IJNHCR Journal

Introduction

Dementia affects approximately 6.5 million people in the US, the number is expected to grow to 13.8 million by 2050, which will increase the demand for caregiving [1]. Family caregivers provide about 80% of dementia care, often performing over 90% of caregiving hours in community settings [2]. However, family caregivers are often the “hidden victims” of dementia, experiencing high stress, physical exhaustion, financial burden, and social isolation, which can negatively impact their physical and mental  health, including higher cardiometabolic risks, more chronic pain, depression and anxiety [3-5]. This distress can further deteriorate caregiver relationships with persons living with dementia (PLWD), lowering the quality of care [6]. Thus, understanding these stressors and developing supportive interventions are imperative to alleviate caregiver burden and improve their well-being.

Stress and Maladaptive Coping

The Stress and Coping Model explains how family caregivers appraise stress and select coping strategies in response to caregiving demands [7]. Family caregivers may choose maladaptive emotionfocused coping, such as denial, avoidance, and toxic positivity, when they feel overwhelmed by the caregiving demands and lack of control over the situation [8]. This type of coping may provide temporary emotional relief, but also can lead to negative longterm health outcomes [9]. Previous research has identified that caregivers with higher caregiving burden, lower socioeconomic status, limited social support, and poor physical health are more prone to unhealthy behaviors like smoking, binge drinking, and poor dietary habits, which further exacerbate their stress and health decline [10-12]. Therefore, it is essential to identify risk factors for negative coping behaviors among dementia caregivers to provide tailored support that equips them with the necessary resources and more adaptive coping strategies.

Smoking and Alcohol Use Among Dementia Caregivers

Previous studies suggest that dementia caregivers are at a higher risk of smoking and alcohol consumption compared to non-caregivers, though findings vary by specific behaviors and caregiver characteristics [13,24]. While dementia caregivers and non-caregivers report similar levels of heavy alcohol use, adult children and those providing fewer daily caregiving hours exhibit higher prevalence [24]. Additionally, younger dementia caregivers show higher smoking rates compared to older caregivers, who report lower smoking rates than the general population [14]. The mechanisms behind these behaviors are complex, as not all caregivers rely on smoking or alcohol for stress relief [15,16]. These variations underscore the complexity of these behaviors and the need for research that accounts for sample differences to identify high-risk groups.

This study also examines e-cigarette use among dementia caregivers, given its rising popularity and potential health risks [17] E-cigarette use among U.S. adults reached 4.5% in 2021, with health risks including lung injury, cardiovascular disease, and nicotine dependence [18-20]. However, there is limited research on e-cigarette use and the related factors among dementia caregivers. Identifying its prevalence among dementia caregivers is crucial because these products are often marketed as less harmful alternatives to traditional tobacco and may be more appealing to caregivers seeking perceived lower-risk coping strategies [21].

The Current Study

In summary, individuals’ coping behaviors are shaped by various factors, including age, gender, education level, income, and social support. When these background factors intersect with dementia caregiving, the impact on coping strategies can be more pronounced. For example, dementia caregivers with low socioeconomic status tend to exhibit higher rates of physical inactivity due to limited supportive resources [22], while those with poor mental health report greater tendencies toward poor dietary habits [23]. Thus, this exploratory study investigates two research questions: a) how do dementia caregivers’ rates of smoking tobacco, e-cigarette use, and alcohol drinking compared to non-caregivers? And b) what demographic, socioeconomic, and health contextual factors contribute to increased smoking and drinking behaviors among dementia caregivers?

This study is crucial for improving understanding of coping patterns among dementia caregivers and identifying subgroups at higher risk of adopting unhealthy behaviors. By identifying contributors to smoking, e-cigarette use, and alcohol consumption, results can inform tailored interventions and policy strategies to mitigate health disparities and improve support for dementia caregivers.

Methods

Study Design

This study is a cross-sectional analysis utilizing data from the 2017 Behavioral Risk Factor Surveillance System (BRFSS), a nationwide health-related telephone survey conducted annually by the Centers for Disease Control and Prevention (CDC) that collects data on health behaviors, chronic conditions, and the use of preventive services. It includes both core modules, which are asked across all states, and optional modules, which states can choose based on local priorities [25]. In 2017, five states (Alaska, Hawaii, New Mexico, Oregon, and Rhode Island) including 28,479 records that administered the modules on tobacco and e-cigarettes use, alcohol consumption, caregiving, and chronic health-related conditions.

Data

This study compares dementia caregivers and non-caregivers on tobacco, e-cigarette, and alcohol use. An examination of missing data on key variables revealed an overall missing rate of 4.8%, with only 0.1% missing among dementia caregivers; thus, cases with missing values were excluded without imputation [26]. The final sample comprised 19,813 respondents, including 19,402 noncaregivers and 511 dementia caregivers.

Measurements

Outcome variables: Tobacco and e-cigarette use were classified into four ordinal categories: every day smoker, some-day smoker (within the past 30 days), former smoker (quit), and never smoker. Alcohol consumption was measured using two numeric variables: the number of alcohol drinking days in the past 30 days and the average number of drinks consumed per drinking day (with 1 drink defined as a 12-ounce beer, a 5-ounce glass of wine, or a drink containing one shot of liquor).

Dementia caregiving status: Dementia caregivers were identified through BRFSS caregiving module responses. Participants who provided care for a family member or friend due to dementia and other cognitive impairment disorders such as Alzheimer’s disease (AD) were classified as dementia caregivers.

Socio-demographic characteristics of participants included: age (18-54, 55-64, 65-74, and 75+ years), sex (female/male), race/ ethnicity (non-Hispanic White, Black, Hispanic, Asian, and others), marital status (married/not married), education level (high school or below, some college, and bachelor’s degree or higher), housing status (own, rent, and other arrangements), employment status (employed, unemployed, and retired), annual household income (<$25,000, $25,000-$50,000, and >$50,000).

Health characteristics. Self-rated health was assessed on a 5-point scale ranging from poor to excellent and, following established categorization methods [27], was classified as good/better versus fair/poor. Depression was determined by self-reported diagnosis of a depressive disorder (yes/no). The number of chronic conditions was based on the presence of 12 chronic diseases, including hypertension, high cholesterol, heart attack, coronary heart disease, stroke, asthma, skin cancer, other types of cancer, chronic obstructive pulmonary disease (COPD), arthritis, kidney disease, and diabetes.

Analytic Strategies

All analyses were conducted using Stata 18 [28]. BRFSS sampling weights were applied to account for the complex sampling design [39], to generalize estimates from 19,813 participants to a weighted population of 4,943,993 across five selected states. Descriptive statistics were calculated, stratified by dementia caregiving status, with categorical variables reported as frequencies/percentages and continuous variables as means/standard deviations (SDs). Bivariate comparisons were conducted using chi-square tests for categorical variables and t-tests for continuous variables.

Multinomial logistic regression models examined tobacco and e-cigarette use, with “never smoker” as the reference group, estimating relative risk ratios (RRRs) for other smoking patterns. These models were adjusted for dementia caregiving status, sociodemographic factors, and health characteristics. Interaction terms between caregiving and each predictor were tested in follow-up models. For significant interactions, predicted outcome values were estimated, with figures illustrating group differences. For example, if the interaction between dementia caregiving and sex was significant in the model comparing the RRR of everyday smokers to never smokers, predicted RRR values were estimated for female non-caregivers, female dementia caregivers, male noncaregivers, and male dementia caregivers.

Two zero-inflated negative binomial (ZINB) models analysed the number of drinking days in the past 30 days and the average number of drinks per drinking day. The ZINB models addressed the high frequency of zero values and over-dispersion [29]. Similar to the smoking models, interaction terms examined whether dementia caregiving status increased the risk of alcohol consumption across sociodemographic and health backgrounds.

Results

Participant Characteristics

The study sample included 19,813 participants, of whom 511 were dementia caregivers. After applying sampling weights, the sample represents approximately 5 million people, including 112,750 dementia caregivers (Table 1). Dementia caregivers significantly differed from non-caregivers in several sociodemographic and health characteristics. Dementia caregivers were older, with a higher proportion aged 55 years and older (40.9% vs. 51.5%; χ² = 32.15, p < 0.001), predominantly female (62.9% vs. 49.7%; χ² = 30.60, p < 0.001), and more likely to self-identify as Asian (12.5% vs. 8.7%) or other racial/ethnic groups (13.6% vs. 9.5%; χ² = 38.87, p < 0.01). Due to sampling limitations, no Black participants were identified in the dementia caregiver group; thus, subsequent analyses do not assess racial differences for this group. Dementia caregivers were more likely to be married (68.2% vs. 55.5%; χ² = 28.98, p < 0.01), and homeowners (83.9% vs. 66.9%; χ² = 59.07, p < 0.001). They also reported greater health burdens than non-caregivers, including higher rates of fair or poor selfrated health (12.5% vs. 17.4%; χ² = 7.77, p < 0.05), depression (23.8% vs. 19.9%; χ² = 13.64, not significant), and more chronic conditions (1.62 vs. 1.34; t = 3.42, p < 0.01).

Dementia caregivers showed similar patterns in smoking and e-cigarette use compared to non-caregivers. Although not statistically significant, dementia caregivers had higher rates of everyday smoking (14.8% vs. 10.1%) and former smoking (28.3% vs. 26.7%), with fewer never smokers (54.1% vs. 58.4%). E-cigarette use patterns were nearly identical between caregivers (1.8%) and non-caregivers (1.8%). Alcohol consumption is also similar, with dementia caregivers reporting a slightly fewer drinking days in the past 30 days (4.95 vs. 5.58 days; t = 1.23, not significant), and fewer drinks per day (1.23 vs. 1.37; t = 0.85, not significant).

Table 1: Descriptive Statistics for the Total Sample by Dementia Caregiving Status.

Multivariate Analyses for Smoking and E-Cigarette Use

Multinomial regression results (Table 2, Model 1) showed that dementia caregivers had a tendency toward higher everyday smoking rates compared to never smokers, though the association was not significant (RRR = 1.94, 95% CI: 0.96, 3.92). Significant predictors of everyday smoking included older age (participants aged 75+ were less likely to smoke every day; RRR = 0.35, 95% CI: 0.22, 0.56; p < 0.001), sex (females were less likely to smoke every day; RRR = 0.62, 95% CI: 0.50, 0.76; p < 0.001), and race/ethnicity (Hispanic participants were significantly less likely to smoke every day compared to non-Hispanic Whites; RRR = 0.33, 95% CI: 0.24, 0.44; p < 0.001). Higher levels of education and income were associated with lower smoking risks, with participants holding a bachelor’s degree or higher (RRR = 0.17, 95% CI: 0.13, 0.22; p < 0.001) and those with household incomes over $50K (RRR = 0.52, 95% CI: 0.39, 0.70; p < 0.001) being less likely to smoke daily. In contrast, more chronic conditions (RRR = 1.17, 95% CI: 1.08, 1.28; p < 0.001) or depression (RRR = 1.99, 95% CI: 1.56, 2.53; p < 0.001) were linked to increasing smoking.

Interaction analyses identified significant effects of income level (being dementia caregiver × low income: RRR=7.03, p=0.001; see Supplementary Table 1) and employment status (being dementia caregiver × not employed/retired: RRR = 4.89, p = 0.03; see Supplementary Table 2). Figures 1A and 1B illustrate that low-income and unemployed dementia caregivers had the highest predicted probability of everyday smoking compared to their non-caregiving counterparts.

For e-cigarette use, multinomial regression results revealed no significant association between dementia caregiving status and everyday e-cigarette use (RRR = 0.92, 95% CI: 0.27, 3.13). Most covariates showed similar relationships to those observed in the tobacco smoking models, including age, gender, and depression. Participants living in rented housing were more likely to report everyday e-cigarette use compared to homeowners (RRR = 1.99, 95% CI: 1.22, 3.25; p < 0.01).

Significant interaction effects were found between dementia caregiving status and race/ethnicity (being dementia caregiver × being Asian: RRR = 5.44, p < 0.001; see Supplementary Table 3) and gender (being dementia caregiver × being female: RRR = 1.96, p < 0.001; see Supplementary Table 4). Figures 2A and 2B show that Asian dementia caregivers had the highest probability of everyday e-cigarette use, while female caregivers exhibited higher predicted probabilities of someday e-cigarette use than male caregivers.

Note: The regression models do not include African American participants due to a lack of representativeness in the dementia caregiver subsample.

* p < 0.05, ** p < 0.01, *** p < 0.001.

Table 2: Multinomial Regression Models for Tobacco Smoking and E-Cigarettes Use

Figure 1: Predicted probability of being an everyday cigarette smoker by income and employment status.

Description: This figure shows the predicted probability of being an everyday cigarette smoker among dementia caregivers and noncaregivers, stratified by (A) annual household income levels and (B) employment status. The results indicate that low-income and unemployed dementia caregivers are at higher risk for daily smoking.

Figure 2: Predicted probability of being an everyday or someday e-cigarette user by race/ethnicity and gender.

Description: This figure illustrates the predicted probability of being an everyday or someday e-cigarette user among dementia caregivers and non-caregivers, stratified by (A) race/ethnicity and (B) gender. The findings indicate that Asian and female dementia caregivers have a significantly higher likelihood of e-cigarette use.

Multivariate Analyses for Alcohol Consumption

ZINB regression models (Table 3) found no significant association between dementia caregiving status and the number of drinking days (IRR = 0.90, 95% CI: 0.70, 1.16) or the number of drinks per drinking day (IRR = 1.29, 95% CI: 0.87, 1.91). However, several covariates showed notable associations. Women drank on fewer days (IRR = 0.73, 95% CI: 0.67, 0.79; p < 0.001), and consumed fewer drinks per day (IRR = 0.67, 95% CI: 0.61, 0.74; p < 0.001). Being in older age groups (55-64, 65-74, and 75+) drank more frequently (e.g., IRR = 1.56 for age 75+; p < 0.001), but consumed fewer drinks per day (e.g., IRR = 0.60 for age 75+; p < 0.001). Participants living in rented or other non-owned housing arrangements reported higher drinking intensity compared to homeowners (e.g., IRR = 1.26 for renters; p < 0.001).

Interaction analyses identified significant effects between dementia caregiving and race/ethnicity for drinking days (being dementia caregiver × being Hispanic: IRR = 0.37, p = 0.02; see Supplementary Table 5) and between dementia caregiving and marital status for the number of drinks per day (being dementia caregiver × being not married: IRR = 2.23, p = 0.04; see Supplementary Table 6). Figure 3A shows that Hispanic dementia caregivers reported fewer drinking days than their non-caregiver counterparts, while Whites showed no difference. Figure 3B illustrates that unmarried dementia caregivers consumed significantly more drinks per day than both married caregivers and non-caregivers across marital statuses.

 Note: The regression models do not include African American participants due to a lack of representativeness in the dementia caregiver subsample.

* p < 0.05, ** p < 0.01, *** p < 0.001.

Table 3: Zero-Inflated Negative Binomial Regression Models for Alcohol Consumption.

Figure 3: Alcohol consumption patterns among dementia caregivers and non-caregivers.