohol. Clin. Exp. Res. 42, 1939–1950. 10.1111/acer.13859. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Gutin I, Hummer RA, 2020. Occupation, employment status, and “despair”-associated mortality risk among working-aged U.S. adults, 1997–2015. Prev. Med 137 10.1016/J.YPMED.2020.106129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Harris KM, Majmundar M, Becker T, 2021. High and Rising Mortality Rates Among Working-Age Adults. The National Academies Press, Washington D.C.. 10.17226/25976 [DOI] [PubMed] [Google Scholar]
  27. Heikkilä K, Nyberg ST, Fransson EI, Alfredsson L, de Bacquer D, Bjorner JB, Bonenfant S, Borritz M, Burr H, Clays E, Casini A, Dragano N, Erbel R, Geuskens GA, Goldberg M, Hooftman WE, Houtman IL, Joensuu M, Jöckel KH, Kittel F, Knutsson A, Koskenvuo M, Koskinen A, Kouvonen A, Leineweber C, Lunau T, Madsen IEH, Hanson LLM, Marmot MG, Nielsen ML, Nordin M, Pentti J, Salo P, Rugulies R, Steptoe A, Siegrist J, Suominen S, Vahtera J, Virtanen M, Väänänen A, Westerholm P, Westerlund H, Zins M, Theorell T, Hamer M, Ferrie JE, Singh-Manoux A, Batty GD, Kivimäki M, 2012. Job strain and alcohol intake: a collaborative meta-analysis of individual-participant data from 140 000 men and women. PLoS One 7, 1–7. 10.1371/journal.pone.0040101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Holmila M, Raitasalo K, 2005. Gender differences in drinking: why do they still exist? Addiction 100, 1763–1769. 10.1111/j.1360-0443.2005.01249.x. [DOI] [PubMed] [Google Scholar]
  29. Hunt G, Asmussen Frank V, Moloney M, 2015. Rethinking gender within alcohol and drug research. Subst. Use Misuse 50, 685–692. 10.3109/10826084.2015.978635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Jackson CL, Hu FB, Kawachi I, Williams DR, Mukamal KJ, Rimm EB, 2015. Black-white differences in the relationship between alcohol drinking patterns and mortality among US men and women. Am. J. Public Health 105, S534–S543. 10.2105/AJPH.2015.302615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jenkins SP, 1995. Easy estimation methods for discrete-time duration models. Oxf. Bull. Econ. Stat. 57, 129–136. 10.1111/j.1468-0084.1995.tb00031.x. [DOI] [Google Scholar]
  32. Jørgensen MB, Pedersen J, Thygesen LC, Lau CJ, Christensen AI, Becker U, Tolstrup JS, 2019. Alcohol consumption and labour market participation: a prospective cohort study of transitions between work, unemployment, sickness absence, and social benefits. Eur. J. Epidemiol. 34, 397–407. 10.1007/s10654-018-0476-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kessler RC, Barker PR, Colpe LJ, Epstein JF, Gfroerer JC, Hiripi E, Howes MJ, Normand S-LT, Manderscheid RW, Walters EE, Zaslavsky AM, 2003. Screening for serious mental illness in the general population. Arch. Gen. Psychiatry 60, 184–189. 10.1001/archpsyc.60.2.184. [DOI] [PubMed] [Google Scholar]
  34. Keyes KM, 2022. Age, period, and cohort effects in alcohol use in the United States in the 20th and 21st centuries: implications for the coming decades. Alcohol Res 42, 02. 10.35946/arcr.v42.1.02. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Keyes KM, Li G, Hasin DS, 2011. Birth cohort effects and gender differences in alcohol epidemiology: a review and synthesis. Alcohol. Clin. Exp. Res. 35, 2101–2112. 10.1111/j.1530-0277.2011.01562.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Keyes KM, Jager J, Mal-Sarkar T, Patrick ME, Rutherford C, Hasin D, 2019a. Is there a recent epidemic of women’s drinking? A critical review of national studies. Alcohol. Clin. Exp. Res. 43, 1344–1359. 10.1111/acer.14082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Keyes KM, Jager J, Mal-Sarkar T, Patrick ME, Rutherford C, Hasin D, 2019b. Is there a recent epidemic of women’s drinking? A critical review of national studies. Alcohol. Clin. Exp. Res. 43, 1344–1359. 10.1111/acer.14082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Knott CS, Coombs N, Stamatakis E, Biddulph JP, 2015. All cause mortality and the case for age specific alcohol consumption guidelines: Pooled analyses of up to 10 population based cohorts 350. 10.1136/bmj.h384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Krueger PM, Tran MK, Hummer RA, Chang VW, 2015. Mortality attributable to low levels of education in the United States. PLoS One 10, 1–13. 10.1371/journal.pone.0131809. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Liu H, Umberson D, Xu M, 2020. Widowhood and mortality: gender, race/ethnicity, and the role of economic resources. Ann. Epidemiol. 45, 69–75. 10.1016/j.annepidem.2020.02.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lundin A, Lundberg I, Hallsten L, Ottosson J, Hemmingsson T, 2010. Unemployment and mortality - a longitudinal prospective study on selection and causation in 49321 Swedish middle-aged men. J. Epidemiol. Community Health 64, 22–28. 10.1136/jech.2008.079269. [DOI] [PubMed] [Google Scholar]
  42. Marchand A, 2008. Alcohol use and misuse: what are the contributions of occupation and work organization conditions? BMC Public Health 8, 1–12. 10.1186/1471-2458-8-333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Mathur MB, Ding P, Riddell CA, VanderWeele TJ, 2018. Website and R package for computing e-values. Epidemiology 29, e45–e47. 10.1097/EDE.0000000000000864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. McKetta S, Prins SJ, Hasin D, Patrick ME, Keyes KM, 2022. Structural sexism and women’s alcohol use in the United States, 1988–2016. Soc. Sci. Med. 301, 114976 10.1016/j.socscimed.2022.114976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. McPherson M, Casswell S, Pledger M, 2004. Gender convergence in alcohol consumption and related problems: Issues and outcomes from comparisons of New Zealand survey data. Addiction 99, 738–748. 10.1111/j.1360-0443.2004.00758.x. [DOI] [PubMed] [Google Scholar]
  46. Mulia N, Karriker-Jaffe KJ, Witbrodt J, Bond J, Williams E, Zemore SE, 2017. Racial/ethnic differences in 30-year trajectories of heavy drinking in a nationally representative U.S. sample. Drug Alcohol Depend. 170, 133–141. 10.1016/j.drugalcdep.2016.10.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Mustonen H, Paakkanen P, Simpura J, 1994. Drinking habits among the employed and unemployed. Nordisk Alkoholtisdkrift (Nordic Alcohol Studies) 11, 21–34. 10.1177/145507259401101s06. [DOI] [Google Scholar]
  48. Nandi A, Glymour MM, Subramanian SV, 2014. Association among socioeconomic status, health behaviors, and all-cause mortality in the United States. Epidemiology 25, 170–177. 10.1097/EDE.0000000000000038. [DOI] [PubMed] [Google Scholar]
  49. 2013 National Health Interview Survey (NHIS) Public Use Data Release, 2014. National Center for Health Statistics. [Google Scholar]
  50. Nie J, Wang J, Aune D, Huang W, Xiao D, Wang Y, Chen X, 2020. Association between employment status and risk of all-cause and cause-specific mortality: a population-based prospective cohort study. J. Epidemiol. Community Health 74, 428–436. 10.1136/jech-2019-213179. [DOI] [PubMed] [Google Scholar]
  51. Ohlsson A, Eckerdal N, Lindahl B, Hanning M, Westerling R, 2021. Non-employment and low educational level as risk factors for inequitable treatment and mortality in heart failure: a population-based cohort study of register data. BMC Public Health 21, 1–12. 10.1186/s12889-021-10919-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Plant ML, 2008. The role of alcohol in women’s lives: a review of issues and responses. J. Subst. Abus. 13, 155–191. 10.1080/14659890802040880. [DOI] [Google Scholar]
  53. Popovici I, French MT, 2013. Does unemployment lead to greater alcohol consumption? Ind. Relat. 52, 444–466. 10.1111/irel.12019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Pratt LA, 2009. Serious Psychological Distress, as Measured by the K6, and Mortality. Ann. Epidemiol. 19, 202–209. 10.1016/j.annepidem.2008.12.005. [DOI] [PubMed] [Google Scholar]
  55. Ransome Y, Carty DC, Cogburn CD, Williams DR, 2017. Racial disparities in the association between alcohol use disorders and health in black and white women. Biodemogr. Social Biol. 63, 236–252. 10.1080/19485565.2017.1335589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Rehm J, Baliunas D, Borges GLG, Graham K, Irving H, Kehoe T, Parry CD, Patra J, Popova S, Poznyak V, Roerecke M, Room R, Samokhvalov AV, Taylor B, 2010. The relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction 105, 817–843. 10.1111/j.1360-0443.2010.02899.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Roche AM, Deehan A, 2002. Women’s alcohol consumption: emerging patterns, problems and public health implications. Drug Alcohol Rev. 21, 169–178. 10.1080/09595230220139073. [DOI] [PubMed] [Google Scholar]
  58. Rogers RG, Krueger PM, Miech R, Lawrence EM, 2013. Lifetime abstainers and mortality risk in the United States. 10.1553/POPULATIONYEARBOOK2013S165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA, 2011. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. Bmj 342, 479. 10.1136/bmj.d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Roos E, Lahelma E, Rahkonen O, 2006. Work-family conflicts and drinking behaviours among employed women and men. Drug Alcohol Depend. 83, 49–56. 10.1016/j.drugalcdep.2005.10.009. [DOI] [PubMed] [Google Scholar]
  61. Rosenbaum WL, Sterling TD, Weinkam JJ, 1998. Use of multiple surveys to estimate mortality among never, current, and former smokers: changes over a 20-year interval. Am. J. Public Health 88, 1664–1668. 10.2105/AJPH.88.11.1664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Sacco RL, Elkind M, Boden-Albala B, Lin IF, Kargman DE, Hauser WA, Shea S, Paik MC, 1999. The protective effect of moderate alcohol consumption on ischemic stroke. JAMA 281, 53–60. 10.1001/JAMA.281.1.53. [DOI] [PubMed] [Google Scholar]
  63. Singer JD, Willett JB, 1993. It’s about time: using discrete-time survival analysis to study duration and the timing of events. J. Educ. Stat. 18, 155–195. 10.3102/10769986018002155. [DOI] [Google Scholar]
  64. U.S. Bureau of Labor Statistics, 2020. Labor force characteristics (CPS) [WWW Document]. United States Department of Labor. URL. https://www.bls.gov/cps/lfcharacteristics.htm#emp(accessed 10.8.20). [Google Scholar]
  65. VanderWeele TJ, Ding P, 2017. Sensitivity analysis in observational research: introducing the e-value. Ann. Intern. Med. 167, 268–274. 10.7326/M16-2607. [DOI] [PubMed] [Google Scholar]
  66. Wagner FA, Anthony JC, 2007. Male-female differences in the risk of progression from first use to dependence upon cannabis, cocaine, and alcohol. Drug Alcohol Depend. 86, 191–198. 10.1016/j.drugalcdep.2006.06.003. [DOI] [PubMed] [Google Scholar]
  67. Watts BV, Gottlieb DJ, Riblet NB, Gui J, Shiner B, 2022. Association of medication treatment for opioid use disorder with suicide mortality. AJP 179, 298–304. 10.1176/appi.ajp.2021.21070700. [DOI] [PubMed] [Google Scholar]
  68. White A, Castle I-JP, Chen CM, Shirley M, Roach D, Hingson R, 2015. Converging patterns of alcohol use and related outcomes among females and males in the United States, 2002 to 2012 Alcohol. Clin. Exp. Res. 39, 1712–1726. 10.1111/acer.12815. [DOI] [PubMed] [Google Scholar]
  69. White AM, Castle IJP, Hingson RW, Powell PA, 2020. Using death certificates to explore changes in alcohol-related mortality in the United States, 1999 to 2017 Alcohol. Clin. Exp. Res. 44, 178–187. 10.1111/acer.14239. [DOI] [PubMed] [Google Scholar]
  70. Wilsnack SC, Wilsnack RW, 1995. Drinking and problem drinking in US women: patterns and recent trends. In: Galanter M(Ed.), Recent Developments in Alcoholism. Plenum Press, New York, pp. 29–60. [DOI] [PubMed] [Google Scholar]
  71. Wilsnack RW, Wilsnack SC, Obot IS, 2005. Why study gender, alcohol and culture? In: Obot IS, Room R (Eds.), Alcohol, Gender and Drinking Problems: Perspective from Low and Middle Income Countries. World Health Organization, Geneva, Switzerland, pp. 1–24. [Google Scholar]
  72. Wilsnack RW, Wilsnack SC, Kristjanson AF, Vogeltanz-Holm ND, Gmel G, 2009. Gender and alcohol consumption: patterns from the multinational GENACIS project. Addiction 104, 1487–1500. 10.1111/j.1360-0443.2009.02696.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Xi B, Veeranki SP, Zhao M, Ma C, Yan Y, Mi J, 2017. Relationship of alcohol consumption to all-cause, cardiovascular, and cancer-related mortality in U.S. adults. J. Am. Coll. Cardiol. 70, 913–922. 10.1016/j.jacc.2017.06.054. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary tables
NIHMS1835340-supplement-Supplementary_tables.docx (47.7KB, docx)

RESOURCES