Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 10.
Published in final edited form as: GeroPsych (Bern). 2022 Feb 25;35(4):211–225. doi: 10.1024/1662-9647/a000285

Mealtime Behavior and Depressive Symptoms in Late-Life Marriage

Talha Ali 1, Gail McAvay 1, Joan K Monin 2
PMCID: PMC9912988  NIHMSID: NIHMS1827048  PMID: 36777454

Abstract

This study examined whether one spouse’s mealtime behaviors were associated with their own and their partner’s depressive symptoms among older, married couples. We examined gender differences in these associations and tested marital satisfaction as a mediator of these associations. 101 couples self-reported mealtime behavior (number of meals, snacks, fast-food meals, and meals eaten alone), depressive symptoms, and marital satisfaction. Results of the Actor Partner Interdependence Model revealed a statistically significant actor effect of number of fast-food meals on depressive symptoms and a significant partner effect of number of fast-food meals and number of meals eaten alone on depressive symptoms. There were gender differences. Husbands’ marital satisfaction mediated the effect of meals eaten alone on depressive symptoms. Wife’s marital satisfaction mediated the effect of the husband’s meals eaten alone, and wife’s number of fast-food meals on the wife’s depressive symptoms. Findings have implications for dyadic interventions to improve depressive symptoms.

Keywords: depression, mealtime behavior, marriage, gender

A growing body of research shows that older married spouses’ mental health is linked (Kim et al., 2006; Meyler et al., 2007; Peek & Markides, 2003). For example, there is overwhelming evidence that couples are concordant in depressive symptomatology and emotional well-being (Meyler et al., 2007; Peek et al., 2006). One argument for this concordance is that spouses share similar health behaviors (Franks et al., 2012; Meyler et al., 2007; Smith & Zick, 1994; Umberson, 1992; Waite, 1995). An important health behavior that has implications for mental health within individuals (Fulkerson et al., 2007; Hoang et al., 2019) and has also been shown to be concordant among spouses is eating behavior (Pachucki et al., 2011). In particular, consumption of fruit, vegetables, eggs, and milk are concordant among couples (Barrett-Connor et al., 1982; Macario & Sorensen, 1998; Meyler et al., 2007). Spouses are frequently involved in their partners’ dietary practices through meal planning, food preparation, and diet plans (August & Sorkin, 2010).

Among couples, social transmission of mealtime behaviors can happen for a number of reasons. According to the shared resources hypothesis, spouses share a similar environment, including physical environment characterized by available resources in the environment (e.g., easy access to snacks because of proximity to a vending machine), economic environment (e.g., financial resources to purchase healthy foods), and sociocultural environment (e.g., social ties and values and beliefs towards behaviors such as eating fast food) (Smith & Zick, 1994). This shared environment translates into shared health behaviors and risks; for example, if one spouse consumes more fast-food meals, the other spouse is also likely to eat more fast-food meals (Waite, 1995). Spouses are also in a unique position to influence their partner’s health behaviors by providing health-related social support or social control as posited by the social control theory (Franks et al., 2012; Umberson, 1992). In fact, among married couples, the most frequently reported source of social persuasion and pressure to maintain healthy behaviors is their spouse (August & Sorkin, 2010). Behavior diffusion theory suggests that spouses can influence each other’s health behaviors directly through mutual reinforcement (Holway et al., 2018; Lewis et al., 2006).

Spousal influences on mealtime behaviors are likely gendered with wives having a greater influence on the husbands’ diet (Conklin et al., 2014) as women are more likely to exercise social control of health behaviors (Kiecolt-Glaser & Newton, 2001; Reczek & Umberson, 2012). Given social norms, wives compared to husbands are more likely to dominate decisions related to food shopping and preparation (Cornelius et al., 2016). For example, married, compared to unmarried, older men have a higher diet quality consisting of fruits, vegetables, vitamins, and fiber (Tucker et al., 1995). Because of a stronger reliance on wives for social control of health behaviors, any changes in health behaviors tend to have stronger influences on husbands’ compared to wives’ mental health (Kiecolt-Glaser & Newton, 2001). Although there has been some work on the concordance of mealtime behaviors, no research to our knowledge has examined whether mealtime behavior is associated with mental health among spouses.

Changes in the health behaviors of one spouse, including diet or other lifestyle changes, have been linked to psychological distress and decline in marital satisfaction (August et al., 2013; Yorgason & Choi, 2016). In contrast, supportive behaviors including sharing meals together and support for change in diet have been linked to positive effects on marital satisfaction and health outcomes (August et al., 2013; Yorgason & Choi, 2016). Unhealthy behaviors, in one or both spouses, are related to lower marital satisfaction (Torvik et al., 2015); and lower marital satisfaction is associated with the development of depressive symptoms (Robles et al., 2014), suggesting that marital satisfaction is a potential mechanism through which mealtime behaviors may influence depressive symptoms. Prior literature shows that wives are more sensitive to marital quality and to emotional transmission and are therefore more likely to experience greater declines in mental health due to marital dissatisfaction compared to husbands (Kiecolt-Glaser & Wilson, 2017). However, it should be noted that marital satisfaction is likely one of the multiple pathways through which mealtime behaviors influence depressive symptoms. As mentioned earlier, there are more direct pathways through which mealtime behaviors affect depressive symptoms, for example through social control. Health behaviors are more likely to transmit from wives to husbands (Holway et al., 2018; Markey et al., 2008; Westmaas et al., 2002) and since wives have been shown to exercise more social control of health behaviors (Holway et al., 2018; Reczek & Umberson, 2012), we expect to see a stronger effect of wives’ mealtime behaviors on husbands’ depressive symptoms.

Among individuals, an association between diet and depression has been confirmed in prospective epidemiologic studies (Ljungberg et al., 2020). Individuals who make unhealthy mealtime choices, such as higher consumption of fast-food meals and processed snacks, are at an increased risk of depression (Akbaraly et al., 2009; Crawford et al., 2011; Fowles et al., 2011; Sánchez-Villegas et al., 2012). In contrast, eating fewer fast-food meals is associated with reduced odds of depression (Liu et al., 2007). Although it is clear that there is an association between the amount of food consumed and depression, findings regarding the exact nature of this association are mixed. Eating too much and not eating enough have both been linked to mental and emotional well-being (Ackard et al., 2003; Polivy, 1996). Among studies examining the reverse association, some suggest that individuals with depression are more likely to engage in overeating due to emotional eating (Ouwens et al., 2009), whereas, others suggest that depressed individuals are more likely to undereat due to loss of appetite (Maxwell & Cole, 2009).

Aside from the biological health influences, the psychological benefit of sharing meals has been a topic of great interest. Eating with others is recognized as a context for social connectedness as it provides opportunities for interaction, communication, and creating a sense of belonging (Fiese et al., 2006). Among adolescents, eating more meals together as a family is associated with reduced rates of depression and suicide attempts (Fulkerson et al., 2007). Eating alone deprives older adults of the opportunity to exchange information and support, and to enjoy intimate interactions (Kuroda et al., 2015). Among older adults, eating alone is associated with higher depressive symptoms for those who live alone as well as those who live with others, indicating that eating alone may be a risk factor for depression (Kuroda et al., 2015; Tani et al., 2015).

Empirical studies and interdependence theory suggest that health behaviors not only affect the health and well-being of adults who directly experience them but also affect the health and well-being of other close individuals such as spouses (Kelley & Thibaut, 1978; Lewis et al., 2006; Roberson et al., 2018). There has been little empirical investigation, however, of the extent to which older adults’ mealtime behaviors affect not just their own but also their partner’s mental health and of the mechanisms of this potential association. In this study, we used the Actor Partner Interdependence Model to examine actor effects (the effect of one’s own meal behaviors on one’s own health) and partner effects (the effects of one’s partner’s meal behaviors on one’s health) on depressive symptoms. We also explored gender differences in these effects and examined marital satisfaction as a potential mechanism of this association. We hypothesized the following:

Hypothesis 1a-d: One’s own number of daily (a) meals, (b) snacks, (c) fast-food meals, and (d) meals eaten alone will be positively correlated with one’s spouse’s number of daily meals, snacks, fast-food meals and meals eaten alone.

Hypothesis 2a-d: Eating fewer total (a) meals each day, and more daily (b) snacks, (c) fast-food meals, and (d) meals eaten alone will be associated with one’s own higher depressive symptoms (actor effects).

Hypothesis 3a-d: One’s spouse eating fewer total (a) meals each day, and more daily (b) snacks, (c) fast-food meals, and (d) meals eaten alone will be associated with one’s own higher depressive symptoms (partner effects).

Hypothesis 4a-d: All actor or partner effects between number of daily (a) meals, (b) snacks, (c) fast-food meals, and (d) meals eaten alone and depressive symptoms will be stronger for husbands than wives.

Hypothesis 5a-d: All actor or partner effects between number of daily (a) meals, (b) snacks, (c) fast food meals, and (d) meals eaten alone and depressive symptoms will be mediated by marital satisfaction.

The conceptual models for each hypothesis are presented in the appendix (Figures A1–A3).

Method

Participants and Procedure

We recruited 101 married couples, 50 years and older, from newspaper advertisements and community bulletins. The aim of the parent experimental study was to examine how spouses support each other with their health concerns and the effect of spousal support on cardiovascular reactivity. Participants had to be in a heterosexual marriage or marriage-like relationship, living together for at least 6 months, and not taking beta-blockers as one of the aims of the parent study was to examine the effect of support on heart rate. Participants completed a background questionnaire at home, before participating in a laboratory session. All measures in this study are from the background questionnaires. The parent study was approved by Yale University’s institutional review board (HIC 1210011003).

Mealtime Behavior Measures

The independent variables in this study were four self-reported mealtime behaviors, including the (1) total number of meals eaten (range: 1 to 4), (2) total number of snacks eaten (range: 0 to 10), (3) number of meals eaten alone (range: 0 to 5), and the (4) number of fast-food meals eaten (range: 0 to 8) on a typical weekday.

Depressive Symptoms

The dependent variable of depressive symptoms was assessed using the 20-item Center for Epidemiologic Studies Depression Scale (CES-D) (Lewinsohn et al., 1997). Participants rated how often over the past week they experienced symptoms associated with depression such as poor appetite, restless sleep, and feelings of fear and loneliness. Response options range from 0 to 3 for each item (0 = rarely or none of the time i.e., less than 1 day; 1 = some or a little of the time i.e., 1–2 days; 2 = occasionally or a moderate amount of time i.e., 3–4 days; 3 = most or all of the time i.e., 5–7 days). The total score ranges from 0 to 60, with higher scores indicating greater depressive symptoms. In determining the CES-D score, we removed the appetite item, “I did not feel like eating; appetite was poor”, from our calculation, and controlled for this item separately in the analysis so that the association between mealtime behaviors and depressive symptoms was not driven by this intrinsic aspect of the outcome. In our sample, the mean CES-D score for husbands was 27.31 (SD = 7.14) and for wives was 27.43 (SD = 7.60).

Marital Satisfaction

Marital satisfaction was measured with the 16-item Locke and Wallace Marital Adjustment Test (MAT) (Jiang et al., 2013; Locke & Wallace, 1959). The test measures several aspects of marital quality including (1) 1 item that captures the overall level of marital happiness on a scale of 1 (very unhappy) to 7 (perfectly happy); (2) 8 items that measure agreement between spouses on different relationship aspects such as handling of family finances, demonstration of affection, and philosophy of life on a scale of 1 (always disagree) to 6 (always agree); and (3) 7 items on how spouses handle disagreements, their engagement in outside interests, and their views of the current partner. Scores range from 26 to 116 in our sample, with higher scores indicative of a higher level of marital satisfaction. For husbands, the mean MAT score was 94.27 (SD = 15.64) and for wives, the mean score was 93.19 (SD = 16.06).

Covariates

We adjusted analyses for potential confounders including age (in years), education (1 = less than high school (n = 1) or high school graduate, 2 = some college, associate’s degree, bachelor’s degree, 3 = some graduate school, professional degree), number of chronic conditions, and whether the couple had any children. Comorbid conditions were assessed using the 24-item Physical Comorbidity Index (Katz et al., 1996) which includes heart disease, stroke, and cancer among other conditions that cause an individual to have difficulty in performing activities of daily living or instrumental activities of daily living. We also adjusted the analyses for the lack of appetite item from the CES-D scale (0 = rarely/none of the time to 3 = most/all of the time).

Statistical Analysis

Participant characteristics were summarized using frequencies and percentages. Paired t-tests comparing mean differences in scores of continuous variables between husbands and wives were conducted. For categorical variables of race and education, Fischer’s exact test was conducted and for the dichotomous variable of whether the couple has any children, McNemar’s chi-square test was conducted to compare differences between husbands and wives. Zero-order correlations between the husbands’ and wives’ mealtime behaviors (H1a-d), depression, covariates, and marital satisfaction were estimated. Covariates that were associated with both mealtime behaviors and depressive symptoms were included in the models if significant at the .10 level.

An Actor-Partner Interdependence model (APIM) (Kenny et al., 2006) was estimated using structural equation modeling, to examine the association between both partners’ mealtime behaviors and their depressive symptoms. APIM estimates two effects: “actor effect”—the effect of an individual’s independent variable on his/her own dependent variable (e.g., the effect of a wife’s diet on the wife’s depressive symptoms)—and “partner effect”—the effect of an individual’s independent variable on his/her partner’s dependent variable (e.g., the effect of a wife’s diet on the husband’s depressive symptoms). Partner effects are labeled in reference to the dyad member outcome. Specifically, we label effects from husbands to wives as the wife partner effect and the effect from wives to husbands as the husband partner effect. Dyads were treated as distinguishable by gender with separate effects estimated for husbands and wives.

To match the ordering of the hypotheses, we first estimated equal actor (H2a-d) and partner (H3a-d) effects in Model 1, assuming there were no differences between husbands and wives. Specifically, we constrained the actor and partner effects to be equal. In a second model (Model 2) we relaxed the assumption of equal actor and partner effects for men and women (H4a-d). We calculated a chi-square difference test between the two models, to determine if Model 2 (df = 4) provided a better fit than Model 1 (df = 6). We also calculated Wald tests to contrast husband and wife actor and partner effects using the TEST statement in Mplus. The covariance between husbands’ and wives’ depressive symptoms was taken into account in Model 2.

Model fit was examined using the chi-square test of model fit, the root mean square error of approximation (RMSEA), the standardized root-mean-square residual (SRMR), the comparative fit index (CFI), and the Tucker-Lewis index (TLI).

To test hypothesis 5, the potential mediating effect of marital satisfaction on the association between mealtime behavior and depressive symptoms, we estimated the actor-partner interdependence mediation model or APIMeM (Ledermann et al., 2011). A mediator effect for marital satisfaction was added to the previously estimated APIM.

The structural equation models were estimated using the full information maximum likelihood (FIML) method, as implemented in Mplus (Version 8) software (Muthén & Muthén, 2017). FIML allows for missing data, which was minimal in our sample. At most 9 couples were missing data: three were missing the complete mealtime data (2.9%), five (4.9%) were missing data on one variable, and one couple (0.99%) was missing data on three items. Statistical significance of model parameters was determined using 95% confidence intervals based on 5,000 bootstrap samples.

Sensitivity Analysis

Given the cross-sectional nature of the data, it is possible that the associations operate in the opposite direction. Therefore, in a sensitivity analysis, we examined whether there were actor and partner associations between depression predicting mealtime behaviors.

Results

Descriptive Analyses

Data from 98 married couples (N = 196) who met the eligibility criteria and had complete data were used in the current analysis. Husbands in our sample were on average 70 years old (SD = 7.66) and wives were 67 years old (SD = 7.14). The majority of the participants were white (92% husbands and 97% wives) and more than two-thirds of husbands and wives had earned a Bachelor’s degree or higher. On average, participants had at least three or more chronic conditions. On average, husbands and wives ate between two and three meals a day, less than two snacks per day, and less than one fast-food meal per day. Both husbands and wives on average ate at least one meal alone each day. Detailed participant characteristics and correlations between study variables are presented in Tables 1 and 2 respectively.

Table 1.

Participant Characteristics

Variable Husbands Wives
N (%)a Range N (%)a Range p-valueb
Mealtime Behaviors
 Number of meals per day 2.72 ± (0.52) 1–4 2.77 ± (0.49) 1–4 0.383
 Number of snacks per day 1.59 ± (1.08) 0–7 1.76 ± (1.39) 0–10 0.300
 Number of meals eaten alone per day 1.14 ± (0.79) 0–3 1.21 ± (0.96) 0–5 0.371
 Number of fast-food meals per day 0.55 ± (1.04) 0–8 0.34 ± (0.91) 0–7 0.097
Depressive symptoms (CES-D score without appetite item) 27.31 ± (7.14) 19–49 27.43 ± (7.60) 18–60 0.888
I did not feel like eating; appetite was poor 1.22 ± (0.55) 1–4 1.16 ± (0.55) 1–4 0.441
Demographic Characteristics
 Age (in years) 70.03 ± (7.66) 56–90 67.40 ± (7.14) 51–89 0.000
 Race 0.484
  White 91 (92) 95 (97)
  Black 1 (1) 0 (0)
  Other 6 (6) 3 (3)
 Education 0.832
  Less than high school 1 (1) 0 (0)
  High school 10 (10) 13 (13)
  Some college credit 15 (15) 13 (13)
  Associate’s degree 7 (7) 8 (8)
  Bachelor’s degree 21 (21) 15 (15)
  Same graduate school 13 (13) 17 (17)
  Professional degree (e.g., PhD, MD) 31 (32) 32 (33)
 Number of chronic conditions 3.36 ± (2.14) 0–9 3.24 ± (2.05) 0–12 0.689
 Have children 76 (75) 73 (72) 0.625
Open in a new tab

Note. CES-D = Center for Epidemiologic Studies Depression Scale.

a

Means and standard deviations (SD) are given for mealtime behaviors, CES-D score, age, and number of chronic conditions

b

p-values comparing mean differences between husbands and wives derived from paired t-tests for continuous variables, from the McNemar test for “have children”, and the Fischer’s exact test for race and education.

Table 2.

Correlations Between Husbands’ and Wives’ Meal Behaviors, Depressive Symptoms, and Covariates

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
1. Wife’s no. of total meals -
2. Wife’s no. of snacks −.10 -
3. Wife’s no. of meals alone −.01 .15 -
4. Wife’s no. of fast-food meals −.05 .12 .12 -
5. Wife’s marital satisfaction .08 −.16 −.06 −.19 -
6. Wife’s age −.02 −.23* −.25* −.06 .14 -
7. Wife’s education level .11 −.06 .21* −.01 .02 .08 -
8. Wife’s chronic conditions −.24* .22* .06 .25* .03 .00 −.05 -
9. Husband’s no. of meals .35* −.06 .09 −.09 .05 −.05 −.06 −.15 -
10. Husband’s no. of snacks .23* −.07 .17 .05 −.23* −.30* .08 .08 −.10 -
11. Husband’s meals alone .02 −.04 .38* .15 −.27* −.07 −.04 −.11 .18 .11 -
12. Husband’s fast-food meals −.23* .08 .20 .33* −.01 −.12 −.21* .45* .01 .08 .05 -
13. Husband’s marital satisfaction −.01 −.08 −.33* −.12 .37* .10 .15 .05 .07 −.22* −.42* −.24* -
14. Husband’s age .05 −.24* −.38* −.11 .18 .79* −.03 −.12 .02 −.22* −.15 −.09 .16 -
15. Husband’s educational level .04 .05 .02 −.17 .05 .15 .48* −.09 −.10 −.20 .03 −.30* .12 .03 -
16. Husband’s chronic conditions .07 .06 −.03 −.02 .05 .07 −.12 .13 .14 .06 .13 .23* −.12 .06 −.07 -
17. Marital length −.11 −.15 −.21* −.07 .11 .50* .04 .01 −.07 −.06 −.26* −.11 .04 .41* .14 −.01 -
18. Any children −.01 .05 .13 −.03 −.23* −.12 .09 −.27* −.00 −.05 .26* −.08 −.20 −.11 −.15 .03 −.43* -
19. Household income −.08 −.03 .10 −.08 .22* .00 .32* −.03 −.03 −.12 −.17 −.22* .20 −.01 .46* −.14 .10 −.07 -
20. Wife’s CES-D score −.34* .21* −.01 .21* −.12 −.17 −.22* .42* −.00 −.03 .02 .39* −.06 −.17 −.17 −.02 .02 −.07 −.05 -
21. Wife’s loss of appetite −.32* .12 −.05 .01 .13 −.14 −.27* .15 −.09 .07 −.01 .21* −.10 .02 −.19 .01 .01 −.02 −.08 .44* -
22. Husband’s CES-D score −.14 .20* .32* .41* −.23* −.16 −.09 .17 −.00 .13 .25* .35* −.39* −.20* −.38* .15 −.20 .35* −.19 .26* .22* -
23. Husband’s loss of appetite −.11 .26* .15 .42* −.21* −.11 .01 .15 −.17 −.05 .02 .18 −.13 −.08 −.10 .10 −.22* .15 −.14 −.01 −.02 .41* -
Open in a new tab

Note. Correlation coefficients shown are Pearson product-moment correlations. Data on marital length, any children, and household income were collected from both husbands and wives but only the husbands’ responses were used in estimating the correlations since both the husbands’ and wives’ reports were identical.

*

p<.05.

Main Hypothesis Testing

Hypothesis 1a-d

As hypothesized, there was a significant correlation between wife’s number of total meals and husband’s number of total meals (r = .35, p = .000); between wife’s number of meals eaten alone and husband’s number of meals eaten alone (r = .38, p = .000); and between wife’s number of fast-food meals and husband’s number of fast-food meals (r = .33, p = .001). The correlation between husbands and wives in the number of snacks consumed was not significant (r = −0.07, p = .523).

Hypothesis 2a-d

We observed an actor effect of fast-food meals such that individuals who ate more fast-food meals reported higher depressive symptoms (β = 1.09, 95% CI: 0.12, 2.29; p = .04), as shown in Table 3-Model 1. The model fit was good by the SRMR and CFI for all meal variables, by the Chi-Square fit for the number of snacks and the number of fast-food meals. But the model fit was poor by the TLI and RMSEA criterion, for all meal variables except fast-food meals.

Table 3.

Structural Equation Model Estimates for Hypothesized Actor and Partner Effects for Mealtime Behaviors Predicting Depressive Symptoms

Model 1: Equal Effects Model 2: Separate Effects
Variable Husbands Wives Contrast
β 95% CI p β 95% CI p β 95% CI p p-value a
Number of meals
  Actor effect −0.94 −3.26, 1.15 .40 0.91 −1.81, 3.56 .51 −3.34 −6.72, −0.11 .05 .03
  Partner effect 0.28 −1.78, 2.22 .78 −1.61 −4.62, 1.14 .26 2.17 −0.47, 4.75 .10 .04
Number of snacks
  Actor effect 0.64 −0.12, 1.73 .18 0.78 −0.37, 2.08 .21 0.48 −0.36, 2.63 .51 .68
  Partner effect 0.26 −0.42, 1.09 .48 0.65 −0.05, 2.30 .24 −0.51 −2.31, 0.76 .53 .11
Number of meals alone
  Actor effect 0.31 −0.63, 1.32 .53 1.04 −0.64, 2.63 .21 −0.19 −1.39, 1.09 .76 .27
  Partner effect 1.30 0.18, 2.28 .02 1.57 0.20, 2.74 .02 0.56 −1.42, 2.53 .57 .37
Number of fast-food meals
  Actor effect 1.09 0.12, 2.29 .04 1.31 −0.17, 2.73 .06 0.74 −1.08, 3.34 .46 .55
  Partner effect 1.53 0.079, 2.64 .02 1.79 −0.23, 3.42 .06 1.29 −0.86, 3.10 .21 .62
Model Fit
Model 1: Equal Actor and Partner Effects Model 2: Separate Effects for Husbands and Wives
SRMR CFI TLI Chi-square, df, p-value RMSEA SRMR CFI TLI Chi-square, df, p-value RMSEA
Number of meals .05 .91 .80 χ2=14.76, df=7, p=.04 .11 .03 .96 .87 χ2=8.63, df=5, p=.13 .09
Number of snacks .04 .94 .86 χ2=12.24, df=7, p=.09 .09 .04 .95 .84 χ2=9.39, df=5, p=.09 .10
Number of meals alone .04 .92 .82 χ2=14.42, df=7, p=.04 .10 .04 .94 .82 χ2=10.40, df=5, p=.13 .10
Number of fast-food meals .03 .98 .97 χ2=8.49, df=7, p=.29 .05 .03 .97 .92 χ2=7.60, df=5, p=.18 .07
Open in a new tab

Note. SRMR = Standardized Mean Square Residual; CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; RMSEA = Root Men Square Error of Approximation.

a

p-values presented for the Wald Test.

Hypothesis 3a-d

As hypothesized, there were significant partner effects between mealtime behavior and depressive symptoms (see Table 3-Model 1). A higher number of meals eaten alone by the partner (β = 1.30, 95% CI: 0.18, 2.28; p = .02), and a higher number of fast-food meals eaten by the partner (β = 1.53, 95% CI: 0.08, 2.64; p = .02), were both associated with more depressive symptoms in the spouse. There was no significant partner effect of the total number of meals or snacks eaten in a day on depressive symptoms. As mentioned under Hypothesis 2a-d, the model fit was good for fast-food meals, but poorer for the other meal variables. This suggested that a model allowing separate effects for husbands and wives might improve fit.

Hypothesis 4a-d

Although under the equal effects assumption there was no actor effect of the total number of meals eaten in a day and depressive symptoms, in model 2, with separate estimates for husbands and wives (see Table 3-Model 2) we found a statistically significant actor effect of the number of meals on depressive symptoms for wives but not husbands, the chi-square difference test between models 1 and 2 was significant (Chi-Square = 6.13, df = 2, p = .047) and model fit was improved. Specifically, wives who ate fewer meals in a day had significantly more depressive symptoms (β = −3.34, 95% CI: −6.72, −0.11; p = .05). The association between the number of meals and depressive symptoms was significantly stronger for wives (β = −3.34 , than husbands versus β = 0.91, Wald Test = 4.87, df = 1, p = .03). There were no significant partner effects for the number of meals.

In regards to the number of snacks and number of meals eaten alone, the chi-square difference tests between Model 1 and Model 2 were not significant (number of snacks: Chi-Square = 2.84, df = 2, p = .242; meals eaten alone: Chi-Square = 4.02, df = 2, p = .134). However, we found a statistically significant partner association for meals eaten alone with depressive symptoms for husbands but not wives, with good indices of fit except for the TLI and RMSEA. Husbands whose wives ate more meals alone in a day (β = 1.57; 95% CI: 0.20, 2.74; p = .02) reported higher depressive symptoms. Lastly, in regards to the fast-food meals measure the chi-square difference test between Model 1 and Model 2 was not significant (Chi-Square = 0.90, df = 2, p = .638) and there were no significant actor or partner effects.

Hypothesis 5a-d

Table A1 presents the total effects, total indirect effects (total IE), specific indirect effects (specific IE), and direct effects for the APIMeM specified for distinguishable dyad members.

In regards to the number of meals, no actor or partner total, indirect, or direct effects were statistically significant for husbands or wives. In regards to the total number of snacks, although the total actor effect of snack consumption was not statistically significant for husbands, the indirect effect of husbands’ snack consumption through husband’s marital satisfaction on husband’s depressive symptoms was statistically significant (βspecific actor IE = 0.39; 95% CI: 0.03, 1.13; p = .12). The direct effect of the husband’s snack consumption on the husband’s depressive symptoms was not significant. Similarly, none of the partner effects of snack consumption on the husband’s depressive symptoms were significant. For wives, the total actor and partner effects of snack consumption were not statistically significant. However, the indirect effect of husbands’ snack consumption through wife’s marital satisfaction on wife’s depressive symptoms was statistically significant (β specific partner IE = 0.35; 95% CI: 0.04, 1.05; p = .13). For wives, none of the actor or partner direct effects were statistically significant.

In regards to the total number of meals eaten alone, although the total actor effect was not statistically significant for husbands, the indirect effect of husband’s meals eaten alone on husband’s depressive symptoms through husband’s marital satisfaction (β specific actor IE = 0.61; 95% CI: 0.03, 1.73; p = .14) was statistically significant. The total partner effect of wife’s meals eaten alone on husband’s depressive symptoms was also statistically significant (β total partner = 1.54; 95% CI: 0.24, 2.75; p = .02). The indirect partner effects were not statistically significant for husbands. For wives, the total, indirect, or direct actor effects were not statistically significant. Although the total partner effect was not significant, the indirect effect of husband’s meals eaten alone on wife’s depressive symptoms through wife’s marital satisfaction (β specific partner IE =0.53; 95% CI: 0.11, 1.48; p = .09) was statistically significant. The total partner effect and the direct partner effect of husband’s meals eaten alone on wife’s depressive symptoms were not statistically significant.

In regards to the number of fast-food meals eaten, none of the actor or partner total, indirect, or direct effects were significant for husbands. For wives, the total actor effect was not statistically significant. However, the specific effect of wives’ number of fast-food meals on wives’ depressive symptoms through wives’ marital satisfaction was statistically significant (β specific actor IE = 0.35; 95% CI: 0.01, 1.31; p = .23). The direct actor effect of wives’ number of fast-food meals on wives’ depressive symptoms was not statistically significant. Similarly, the partner total, indirect, or direct effects of fast-food meals were not significant for wives.

Sensitivity Analysis

In Model 1, individuals and their spouses who reported higher depressive symptoms, ate significantly more fast-food meals, (see Table A2 for detailed results). In addition, there was a partner effect of depression on meals eaten alone. Number of meals and number of snacks did not have significant actor or partner associations with depression. In Model 2, looking at wives and husbands separately, we found that wives whose husbands reported higher depressive symptoms ate more meals alone, and this gender difference was significant.

Discussion

This study examined the correlation between mealtime behaviors of spouses, and the actor and partner associations between mealtime behaviors and depressive symptoms, with depressive symptoms as the outcome. We found that husbands’ and wives’ mealtime behaviors were positively correlated (H1a, c-d). When we assumed equal effects for husbands and wives we found that one’s own (Model 1; H2c) and one’s spouses’ (Model 1; H3c) number of fast-food meals consumed is associated with one’s own depressive symptoms. Additionally, there was a significant positive association between one’s partner’s number of meals eaten alone and one’s own depressive symptoms (Model 1; H3d). When we estimated separate effects for husbands and wives, fewer meals eaten by wives was associated with more depressive symptoms among wives (Model 2; H4a). Furthermore, more meals eaten alone by wives was associated with more depressive symptoms among husbands (Model 2; H4d). There was also evidence of the wife’s marital satisfaction mediating the association between the husband’s meals eaten alone (H5d) and the wife’s depressive symptoms, and the association between the wife’s number of fast-food meals (H5c) and the wife’s depressive symptoms.

A Summary of the Evidence: What Does This Study Establish?

Spouses commonly share health behaviors including wearing seatbelts, smoking patterns, drinking habits, and regularity and quality of meals eaten (Meyler et al., 2007; Rokach, 1998; Stimpson et al., 2006) as a result of shared resources and through the provision of social support or social control. Our findings contribute to the extant literature by showing that husbands’ and wives’ number of meals eaten, meals eaten alone, and fast-food meals eaten are also interdependent (r = .33 to .38); however, contrary to our hypothesis, husbands’ and wives’ number of snacks were not correlated.

As hypothesized, we found an actor effect between the total number of meals and depressive symptoms among wives. Specifically, among wives, eating fewer meals per day—with a range of 1 to 4 meals—was associated with higher depressive symptoms. This corroborates previous findings that suggest that less frequent meal consumption leads to feelings of low energy which in turn can contribute to feelings of depressed mood, whereas adequate meal consumption can avert feelings of depressed mood through feelings of increased energy (Fulkerson et al., 2004). The reverse is also possible: individuals with depression may eat fewer meals because they do not have the motivation or energy to prepare meals and may suffer from loss of appetite. Especially among women, changes in appetite appear to be a more common symptom of depression (Kim et al., 2015). However, results from the models in the alternate direction did not find an association between depressive symptoms and the total number of meals eaten. Furthermore, to account for potential confounding by loss of appetite, we calculated the depressive symptoms score by removing the appetite item and adjusted for the appetite item in all models, and results did not change.

Previous studies suggest that unhealthy dietary choices, such as fast-food and snack consumption, are linked to worse mental health (Crawford et al., 2011; Fowles et al., 2011; Sánchez-Villegas et al., 2012). In line with our hypothesis, we found that individuals who consumed more fast-food meals also reported higher depressive symptoms. However, we did not find an association between snack consumption and depressive symptoms. Results from the sensitivity analysis revealed that it is also possible that individuals with depression may increase their consumption of fast foods due to a lack of motivation and energy to prepare healthy meals.

Additionally, we expected fewer meals eaten alone by an individual to be associated with reduced depression, as sharing meals with others, whether a spouse or other family or friends, provides opportunities for interactions which is associated with better mental health. However, we did not find any actor effects between the number of meals eaten alone and depressive symptoms.

In addition to actor effects, we hypothesized that there would be partner effects such that one’s depressive symptoms would be associated with one’s partner’s mealtime behaviors. We found that individuals whose spouses ate more fast-food meals and more meals alone had higher depressive symptoms. In particular, husbands whose wives ate more meals alone in a day reported higher depressive symptoms. Meals eaten alone by one’s partner may reflect feelings of loneliness and isolation which are associated with depression (Adams et al., 2004; Erzen & Çikrikci, 2018). Additionally, eating more meals alone may reflect poor communication. An alternative explanation for these findings is that higher depressive symptoms in one spouse affect the couple’s mealtime behaviors. Prior research suggests that wives are more likely to be affected by their husbands’ health than vice versa (Kiecolt-Glaser & Wilson, 2017). If the husband is feeling depressed the couple may eat fewer meals together resulting in the wives eating more meals alone. Results from the sensitivity analysis support this hypothesis as wives whose husbands reported higher depressive symptoms ate more meals alone. Although the current study did not collect data on whether meals eaten with others were shared with the spouse or other friends and family, given the demographics of our sample, individuals not eating alone were most likely eating meals with their spouses. Given the average age of the sample (husbands=70 years and wives=67 years) and their overall high functional ability, they were also unlikely to be living with their children.

We also found evidence of marital satisfaction as a potential mediator of the associations between mealtime behaviors and depressive symptoms. Healthy and concordant mealtime behaviors are associated with higher marital satisfaction, whereas, unhealthy mealtime behaviors are associated with lower marital satisfaction (Yorgason & Choi, 2016). Furthermore, lower marital satisfaction is associated with higher depressive symptoms (Robles et al., 2014). As hypothesized, we observed that the husband’s marital satisfaction mediated the association between the husband’s number of meals alone and the husband’s own depressive symptoms. Similarly, the wife’s marital satisfaction mediated the association between the husband’s number of meals eaten alone and wife’s depressive symptoms, and the association between the wife’s number of fast-food meals and the wife’s depressive symptoms.

Strengths and Limitations

The present study has strengths and limitations. It extends previous research examining the influence of mealtime behaviors on adolescents’ and children’s psychological well-being to the context of late-life. The current study also adds to the literature by examining interpersonal, in addition to intrapersonal, associations between mealtime behavior and depressive symptoms. With the examination of marital satisfaction as a mediator, the current study also addresses a potential transactional mechanism for dyadic findings. Another contribution of this study is the examination of gender differences in health behaviors and psychological well-being – it supports existing literature that suggests a stronger influence of wives’ health behaviors on husbands’ health. Finally, the use of a non-clinical sample makes the findings generalizable to a large population of older adults.

The main limitation of the study is its cross-sectional design which prevents us from making any conclusions about the causal ordering of the study variables. Health behaviors of spouses can result from social control, social promotion, or mutual reinforcement of behaviors, as hypothesized in this study. However, health behaviors can also develop as a means of coping with marital stress (Krueger & Chang, 2008; Roberson et al., 2018) or as a result of changes in mental health (Kiecolt-Glaser et al., 2010). There is evidence to suggest that individuals with depression are at risk of undereating, overeating, or eating less nutritiously (Kiecolt-Glaser et al., 2010). Previous literature using longitudinal design suggests that mealtime behavior is an important modifiable risk factor for depression (Kiecolt-Glaser et al., 2015; Li et al., 2017; Molendijk et al., 2018), yet, the possibility that the observed results reflect the effect of depressive symptoms on mealtime behaviors cannot be dismissed. More longitudinal research is needed to establish the direction of association between mealtime behaviors and depressive symptoms. Lastly, there was a lack of sample diversity. Future research should try to replicate these findings using a representative range of racial and ethnic groups, socioeconomic status, and same sex and LGBTQ long-term couples.

Conclusions

Findings, from this study, raise implications for clinical practice. Because previous research suggests that married couples share similar health statuses and display similar risk factors, treatments and changes in health behaviors that focus on the couple are important. Our findings indicate that spouses play an important role in influencing one another’s mealtime behaviors. Additionally, spouses’ mealtime behaviors, in particular fast-food consumption and meals eaten alone, may play a significant role in affecting their partner’s depressive symptoms. Thus, spouses should be involved in interventions that require making changes to one’s diet or other health behaviors.

Funding and Acknowledgements

This work was supported by the National Institute on Aging (K01 AG042450–01 to Dr. Monin, T32AG019134 to Dr. Ali, and P30AG021342 to the Claude D. Pepper Older Americans Independence Center at Yale). Data, analytic methods, and study materials will be made available to other researchers upon request. This study was not preregistered.

Appendix

Figure A.1.

Figure A.1

Open in a new tab

Proposed Actor-Partner Interdependence Model (Equal Effects Assumption).

Note. Actor’s: a_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; a_cesdsum = depression. Partner’s: p_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; p_cesdsum = depression. Boxes reflect observed variables. Hypotheses 2a-d are testing actor effects represented by pathways c’1. Hypotheses 3a-d are testing partner effects represented by pathways c’2.

Figure A.2.

Figure A.2

Open in a new tab

Proposed Actor-Partner Interdependence Model (Separate Effects Assumption Under Hypothesis 4a-d).

Note. Husband’s: h_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; h_cesdsum = depression. Wife’s: w_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; w_cesdsum = depression. Boxes reflect observed variables. Husband actor effect is represented by pathway c’1. Husband partner effect is represented by pathway c’3. Wife partner effect is represented by pathway c’2. Wife actor effect is represented by pathway c’4.

Figure A.3.

Figure A.3

Open in a new tab

Proposed Actor-Partner Interdependence Mediation Model (APIMeM; Hypothesis 5a-d).

Note. Husband’s: h_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; h_marsat = marital satisfaction; h_cesdsum = depression. Wife’s: w_meals = number of total meals, snacks, fast-food meals, and meals eaten alone in a day; w_marsat = marital satisfaction; w_cesdsum = depression. Boxes reflect observed variables. Husband actor effects: Total effect = a1 + b1 + a2 + b3 + c’1; Total indirect effect = a1 + b1 + a2 + b3; Actor-actor simple indirect effect = a1 + b1; Partner-partner simple indirect effect = a2 + b3; Direct effect = c’1. Wife actor effects: Total effect = a4 + b4 + a3 + b2 + c’4; Total indirect effect = a4 + b4 + a3 + b2; Actor-actor simple indirect effect = a4 + b4; Partner-partner simple indirect effect = a3 + b2; Direct effect = c’4. Husband partner effects: Total effect = a4 + b3 + a3 + b1 + c’3; Total indirect effect = a4 + b3 + a3 + b1; Actor-partner simple indirect effect = a4 + b3; Partner-actor simple indirect effect = a3 + b1; Direct effect = c’3. Wife partner effects: Total effect = a1 + b2 + a2 + b4 + c’2; Total indirect effect = a1 + b2 + a2 + b4; Actor-partner simple indirect effect = a1 + b2; Partner-actor simple indirect effect = a2 + b4; Direct effect = c’2.

Table A.1.

Actor-Partner Interdependence Mediation Model (APIMeM) of Mealtime Behaviors Predicting Depression Mediated by Marital Satisfaction

Husbands’ Depressive Symptoms
Variable Number of meals Number of snacks Number of meals alone Number of fast-food meals
Estimate 95% CI p Estimate 95% CI p Estimate 95% CI p Estimate 95% CI p
Total actor effect 0.91 −1.89, 3.60 .52 0.77 −0.39, 2.13 .22 1.10 −0.55, 2.67 .18 1.29 −0.29, 2.59 .06
Total actor indirect effect −0.34 −1.45, 0.59 .50 0.34 −0.12, 1.07 .23 0.63 −0.14, 1.74 .18 0.35 −0.07, 1.64 .44
Specific actor indirect effects
 H_MEALS to H_MARSAT to H_CESD −0.34 −1.40, 0.55 .47 0.39 0.03, 1.13 .12 0.61 0.03, 1.73 .14 0.34 −0.02, 1.62 .43
 H_MEALS to W_MARSAT to H_CESD 0.01 −0.44, 0.58 .98 −0.05 −0.45, 0.27 .77 .02 −0.49, 0.74 .95 0.01 −0.12, 0.26 .94
Actor direct effect 1.25 −1.21, 4.06 .36 0.43 −0.76, 1.65 .50 0.48 −1.35, 2.10 .59 0.94 −0.73, 2.16 .20
Total partner effect −1.66 −4.70, 1.10 .25 0.67 −0.07, 2.37 .25 1.54 0.24, 2.75 .02 1.81 −0.30, 3.39 .05
Total partner indirect effect 0.18 −1.07, 1.24 .76 0.10 −0.18, 0.80 .65 0.30 −0.15, 1.29 .42 0.08 −0.46, 1.11 .84
Specific partner indirect effects
 W_MEALS to W_MARSAT to H_CESD 0.02 −0.33, 0.66 .94 −0.03 −0.53, 0.14 .83 −0.01 −0.24, 0.17 .98 −0.03 −0.61, 0.37 .89
 W_MEALS to H_MARSAT to H_CESD 0.16 −1.07, 1.19 .77 0.13 −0.08, 0.72 .49 0.30 −0.06, 1.32 .40 0.11 −0.21, 1.10 .75
Partner direct effect −1.84 −4.78, 0.65 .18 0.57 −0.08, 2.14 .26 1.24 −0.25, 2.73 .10 1.73 −0.68, 3.08 .09
Wives’ Depressive Symptoms
Variable Number of meals Number of snacks Number of meals alone Number of fast-food meals
Estimate 95% CI p Estimate 95% CI p Estimate 95% CI p Estimate 95% CI p
Total actor effect −3.13 −6.64, 0.11 .07 0.45 −0.39, 2.57 .54 −0.14 −1.34, 1.12 .82 0.77 −0.99, 3.41 .45
Total actor indirect effect −0.20 −1.45, 0.40 .67 0.19 −0.02, 0.73 .28 −0.13 −0.90, 0.34 .67 0.30 −0.07, 1.08 .25
Specific actor indirect effects
 W_MEALS to W_MARSAT to W_CESD −0.19 −1.28, 0.30 .62 0.20 −0.01, 0.82 .28 −0.07 −0.64, 0.18 .72 0.35 0.01, 1.31 .23
 W_MEALS to H_MARSAT to W_CESD −0.01 −0.50, 0.26 .97 −0.01 −0.34, 0.09 .90 −0.07 −0.84, 0.18 .78 −0.05 −1.01, 0.08 .81
Actor direct effect −2.94 −6.34, 0.32 .08 0.25 −0.51, 2.41 .72 −0.01 −1.34, 1.53 .99 0.47 −1.48, 3.09 .64
Total partner effect 2.16 −0.48, 4.72 .10 −0.53 −2.35, 0.76 .52 0.52 −1.48, 2.54 .61 1.25 −0.97, 3.05 .23
Total partner indirect effect −0.05 −0.81, 0.92 .91 0.32 −0.06, 0.87 .17 0.40 −0.43, 1.24 .35 −0.23 −1.51, 0.16 .52
Specific partner indirect effects
 H_MEALS to H_MARSAT to W_CESD 0.02 −0.32, 0.80 .95 −0.03 −0.59, 0.23 .86 −0.14 −1.43, 0.39 .74 −0.15 −1.38, 0.11 .59
 H_MEALS to W_MARSAT to W_CESD −0.07 −1.36, 0.50 .88 0.35 0.04, 1.05 .13 0.53 0.11, 1.48 .09 −0.07 −0.50, .14 .63
Partner direct effect 2.21 −0.66, 4.79 .11 −0.84 −2.78, 0.56 .34 0.12 −1.96, 2.02 .91 1.48 −0.62, 3.18 .14
Model Fit
SRMR CFI TLI Chi-square, df, p-value RMSEA
Number of meals .06 .91 .79 χ2=25.65, df=15, p=.04 .08
Number of snacks .06 .90 .77 χ2=27.25, df=15, p=.03 .09
Number of meals alone .06 .93 .83 χ2=25.01, df=15, p=.05 .08
Number of fast-food meals .06 .94 .86 χ2=23.00, df=15, p=.08 .07
Open in a new tab

Note. H_MEALS = husband’s mealtime behaviors; H_MARSAT = husband’s marital satisfaction; H_CESD = husband’s depressive symptoms; W_MEALS = wife’s mealtime behaviors; W_MARSAT = wife’s marital satisfaction; W_CESD = wife’s depressive symptoms.

SRMR = Standardized Root-Mean-Square Residual; CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; RMSEA = Root Men Square Error of Approximation.

Table A.2.

Estimates for Structural Equation Models: Hypothesized Actor and Partner Effects of Depression Perdicting Mealtime Behaviors

Model 1: Equal Effects Model 2: Separate Effects
Variable Husbands Wives Contrast
β 95% CI p β 95% CI p β 95% CI p p-value a
Number of meals
  Actor effect −0.01 −0.02, 0.01 .24 0.002 −0.02, 0.02 .81 −0.01 −0.03, 0.002 .10 .16
  Partner effect −0.002 −0.01, 0.01 .72 −0.002 −0.02, 0.02 .82 −0.002 −0.02, 0.01 .74 .98
Number of snacks
  Actor effect 0.02 −0.01, 0.05 .13 0.03 −0.01, 0.07 .16 0.02 −0.04, 0.06 .47 .66
  Partner effect −0.001 −0.03, 0.03 .97 −0.02 −0.04, 0.02 .24 0.03 −0.02, 0.08 .24 .06
Number of meals alone
  Actor effect 0.01 −0.01, 0.03 .35 0.03 −0.001, 0.06 .06 −0.01 −0.04, 0.02 .46 .06
  Partner effect 0.02 0.003, 0.03 .03 −0.003 −0.02, 0.02 .75 0.05 0.02, 0.07 .001 .01
Number of fast-food meals
  Actor effect 0.02 0.004, 0.05 .04 0.03 0.006, 0.09 .08 0.01 −0.01, 0.04 .30 .34
  Partner effect 0.04 0.01, 0.09 .03 0.04 −0.01, 0.09 .17 0.05 0.01, 0.11 0.06 .55
Model Fit
Model 1: Equal Actor and Partner Effects Model 2: Separate Effects for Husbands and Wives
SRMR CFI TLI Chi-square, df, p-value RMSEA SRMR CFI TLI Chi-square, df, p-value RMSEA
Number of meals .03 .94 .88 χ2=8.37, df=7, p=.30 .05 .03 .96 .90 χ2=5.85, df=5, p=32 .04
Number of snacks .04 .55 .04 χ2=11.36, df=7, p=.12 .08 .03 .70 .10 χ2=7.94, df=5, p=.16 .08
Number of meals alone .05 .80 .58 χ2=11.42, df=7, p=.12 .08 .02 1.0 1.0 χ2=3.36, df=5, p=.64 .00
Number of fast-food meals .05 .75 .47 χ2=22.45, df=7, p=.002 .15 .05 .74 .20 χ2=21.42, df=5, p=.001 .18
Open in a new tab

Note. SRMR = Standardized Mean Square Residual; CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; RMSEA = Root Men Square

Error of Approximation.

a

p-values presented for the Wald Test

Footnotes

Conflict of Interest: The authors declare that they have no conflict of interest.

References

  1. Ackard DM, Neumark-Sztainer D, Story M, & Perry C (2003). Overeating among adolescents: prevalence and associations with weight-related characteristics and psychological health. Pediatrics, 111(1), 67–74. 10.1542/peds.111.1.67 [DOI] [PubMed] [Google Scholar]
  2. Adams KB, Sanders S, & Auth E (2004). Loneliness and depression in independent living retirement communities: risk and resilience factors. Aging & mental health, 8(6), 475–485. 10.1080/13607860410001725054 [DOI] [PubMed] [Google Scholar]
  3. Akbaraly TN, Brunner EJ, Ferrie JE, Marmot MG, Kivimaki M, & Singh-Manoux A (2009). Dietary pattern and depressive symptoms in middle age. The British journal of psychiatry : the journal of mental science, 195(5), 408–413. 10.1192/bjp.bp.108.058925 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. August KJ, Rook KS, Franks MM, & Stephens MAP (2013). Spouses’ involvement in their partners’ diabetes management: Associations with spouse stress and perceived marital quality. Journal of Family Psychology, 27(5), 712. 10.1037/a0034181 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. August KJ, & Sorkin DH (2010). Marital status and gender differences in managing a chronic illness: The function of health-related social control. Social science & medicine, 71(10), 1831–1838. 10.1016/j.socscimed.2010.08.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barrett-Connor E, Suarez L, & Criqui MH (1982). Spouse concordance of plasma cholesterol and triglyceride. Journal of Chronic Diseases, 35(5), 333–340. 10.1016/0021-9681(82)90004-2 [DOI] [PubMed] [Google Scholar]
  7. Conklin AI, Forouhi NG, Surtees P, Khaw K-T, Wareham NJ, & Monsivais P (2014). Social relationships and healthful dietary behaviour: Evidence from over-50s in the EPIC cohort, UK. Social science & medicine, 100, 167–175. 10.1016/j.socscimed.2013.08.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cornelius T, Desrosiers A, & Kershaw T (2016). Spread of health behaviors in young couples: how relationship power shapes relational influence. Social science & medicine, 165, 46–55. 10.1016/j.socscimed.2016.07.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Crawford GB, Khedkar A, Flaws JA, Sorkin JD, & Gallicchio L (2011). Depressive symptoms and self-reported fast-food intake in midlife women. Preventive medicine, 52(3–4), 254–257. 10.1016/j.ypmed.2011.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Erzen E, & Çikrikci Ö (2018). The effect of loneliness on depression: A meta-analysis. International Journal of Social Psychiatry, 64(5), 427–435. 10.1177/0020764018776349 [DOI] [PubMed] [Google Scholar]
  11. Fiese BH, Foley KP, & Spagnola M (2006). Routine and ritual elements in family mealtimes: Contexts for child well-being and family identity. New directions for child and adolescent development, 2006(111), 67–89. 10.1002/cd.156 [DOI] [PubMed] [Google Scholar]
  12. Fowles ER, Timmerman GM, Bryant M, & Kim S (2011). Eating at fast-food restaurants and dietary quality in low-income pregnant women. Western journal of nursing research, 33(5), 630–651. 10.1177/0193945910389083 [DOI] [PubMed] [Google Scholar]
  13. Franks MM, Sahin ZS, Seidel AJ, Shields CG, Oates SK, & Boushey CJ (2012). Table for two: Diabetes distress and diet-related interactions of married patients with diabetes and their spouses. Families, Systems, & Health, 30(2), 154. 10.1037/a0028614 [DOI] [PubMed] [Google Scholar]
  14. Fulkerson JA, Sherwood NE, Perry CL, Neumark-Sztainer D, & Story M (2004). Depressive symptoms and adolescent eating and health behaviors: a multifaceted view in a population-based sample. Preventive medicine, 38(6), 865–875. 10.1016/j.ypmed.2003.12.028 [DOI] [PubMed] [Google Scholar]
  15. Fulkerson JA, Strauss J, Neumark-Sztainer D, Story M, & Boutelle K (2007). Correlates of psychosocial well-being among overweight adolescents: the role of the family. Journal of consulting and clinical psychology, 75(1), 181. 10.1037/0022-006X.75.1.181 [DOI] [PubMed] [Google Scholar]
  16. Hoang D, Kristoffersen I, & Li IW (2019). All in the mind? Estimating the effect of mental health on health behaviours. Social science & medicine, 225, 69–84. 10.1016/j.socscimed.2019.02.017 [DOI] [PubMed] [Google Scholar]
  17. Holway GV, Umberson D, & Donnelly R (2018). Health and health behavior concordance between spouses in same-sex and different-sex marriages. Social currents, 5(4), 319–327. 10.1177/2329496517734570 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jiang Y, Terhorst L, Donovan HS, Weimer JM, Choi C-WJ, Schulz R, . . . Sherwood PR (2013). Locke-Wallace Short Marital-Adjustment Test: psychometric evaluation in caregivers for persons with primary malignant brain tumor. Journal of nursing measurement, 21(3), 502–515. 10.1891/1061-3749.21.3.502 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Katz JN, Chang LC, Sangha O, Fossel AH, & Bates DW (1996). Can comorbidity be measured by questionnaire rather than medical record review? Medical care, 73–84. 10.1097/00005650-199601000-00006 [DOI] [PubMed] [Google Scholar]
  20. Kelley HH, & Thibaut JW (1978). Interpersonal relations: A theory of interdependence. John Wiley & Sons. [Google Scholar]
  21. Kenny DA, Kashy DA, & Cook WL (2006). Dyadic data analysis. Guilford press. [Google Scholar]
  22. Kiecolt-Glaser JK, Derry HM, & Fagundes CP (2015). Inflammation: depression fans the flames and feasts on the heat. American Journal of Psychiatry, 172(11), 1075–1091. 10.1176/appi.ajp.2015.15020152 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kiecolt-Glaser JK, Gouin J-P, & Hantsoo L (2010). Close relationships, inflammation, and health. Neuroscience & biobehavioral reviews, 35(1), 33–38. 10.1016/j.neubiorev.2009.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kiecolt-Glaser JK, & Newton TL (2001). Marriage and health: his and hers. Psychological bulletin, 127(4), 472. 10.1037/0033-2909.127.4.472 [DOI] [PubMed] [Google Scholar]
  25. Kiecolt-Glaser JK, & Wilson SJ (2017). Lovesick: How couples’ relationships influence health. Annual review of clinical psychology, 13, 421–443. 10.1146/annurev-clinpsy-032816-045111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kim HC, Kang DR, Choi KS, Nam CM, Thomas GN, & Suh I (2006). Spousal concordance of metabolic syndrome in 3141 Korean couples: a nationwide survey. Annals of epidemiology, 16(4), 292–298. 10.1016/j.annepidem.2005.07.052 [DOI] [PubMed] [Google Scholar]
  27. Kim J, Cho MJ, Hong JP, Bae JN, Cho S-J, Hahm B-J, . . . Jeon HJ (2015). Gender differences in depressive symptom profile: results from nationwide general population surveys in Korea. Journal of Korean medical science, 30(11), 1659–1666. 10.3346/jkms.2015.30.11.1659 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Krueger PM, & Chang VW (2008). Being poor and coping with stress: health behaviors and the risk of death. American journal of public health, 98(5), 889–896. 10.2105/AJPH.2007.114454 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kuroda A, Tanaka T, Hirano H, Ohara Y, Kikutani T, Furuya H, . . . Iijima K (2015). Eating Alone as Social Disengagement is Strongly Associated With Depressive Symptoms in Japanese Community-Dwelling Older Adults. Journal of the American Medical Directors Association, 16(7), 578–585. 10.1016/j.jamda.2015.01.078 [DOI] [PubMed] [Google Scholar]
  30. Ledermann T, Macho S, & Kenny DA (2011). Assessing mediation in dyadic data using the actor-partner interdependence model. Structural Equation Modeling: A Multidisciplinary Journal, 18(4), 595–612. 10.1080/10705511.2011.607099 [DOI] [Google Scholar]
  31. Lewinsohn PM, Seeley JR, Roberts RE, & Allen NB (1997). Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychology and aging, 12(2), 277. 10.1037/0882-7974.12.2.277 [DOI] [PubMed] [Google Scholar]
  32. Lewis MA, McBride CM, Pollak KI, Puleo E, Butterfield RM, & Emmons KM (2006). Understanding health behavior change among couples: An interdependence and communal coping approach. Social science & medicine, 62(6), 1369–1380. 10.1016/j.socscimed.2005.08.006 [DOI] [PubMed] [Google Scholar]
  33. Li Y, Lv M-R, Wei Y-J, Sun L, Zhang J-X, Zhang H-G, & Li B (2017). Dietary patterns and depression risk: a meta-analysis. Psychiatry research, 253, 373–382. 10.1016/j.psychres.2017.04.020 [DOI] [PubMed] [Google Scholar]
  34. Liu C, Xie B, Chou C-P, Koprowski C, Zhou D, Palmer P, . . . Sun X (2007). Perceived stress, depression and food consumption frequency in the college students of China Seven Cities. Physiology & behavior, 92(4), 748–754. 10.1016/j.physbeh.2007.05.068 [DOI] [PubMed] [Google Scholar]
  35. Ljungberg T, Bondza E, & Lethin C (2020). Evidence of the Importance of Dietary Habits Regarding Depressive Symptoms and Depression. International journal of environmental research and public health, 17(5), 1616. 10.3390/ijerph17051616 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Locke HJ, & Wallace KM (1959). Short marital-adjustment and prediction tests: Their reliability and validity. Marriage and family living, 21(3), 251–255. 10.2307/348022 [DOI] [Google Scholar]
  37. Macario E, & Sorensen G (1998). Spousal similarities in fruit and vegetable consumption. American Journal of Health Promotion, 12(6), 369–377. 10.4278/0890-1171-12.6.369 [DOI] [PubMed] [Google Scholar]
  38. Markey CN, Gomel JN, & Markey PM (2008). Romantic relationships and eating regulation: An investigation of partners’ attempts to control each others’ eating behaviors. Journal of Health Psychology, 13(3), 422–432. 10.1177/1359105307088145 [DOI] [PubMed] [Google Scholar]
  39. Maxwell MA, & Cole DA (2009). Weight change and appetite disturbance as symptoms of adolescent depression: Toward an integrative biopsychosocial model. Clinical Psychology Review, 29(3), 260–273. 10.1016/j.cpr.2009.01.007 [DOI] [PubMed] [Google Scholar]
  40. Meyler D, Stimpson JP, & Peek MK (2007). Health concordance within couples: a systematic review. Social science & medicine, 64(11), 2297–2310. 10.1016/j.socscimed.2007.02.007 [DOI] [PubMed] [Google Scholar]
  41. Molendijk M, Molero P, Ortuño Sánchez-Pedreño F, Van der Does W, & Angel Martínez-González M (2018). Diet quality and depression risk: A systematic review and dose-response meta-analysis of prospective studies. Journal of affective disorders, 226, 346–354. 10.1016/j.jad.2017.09.022 [DOI] [PubMed] [Google Scholar]
  42. Muthén L, & Muthén B (2017). Mplus (Version 8)[computer software].(1998–2017). Los Angeles, CA: Muthén & Muthén. [Google Scholar]
  43. Ouwens MA, van Strien T, & van Leeuwe JFJ (2009). Possible pathways between depression, emotional and external eating. A structural equation model. Appetite, 53(2), 245–248. 10.1016/j.appet.2009.06.001 [DOI] [PubMed] [Google Scholar]
  44. Pachucki MA, Jacques PF, & Christakis NA (2011). Social network concordance in food choice among spouses, friends, and siblings. American journal of public health, 101(11), 2170–2177. 10.2105/AJPH.2011.300282 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Peek MK, & Markides KS (2003). Blood pressure concordance in older married Mexican-American couples. Journal of the American Geriatrics Society, 51(11), 1655–1659. 10.1046/j.1532-5415.2003.51520.x [DOI] [PubMed] [Google Scholar]
  46. Peek MK, Stimpson JP, Townsend AL, & Markides KS (2006). Well-being in older Mexican American spouses. The Gerontologist, 46(2), 258–265. 10.1093/geront/46.2.258 [DOI] [PubMed] [Google Scholar]
  47. Polivy J (1996). Psychological Consequences of Food Restriction. Journal of the American Dietetic Association, 96(6), 589–592. 10.1016/S0002-8223(96)00161-7 [DOI] [PubMed] [Google Scholar]
  48. Reczek C, & Umberson D (2012). Gender, health behavior, and intimate relationships: Lesbian, gay, and straight contexts. Social science & medicine, 74(11), 1783–1790. 10.1016/j.socscimed.2011.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Roberson PNE, Shorter RL, Woods S, & Priest J (2018). How health behaviors link romantic relationship dysfunction and physical health across 20 years for middle-aged and older adults. Social science & medicine, 201, 18–26. 10.1016/j.socscimed.2018.01.037 [DOI] [PubMed] [Google Scholar]
  50. Robles TF, Slatcher RB, Trombello JM, & McGinn MM (2014). Marital quality and health: A meta-analytic review. Psychological bulletin, 140(1), 140. 10.1037/a0031859 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rokach A (1998). Loneliness in singlehood and marriage. Psychology, 35(2), 2–17. 10.1521/jscp.1998.17.1.75 [DOI] [Google Scholar]
  52. Sánchez-Villegas A, Toledo E, de Irala J, Ruiz-Canela M, Pla-Vidal J, & Martínez-González MA (2012). Fast-food and commercial baked goods consumption and the risk of depression. Public Health Nutrition, 15(3), 424–432. 10.1017/S1368980011001856 [DOI] [PubMed] [Google Scholar]
  53. Smith KR, & Zick CD (1994). Linked lives, dependent demise? Survival analysis of husbands and wives. Demography, 31(1), 81–93. 10.2307/2061909 [DOI] [PubMed] [Google Scholar]
  54. Stimpson JP, Masel MC, Rudkin L, & Peek MK (2006). Shared health behaviors among older Mexican American spouses. American journal of health behavior, 30(5), 495–502. 10.5993/AJHB.30.5.6 [DOI] [PubMed] [Google Scholar]
  55. Tani Y, Sasaki Y, Haseda M, Kondo K, & Kondo N (2015). Eating alone and depression in older men and women by cohabitation status: The JAGES longitudinal survey. Age and Ageing, 44(6), 1019–1026. 10.1093/ageing/afv145 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Torvik FA, Gustavson K, Røysamb E, & Tambs K (2015). Health, health behaviors, and health dissimilarities predict divorce: results from the HUNT study. BMC Psychology, 3(1), 13. 10.1186/s40359-015-0072-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Tucker K, Spiro A, & Weiss ST (1995). Variation in food and nutrient intakes among older men: Age, and other socio-demographic factors. Nutrition Research, 15(2), 161–176. 10.1016/0271-5317(95)92582-5 [DOI] [Google Scholar]
  58. Umberson D (1992). Gender, marital status and the social control of health behavior. Social science & medicine, 34(8), 907–917. 10.1016/0277-9536(92)90259-S [DOI] [PubMed] [Google Scholar]
  59. Waite LJ (1995). Does marriage matter? Demography, 32(4), 483–507. 10.2307/2061670 [DOI] [PubMed] [Google Scholar]
  60. Westmaas JL, Wild TC, & Ferrence R (2002). Effects of gender in social control of smoking cessation. Health psychology, 21(4), 368. 10.1037/0278-6133.21.4.368 [DOI] [PubMed] [Google Scholar]
  61. Yorgason JB, & Choi H (2016). Health contributions to marital quality: Expected and unexpected links. 10.1037/14897-010 [DOI]

RESOURCES