Advertisement

Eating Alone as Social Disengagement is Strongly Associated With Depressive Symptoms in Japanese Community-Dwelling Older Adults

Published:February 14, 2015DOI:https://doi.org/10.1016/j.jamda.2015.01.078

      Abstract

      Objectives

      Depression in later life poses a grave challenge for the aging countries. The reported key risk factors include social disengagement, but the lack of social companionship during mealtimes, namely eating alone, has not been examined extensively, especially in relation to living arrangement. Past studies on changes along geriatric trajectories in the association between social engagement and depression also remain inadequate. This study aims to examine the association between social engagement and depressive symptoms with a particular focus on eating alone and how the association changes along the aging and mental frailty trajectories.

      Design

      A cross-sectional study.

      Setting

      Kashiwa-city, Chiba-prefecture in Japan.

      Participants

      A total of 1856 community-dwelling older adults.

      Measurements

      The 15-item Geriatric Depression Scale was used to measure depressive symptoms. The indicators used to assess social engagement included eating alone, living arrangement, reciprocity of social support, social participation, social stressors and social ties.

      Results

      Social engagement was significantly associated with depressive symptoms. Those who live with their families yet eat alone were found to be at particular risk (odds ratio = 5.02, 95% confidence interval 2.5–9.9 for young-old; odds ratio = 2.41, 95% confidence interval 1.2–4.8 for old-old). Younger and less mentally frail populations showed stronger associations.

      Conclusions

      Eating alone was a key risk factor for depressive symptoms in community-dwelling older adults. The living arrangement in which they eat alone is important in identifying those with the greatest risk. Mental health management for older adults requires comprehensive assessment of their social relations that takes into account their companionship during mealtimes. Social preventive measures need to involve early interventions in order to augment their effectiveness against mental frailty.

      Keywords

      The problem of depression in later life has become a pressing global concern, as the population aging continues worldwide.
      • Djernes J.K.
      Prevalence and predictors of depression in populations of elderly: A review.
      It undermines well-being and quality of life while adding to healthcare costs, with potential consequences on a wide range of health outcomes.
      • Schwarzbach M.
      • Luppa M.
      • Forstmeier S.
      • et al.
      Social relations and depression in late life-A systematic review.
      The problem poses a grave socioeconomic burden on aging countries, not least in Japan where the unprecedented level of aging threatens to undermine its social security system.
      • Muramatsu N.
      • Akiyama H.
      Japan: Super-aging society preparing for the future.
      The prevalence of depression among community-dwelling older adults varies enormously and has been reported to be as high as 35%.
      • Wada T.
      • Ishine M.
      • Sakagami T.
      • et al.
      Depression in Japanese community-dwelling elderly—Prevalence and association with ADL and QOL.
      The key reported predictors of depressive symptoms include female gender, cognitive and functional impairments, medical disorders, low level of education, and social disengagement.
      • Djernes J.K.
      Prevalence and predictors of depression in populations of elderly: A review.
      • Blazer D.G.
      • Hybels C.F.
      Origins of depression in later life.
      • Park N.S.
      • Jang Y.
      • Lee B.S.
      • et al.
      The mediating role of loneliness in the relation between social engagement and depressive symptoms among older Korean Americans: Do men and women differ?.
      • Beekman A.T.F.
      • Deeg D.J.H.
      • vanTilburg T.
      • et al.
      Major and minor depression in later life: A study of prevalence and risk factors.
      • Fukunaga R.
      • Abe Y.
      • Nakagawa Y.
      • et al.
      Living alone is associated with depression among the elderly in a rural community in Japan.
      • Kawachi I.
      • Berkman L.F.
      Social ties and mental health.
      • Heun R.
      • Hein S.
      Risk factors of major depression in the elderly.
      Social engagement is an “umbrella concept for the various components of an individual's social behavior and social structure”
      • Mendes de Leon C.F.
      Social engagement and successful aging.
      and its different aspects have consistently been found to predict mortality, disease outcomes, disability, cognitive decline as well as depressive symptoms.
      • Park N.S.
      The relationship of social engagement to psychological well-being of older adults in assisted living facilities.
      • Tilvis R.S.
      • Routasalo P.
      • Karppinen H.
      • et al.
      Social isolation, social activity and loneliness as survival indicators in old age; A nationwide survey with a 7-year follow-up.
      • Tomaka J.
      • Thompson S.
      • Palacios R.
      The relation of social isolation, loneliness, and social support to disease outcomes among the elderly.
      • Avlund K.
      • Lund R.
      • Holstein B.E.
      • Due P.
      Social relations as determinant of onset of disability in aging.
      • Bassuk S.S.
      • Glass T.A.
      • Berkman L.F.
      Social disengagement and incident cognitive decline in community-dwelling elderly persons.
      While the conceptualization of social engagement lacks a strong consensus,
      • Bath P.A.
      • Deeg D.
      Social engagement and health outcomes among older people: Introduction to a special section.
      this should not be viewed as a weakness but as an invitation to explore its unexamined aspects in a search for the most relevant screening questions to identify older adults at risk.
      • Mendes de Leon C.F.
      Social engagement and successful aging.
      This study, thus, aims to examine new concepts and ideas that remain under-explored, especially in relation to depression.
      One such aspect is the social behavior during mealtimes. Commensality (ie, the act of eating with others) provides opportunities for social interactions and exchange of information and support by facilitating participation in shared social activities of mealtimes.
      • Vesnaver E.
      • Keller H.H.
      Social influences and eating behavior in later life: A review.
      Eating alone deprives older adults such valuable social opportunities. Eating alone has been studied in relation to dietary intake, but research in relation to depression and wider health outcomes remains limited.
      • Kimura Y.
      • Wada T.
      • Okumiya K.
      • et al.
      Eating alone among community-dwelling Japanese elderly: Association with depression and food diversity.
      To our knowledge, none has examined its association with depression in combination with other components of social engagement nor investigated it in relation to the living arrangement. Living alone is often cited as a key risk factor for older adults, as does the Ministry of Health, Labor and Welfare, Japan, but eating alone is rarely discussed. A shared living arrangement may result in increased opportunities for commensality, but does not guarantee it,
      • Vesnaver E.
      • Keller H.H.
      Social influences and eating behavior in later life: A review.
      • Kimura Y.
      • Wada T.
      • Okumiya K.
      • et al.
      Eating alone among community-dwelling Japanese elderly: Association with depression and food diversity.
      requiring independent considerations.
      Furthermore, past studies have not adequately examined how the association between social engagement and depression changes along geriatric trajectories such as aging and frailty. Frailty is not only a physical but a multidimensional concept,
      • Markle-Reid M.
      • Browne G.
      Conceptualizations of frailty in relation to older adults.
      and mental frailty, one important dimension, may manifest as depressive states. The role of social engagement vis-à-vis depression is expected to change as older adults age or become more mentally frail, influencing the effectiveness of social intervention measures.
      The purpose of the present study is 2-fold. The first objective is to examine whether social engagement is associated with depressive symptoms with a particular focus on eating alone and its relation to the living arrangement. Second, effects of geriatric trajectories, namely aging and mental frailty trajectory on the above association, are examined in order to better identify the most effective social intervention sites for depressive symptoms.

      Methods

      Study Design

      The study was cross-sectional.

      Setting and Participants

      This study was based on data from 1856 randomly selected community-dwelling older adults (independent or those requiring support), aged 65–94, who participated in the first year health assessment of a 3-year cohort study between 2012 and 2014 in Kashiwa city, Japan. A total of 2044 persons participated in the assessment and 188 persons were excluded due to missing items of data.

      Measurements

      Depressive symptoms

      The 15-item Geriatric Depression Scale (GDS) was used. Scores of ≥6 were defined as “depressive symptoms,”
      • Schreiner A.S.
      • Hayakawa H.
      • Morimoto T.
      • Kakuma T.
      Screening for late life depression: Cut-off scores for the Geriatric Depression Scale and the Cornell Scale for Depression in Dementia among Japanese subjects.
      6–9 as “mild depression,” and ≥10 as “severe depression.”
      • Wada T.
      • Ishine M.
      • Sakagami T.
      • et al.
      Depression in Japanese community-dwelling elderly—Prevalence and association with ADL and QOL.

      Social engagement

      Seven components were assessed: (1) living arrangement; (2) eating arrangement; (3) reciprocal social support; (4) social participation; (5) social stressors; (6) social ties with family; and (7) social ties with friends. The following questions were asked regarding each item: (1) Do you live with your family: yes or no? (No = living alone); (2) Do you eat your meals with anyone else, at least once a day: yes or no? (No = eating alone); (3) Do you give advice and a helping hand to your family or friends: yes or no? (No = low reciprocal social support); (4) Are you going out less frequently compared to last year: yes or no? (The Kihon Check List, Ministry of Health, Labor and Welfare) (No = fewer frequency of going out); and (5) Did you experience any major changes in life in the past year, such as moving home, retirement, loss of relatives, financial troubles, troubles in the relationships with people: yes or no? (Yes = major change in life). For (6) and (7), the abbreviated Lubben Social Network Scale-6 and its Family and Friends subscales
      • Crooks V.C.
      • Lubben J.
      • Petitti D.B.
      • et al.
      Social network, cognitive function, and dementia incidence among elderly women.
      • Lubben J.
      • Blozik E.
      • Gillmann G.
      • et al.
      Performance of an abbreviated version of the Lubben Social Network Scale among three European community-dwelling older adult populations.
      were used. Living arrangement and eating arrangement were crossed to construct 4 dummy variables: “living and eating alone,” “living alone yet eating with others,” “living with others yet eating alone,” and “living and eating with others” (reference).

      Sociodemographic variables

      Age and the years of education were included in the analysis as continuous variables. Health literacy was measured by 5 items developed for Japanese persons.
      • Ishikawa H.
      • Nomura K.
      • Sato M.
      • Yano E.
      Developing a measure of communicative and critical health literacy: A pilot study of Japanese office workers.
      Information on economic status was obtained as income ranking based on long-term care insurance premiums. Logistic regression was performed with the income ranking and depressive symptoms as the independent and dependent variables, respectively. The odds ratios were plotted to observe changes in the trend and those with less than 1.4 million Japanese Yen per person were categorized as the “low income” group.

      Medical histories

      Medical histories of hypertension, osteoporosis, cerebrovascular diseases, diabetes, heart diseases, and malignant neoplasm were obtained through medical interviews by nurses.

      Number of medications

      The total number of oral medications was recorded as a continuous variable, as polypharmacy is known to be associated with increased depressive symptoms.
      • Liu C.P.
      • Leung D.S.
      • Chi I.
      Social functioning, polypharmacy and depression in older Chinese primary care patients.

      Physical health and functions

      Instrumental activities of daily living (IADL) was measured using the Tokyo Metropolitan Institute of Gerontology index of competence.
      • Koyano W.
      • Shibata H.
      • Nakazato K.
      • et al.
      Measurement of competence—Reliability and validity of the TMIG index of competence.
      Mobility was assessed by Life-Space Assessment,
      • Baker P.S.
      • Bodner E.V.
      • Allman R.M.
      Measuring life-space mobility in community-dwelling older adults.
      • Peel C.
      • Baker P.S.
      • Roth D.L.
      • et al.
      Assessing mobility in older adults: The UAB Study of Aging Life-Space Assessment.
      measured with the Elderly-Status Assessment Set.
      • Shimada H.
      • Sawyer P.
      • Harada K.
      • et al.
      Predictive validity of the classification schema for functional mobility tests in instrumental activities of daily living decline among older adults.
      • Shimada H.
      • Ishizaki T.
      • Kato M.
      • et al.
      How often and how far do frail elderly people need to go outdoors to maintain functional capacity?.
      The highest level of life-spaces (level 5) was used. To assess usual and maximum gait speeds, participants were instructed to walk over an 11-m course and the time spent in the middle 5 m was recorded.
      • Ishii S.
      • Tanaka T.
      • Shibasaki K.
      • et al.
      Development of a simple screening test for sarcopenia in older adults.

      Cognitive function

      The Mini-Mental State Examination was used, and its score was included in the analysis as a continuous variable.

      Oral health and functions

      The Japanese version of the General Oral Health Assessment Index (GOHAI)
      • Atchison K.A.
      • Dolan T.A.
      Development of the geriatric oral health assessment index.
      • Naito M.
      • Suzukamo Y.
      • Nakayama T.
      • et al.
      Linguistic adaptation and validation of the General Oral Health Assessment Index (GOHAI) in an elderly Japanese population.
      was used to measure the oral health-related quality of life. Numbers of remaining teeth were counted by dental hygienists. Occlusal force was assessed by Dental Prescale (Fujifilm, Shizuoka-prefecture, Japan).

      Nutritional and dietary status

      BMI was calculated by dividing the weight by the square of height. Food variety score was calculated from a 10-item questionnaire.
      • Kumagai S.
      • Watanabe S.
      • Shibata H.
      • et al.
      Effects of dietary variety on declines in high-level functional capacity in elderly people living in a community.
      Nutritional status was assessed by Mini-Nutrition Assessment-Short Form (MNA-SF), with scores ≤ 11 indicating possible malnutrition.
      • Rubenstein L.Z.
      • Harker J.O.
      • Salva A.
      • et al.
      Screening for undernutrition in geriatric practice: Developing the Short-Form Mini-Nutritional Assessment (MNA-SF).

      Statistical Analysis

      Binomial multiple logistic regression analysis was performed with depressive symptoms as the dependent variable, stratified by the age groups (65–74 years old indicating “young-old” and ≥75 years old indicating “old-old”). Multinomial multiple logistic regression analysis was performed with different degrees of depressive states (“mild depression” and “severe depression”) as the outcome. The characteristics of the 4 groups by eating and living arrangement were also compared, to explore the reasons behind their differing associations with depressive symptoms. For continuous variables only, multiple comparison test (Dunnett T3) was used to test whether there were significant differences between “living with others yet eating alone” and “living and eating with others.” IBM SPSS statistics v 22 for Windows (IBM Japan, Tokyo, Japan) was used to perform statistical analysis. P value of <.05 was considered to indicate statistical significance.

      Ethical Considerations

      The study was approved by the Ethics Committee of the University of Tokyo. Data received for analysis had the participants' names substituted with ID numbers, and confidential information was excluded to ensure protection of personal information.

      Results

      Sample Characteristics

      Of the total 1856 participants (928 male and 928 female, mean age was 72.9 ± 5.5 years), 1201 (64.7%) were young-old whereas 655 (35.3%) were old-old. Furthermore, 14.7% showed depressive symptoms (14.0% of young-old and 15.9% of old-old, 15.6% of women and 13.7% of men); 10.6% were living alone (8.1% of young-old and 15.1% of old-old, 15.4% of women and 5.7% of men); 14.6% were eating alone (11.1% of young-old and 21.1% of old-old, 17.9% of women and 11.3% of men); and 6.0% were eating alone despite living with family members (4.2% of young-old and 9.2% of old-old, 5.2% of women and 6.8% of men).

      Social Engagement and Depressive Symptoms by Age Groups

      Table 1 shows the comparison of the geriatric characteristics between normal and depressed participants for young-old and old-old, respectively. Based on this result, logistic regression was performed to identify the key risk factors for depressive symptoms (Table 2). The variables independently associated with depressive symptoms for both age-groups were “living with others yet eating alone,” social participation (fewer frequency of going out), social stressors (major change in life), and social ties of friends, health literacy, and GOHAI. Those unique to young-old were low reciprocal social support, social ties with family, low income, mobility, and MNA-SF scores. Risk factor unique to old-old was the number of medications.
      Table 1Geriatric Characteristics of Normal (Nondepressed) and Depressed Study Participants
      χ2 test or Fisher exact test was used for categorical variables and nonpaired t-test was used for continuous variables.
      (n = 1856)
      VariablesYoung-Old (65–74 Years Old)Old-Old (≥75 Years Old)
      Normal (n = 1033)Depressive Symptoms (n = 168)Normal (n = 551)Depressive Symptoms (n = 104)
      Mean ± SD or n (%)P ValueMean ± SD or n (%)P Value
      Sociodemographic variables
       Sex (male)519 (50.2)71 (42.3).055282 (51.2)56 (53.8).618
       Age69.6 ± 2.769.6 ± 2.6.96979.0 ± 3.779.4 ± 4.0.294
       Education (years)13.0 ± 2.512.6 ± 2.6.08912.4 ± 3.111.7 ± 3.3.056
       Health literacy4.03 ± 0.613.71 ± 0.67<.0014.07 ± 0.603.64 ± 0.70<.001
       Low income598 (57.9)126 (75.0)<.001293 (53.2)62 (59.6).227
      Social engagement
       Living alone77 (7.5)20 (11.9).05084 (15.2)15 (14.4).830
       Eating alone91 (8.8)42 (25.0)<.001104 (18.9)34 (32.7).002
      Living and eating with others929 (89.9)124 (73.8)<.001428 (77.7)68 (65.4).007
      Living and eating alone64 (6.2)18 (10.7).03165 (11.8)13 (12.5).839
      Living alone yet eating with others13 (1.3)2 (1.2)1.00019 (3.4)2 (1.9).555
      Living with others yet eating alone27 (2.6)24 (14.3)<.00139 (7.1)21 (20.2)<.001
       Low reciprocal social support45 (4.4)29 (17.3)<.00134 (6.2)18 (17.3)<.001
       Fewer frequency of going out127 (12.3)65 (38.7)<.001107 (19.4)47 (45.2)<.001
       Major change in life225 (21.8)62 (36.9)<.00185 (15.4)28 (26.9).004
       Social ties with family8.33 ± 3.16.58 ± 3.1<.0018.21 ± 3.26.91 ± 3.0<.001
       Social ties with friends8.43 ± 3.56.23 ± 3.4<.0018.43 ± 3.66.30 ± 3.4<.001
      Medical histories
       Hypertension388 (37.6)78 (46.4).029270 (49.0)69 (66.3).001
       Cerebrovascular diseases36 (3.5)16 (9.5)<.00147 (8.5)13 (12.5).198
       Diabetes116 (11.2)17 (10.1).67168 (12.3)14 (13.5).752
       Osteoporosis77 (7.5)21 (12.5).02779 (14.3)23 (22.1).045
       Heart diseases151 (14.6)28 (16.7).489111 (20.1)32 (30.8).016
       Malignant neoplasm152 (14.7)16 (9.5).07292 (16.7)23 (22.1).183
      Number of medications2.21 ± 2.52.85 ± 2.9.0083.80 ± 3.35.72 ± 3.9<.001
      Physical health and functions
       IADL4.90 ± 0.364.77 ± 0.63.0134.85 ± 0.504.61 ± 0.89.007
       Mobility25.8 ± 9.821.1 ± 10<.00124.1 ± 9.920.9 ± 11.003
      Cognitive function: MMSE28.5 ± 1.728.0 ± 1.9.00228.0 ± 1.927.3 ± 2.3.006
      Oral health and functions: GOHAI55.8 ± 5.451.3 ± 7.1<.00154.5 ± 6.349.5 ± 8.9<.001
      Nutritional and dietary status
       BMI (kg/m2)23.0 ± 2.922.6 ± 3.0.07122.7 ± 3.122.6 ± 3.0.625
       Food variety3.63 ± 2.03.04 ± 1.9<.0014.23 ± 2.13.72 ± 2.1.021
       MNA-SF12.7 ± 1.312.1 ± 1.8<.00112.4 ± 1.511.8 ± 1.8.004
      BMI, body mass index; MMSE, Mini-Mental State Examination; SD, standard deviation.
      χ2 test or Fisher exact test was used for categorical variables and nonpaired t-test was used for continuous variables.
      Table 2Association Between Depressive Symptoms and Risk Factors by Binomial Multiple Logistic Regression
      VariablesYoung-Old (65–74 Years Old) (n = 1201)
      Model 1Model 2
      OR (95% CI)P ValueOR (95% CI)P Value
      Social engagement
       Living and eating with others (ref)
       Living and eating alone1.94 (1.1–3.6).0341.53 (0.79–2.9).204
       Living alone yet eating with others1.59 (0.32–7.9).5691.14 (0.19–6.8).885
       Living with others yet eating alone6.33 (3.3–12)<.0015.02 (2.5–9.9)<.001
       Low reciprocal social support2.57 (1.5–4.6).0012.41 (1.3–4.5).006
       Fewer frequency of going out3.79 (2.6–5.6)<.0012.57 (1.7–3.9)<.001
       Major change in life1.78 (1.2–2.6).0041.72 (1.1–2.6).009
       Social ties with family0.901 (0.84–0.96).0020.905 (0.84–0.97).005
       Social ties with friends0.911 (0.86–0.96).0010.940 (0.88–1.0).049
      Sociodemographic variables
       Sex (male)1.29 (0.77–2.2).334
       Health literacy0.691 (0.52–0.93).013
       Low income1.77 (1.0–3.0).038
      Medical histories
       Hypertension1.17 (0.75–1.8).486
       Cerebrovascular diseases1.99 (0.89–4.4).094
       Osteoporosis1.38 (0.74–2.6).308
      Number of medications1.03 (0.96–1.1).402
      Physical health and functions
       IADL0.824 (0.54–1.3).369
       Mobility0.973 (0.96–0.99).007
       Cognitive function: MMSE1.04 (0.92–1.2).521
       Oral health and functions: GOHAI0.944 (0.92–0.97)<.001
      Nutritional and dietary status
       Food variety0.929 (0.84–1.0).163
       MNA-SF0.870 (0.76–0.99).038
      VariablesOld-Old (≥75 Years Old) (n = 655)
      Model 1Model 2
      OR (95% CI)P ValueOR (95% CI)P Value
      Social engagement
       Living and eating with others (ref)
       Living and eating alone1.01 (0.51–2.0).9681.06 (0.48–2.4).889
       Living alone yet eating with others0.753 (0.17–3.4).7120.979 (0.19–5.0).980
       Living with others yet eating alone2.45 (1.3–4.7).0062.41 (1.2–4.8).014
       Low reciprocal social support1.91 (0.95–3.9).0711.04 (0.48–2.3).917
       Fewer frequency of going out2.97 (1.9–4.7)<.0012.09 (1.2–3.6).008
       Major change in life1.98 (1.2–3.4).0122.18 (1.2–3.9).009
       Social ties with family0.981 (0.90–1.1).6510.972 (0.89–1.1).548
       Social ties with friends0.880 (0.82–0.94)<.0010.895 (0.83–0.97).006
      Sociodemographic variables
       Sex (male)1.56 (0.88–2.8).126
       Health literacy0.499 (0.34–0.74)<.001
      Medical histories
       Hypertension1.46 (0.83–2.6).185
       Osteoporosis1.27 (0.63–2.5).505
       Heart diseases1.21 (0.68–2.1).525
      Number of medications1.10 (1.0–1.2).010
      Physical health and functions
       IADL0.842 (0.59–1.2).340
       Mobility1.00 (0.98–1.0).990
      Cognitive function: MMSE0.919 (0.82–1.0).160
      Oral health and functions: GOHAI0.935 (0.90–0.97)<.001
      Nutritional and dietary status
       Food variety0.982 (0.87–1.1).770
       MNA-SF0.929 (0.79–1.1).365
      CI, confidence interval; MMSE, Mini-Mental State Examination; OR, odds ratio.
      Model 1: social engagement.
      Model 2: social engagement, sociodemographic variables, medical histories, number of medications, physical health and functions, cognitive function, oral health and functions, and nutritional and dietary status.

      Social Engagement and Different Degrees of Depression

      Table 3 shows the comparison of the geriatric characteristics between “normal,” “mildly depressed” and “severely depressed” participants. Based on this result, multinomial logistic regression was performed, as shown in Table 4. The variables independently associated with both degrees of depression were eating alone, social participation (fewer frequency of going out), social ties with friends, health literacy, the number of medications, and GOHAI. Those unique for “mild depression” were living alone; they had low reciprocal social support, social stressors (major change in life), social ties with family, age, low income, and mobility. Risk factors unique for “severe depression” were male gender, history of cerebrovascular diseases, and MNA-SF scores.
      Table 3Geriatric Characteristics of Normal, Mildly Depressed and Severely Depressed Participants
      Cochran-Armitage trend test was used for categorical variables and Jonckheere-Terpstra trend test was used for continuous variables.
      (n = 1856)
      Normal (n = 1584)Mild Depression (n = 193)Severe Depression (n = 79)
      VariablesMean ± SD or n (%)P Value
      Sociodemographic variables
       Sex (male)801 (50.6)84 (43.5)43 (54.4).601
       Age72.8 ± 5.472.7 ± 5.674.8 ± 6.0.201
       Education (years)12.8 ± 2.712.3 ± 2.912.2 ± 3.1.007
       Health literacy4.04 ± 0.613.75 ± 0.673.52 ± 0.70<.001
       Low income891 (56.3)137 (71.0)51 (64.6).001
      Social engagement
       Living alone161 (10.2)19 (9.8)16 (20.3).031
       Eating alone195 (12.3)47 (24.4)29 (36.7)<.001
      Living and eating with others (ref)1357 (85.7)146 (75.6)46 (58.2)<.001
      Living and eating alone129 (8.1)19 (9.8)12 (15.2).031
      Living alone yet eating with others32 (2.0)0 (0.0)4 (5.1).681
      Living with others yet eating alone66 (4.2)28 (14.5)17 (21.5)<.001
       Low reciprocal social support79 (5.0)30 (15.5)17 (21.5)<.001
       Fewer frequency of going out234 (14.8)75 (38.9)37 (46.8)<.001
       Major change in life310 (19.6)66 (34.2)24 (30.4)<.001
       Social ties with family8.29 ± 3.16.82 ± 3.16.42 ± 3.0<.001
       Social ties with friends8.43 ± 3.56.42 ± 3.45.86 ± 3.4<.001
      Medical histories
       Hypertension658 (41.5)107 (55.4)40 (50.6).001
       Cerebrovascular diseases83 (5.2)17 (8.8)12 (15.2)<.001
       Diabetes184 (11.6)23 (11.9)8 (10.1).805
       Osteoporosis156 (9.8)31 (16.1)13 (16.5).003
       Heart diseases262 (16.5)43 (22.3)17 (21.5).043
       Malignant neoplasm244 (15.4)27 (14.0)12 (15.2).739
      Number of medications2.77 ± 2.93.84 ± 3.44.20 ± 3.9<.001
      Physical health and functions
       IADL4.88 ± 0.424.73 ± 0.704.66 ± 0.83<.001
       Mobility25.2 ± 9.821.0 ± 1020.9 ± 11<.001
      Cognitive function: MMSE28.3 ± 1.827.7 ± 2.027.7 ± 2.2<.001
      Oral health and functions: GOHAI55.4 ± 5.851.1 ± 7.449.2 ± 8.7<.001
      Nutritional and dietary status
       BMI (kg/m2)22.9 ± 3.022.7 ± 3.122.3 ± 2.9.163
       Food variety3.84 ± 2.03.34 ± 2.03.20 ± 2.1<.001
       MNA-SF12.6 ± 1.412.1 ± 1.711.7 ± 1.9<.001
      BMI, body mass index; MMSE, Mini-Mental State Examination; SD, standard deviation.
      Cochran-Armitage trend test was used for categorical variables and Jonckheere-Terpstra trend test was used for continuous variables.
      Table 4Association Between Mild and Severe Depression and Their Risk Factors by Multinomial Multiple Logistic Regression (n = 1856)
      Mild Depression (n = 193)Severe Depression (n = 79)
      VariablesOR 95%CIP ValueOR 95%CIP Value
      Social engagement
       Living alone0.374 (0.19–0.74).0050.777 (0.33–1.8).566
       Eating alone2.96 (1.8–5.0)<.0013.33 (1.6–6.8).001
       Low reciprocal social support1.73 (1.0–2.9).0451.66 (0.80–3.4).172
       Fewer frequency of going out2.21 (1.5–3.2)<.0012.79 (1.6–4.8)<.001
       Major change in life1.78 (1.2–2.6).0021.63 (0.93–2.9).091
       Social ties with family0.940 (0.88–1.0).0460.935 (0.85–1.0).162
       Social ties with friends0.929 (0.88–0.98).0070.895 (0.82–0.97).009
      Sociodemographic variables
       Sex (male)1.27 (0.78–2.1).3352.46 (1.2–5.0).013
       Age0.950 (0.92–0.98).0050.998 (0.95–1.0).943
       Education (years)1.05 (0.98–1.1).1901.03 (0.93–1.1).582
       Health literacy0.670 (0.52–0.87).0030.440 (0.31–0.63)<.001
       Low income1.72 (1.1–2.8).0241.65 (0.84–3.3).145
      Medical histories
       Hypertension0.743 (0.51–1.1).1181.14 (0.64–2.0).655
       Cerebrovascular diseases1.38 (0.74–2.6).3122.36 (1.1–5.2).033
       Osteoporosis0.712 (0.43–1.2).1840.839 (0.39–1.8).652
       Heart diseases1.00 (0.65–1.5).9941.28 (0.67–2.5).461
      Number of medications1.08 (1.0–1.1).0171.10 (1.0–1.2).027
      Physical health and functions
       IADL0.834 (0.63–1.1).2150.862 (0.59–1.3).446
       Mobility0.983 (0.97–1.0).0440.988 (0.96–1.0).327
      Cognitive function: MMSE0.927 (0.85–1.0).1030.994 (0.87–1.1).930
      Oral health and functions: GOHAI0.943 (0.92–0.97)<.0010.928 (0.90–0.96)<.001
      Nutritional and dietary status
       Food variety0.959 (0.88–1.0).3440.960 (0.84–1.1).531
       MNA-SF0.936 (0.84–1.0).2510.839 (0.72–0.98).029
      CI, confidence interval; MMSE, Mini-Mental State Examination; OR, odds ratio.

      Living Arrangement and Eating Arrangement

      To examine further the role of eating alone and its potential risk factors, living arrangement and eating arrangement were crossed and the physical, mental, oral, cognitive, nutritional and dietary as well as social characteristics of the 4 resultant groups [living and eating alone (n = 160), living alone yet eating with others (n = 36), living with others yet eating alone (n = 111), living and eating with others (n = 1549)] were compared. The results are shown in Table 5.
      Table 5Characteristics by Living and Eating Arrangement (n = 1856)
      Living and Eating Alone (n = 160)Living Alone Yet Eating With Others (n = 36)Living with Others Yet Eating Alone (n = 111)Living and Eating With Others (n = 1549)
      VariablesMean ± SD or n (%)P Value
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
      Social engagement
       Live with spouse61 (55.0)1393 (89.9)<.001
       Live with children74 (66.7)627 (40.5)<.001
       Live with children-in-law21 (18.9)117 (7.6)<.001
       Live with grand-children29 (26.1)171 (11.0)<.001
       Social ties with family7.24 ± 3.48.83 ± 3.57.19 ± 3.28.19 ± 3.1<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Social ties with friends8.08 ± 3.48.86 ± 2.96.86 ± 4.08.19 ± 3.6.003
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
      Sociodemographic variables
       Sex (male)42 (26.3)11 (30.6)63 (56.8)812 (52.4)<.001
       Age74.6 ± 6.075.4 ± 5.275.3 ± 5.772.5 ± 5.3<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Education (years)11.9 ± 2.712.0 ± 2.711.8 ± 3.212.9 ± 2.7<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Health literacy3.87 ± 0.714.13 ± 0.713.86 ± 0.664.01 ± 0.62.015
      Number of medications3.50 ± 3.54.17 ± 4.33.76 ± 3.72.79 ± 2.9.004
      Physical health and functions
       Usual gait speed (m/s)1.43 ± 0.251.44 ± 0.261.41 ± 0.271.48 ± 0.25.026
       Max gait speed (m/s)2.05 ± 0.382.03 ± 0.462.01 ± 0.362.17 ± 0.39<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       IADL4.94 ± 0.304.94 ± 0.234.69 ± 0.844.86 ± 0.46.007
       Mobility23.9 ± 1027.3 ± 1121.0 ± 1124.9 ± 9.9<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
      Mental health
       GDS3.18 ± 3.42.86 ± 3.24.83 ± 4.12.39 ± 2.7<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Depressive symptoms: GDS ≥631 (19.4)4 (11.1)45 (40.5)192 (12.4)<.001
       Severe depression: GDS ≥1012 (7.5)4 (11.1)17 (15.3)46 (3.0)<.001
      Cognitive function: MMSE28.3 ± 1.828.0 ± 1.627.8 ± 1.928.2 ± 1.8.029
      Oral health and functions
       GOHAI53.8 ± 7.353.3 ± 7.853.1 ± 6.654.9 ± 6.2<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Number of remaining teeth20.5 ± 8.019.2 ± 8.417.8 ± 9.721.0 ± 8.3.003
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
       Occlusal force (N)496 ± 333522 ± 365478 ± 345585 ± 361<.001
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
      Nutritional and dietary status
       BMI (kg/m2)22.3 ± 3.324.3 ± 3.622.8 ± 3.322.9 ± 2.9.002
       Food variety3.74 ± 2.03.89 ± 2.03.26 ± 2.13.79 ± 2.0.037
       MNA-SF12.2 ± 1.612.4 ± 1.612.1 ± 1.712.5 ± 1.4.007
      Those continuous variables that showed significant difference between “living with others yet eating alone” and “living and eating with others” in the multiple comparison test (Dunnett T3) are highlighted with “*”.
      Kruskal-Wallis test for continuous variables and χ2 test for categorical variables.
      The participants “living with others yet eating alone” had the poorest scores of social ties with family and friends, years of education, health literacy, physical health and functions (normal and maximum gait speeds, IADL and mobility), cognitive function, oral health and functions (GOHAI, number of remaining teeth, and occlusal force), and nutritional and dietary status (MNA-SF and food variety).
      Furthermore, greater proportion of those who “live with others yet eat alone” live with their children, children-in-law, and grand-children, compared with those who “live and eat with others,” most of whom live with their spouse.

      Discussion

      The main aim of the present study was to examine the association between social engagement and depressive symptoms in community-dwelling Japanese older adults, with a particular focus on eating alone and on the changes in the association along geriatric trajectories of aging and mental frailty.
      The study was carried out on a population sample of Japanese older adults, of whom 14.7% showed depressive symptoms (GDS ≥6). This is on the lower end compared with previous studies that used the same GDS cut-off point, in which the prevalence ranged between 14% and 40%.
      • Wada T.
      • Ishine M.
      • Sakagami T.
      • et al.
      Depression in Japanese community-dwelling elderly—Prevalence and association with ADL and QOL.
      The results highlighted a significant association between depressive symptoms and social engagement variables such as social ties, eating alone, social participation, social stressors and reciprocity of social support. Of particular interest was eating alone, which to our knowledge has not been assessed before in combination with different components of social engagement and in relation to the living arrangement. “Living with others yet eating alone” was a significant predictor of depression for both age groups, with odds ratio reaching as high as 5 times for the young-old. This suggests that eating alone acts as stronger risk factor than living alone, and that the living arrangement in which older adults eat alone can act as a critical determinant of depressive risks. Meals are an important location of socialization whereby older adults enjoy intimate interactions, and when shared with others, they can provide valuable opportunities for companionship and social support.
      • Vesnaver E.
      • Keller H.H.
      Social influences and eating behavior in later life: A review.
      A lack of communication during meals may result in feelings of loneliness and depressed moods.
      • Kimura Y.
      • Wada T.
      • Okumiya K.
      • et al.
      Eating alone among community-dwelling Japanese elderly: Association with depression and food diversity.
      Table 5 suggests that those who eat alone despite living with their families tend to be the most socially withdrawn, with least awareness of their health conditions and the poorest physical, oral, and cognitive functions as well as nutritional status. The fact that they do not share a single meal with their families despite living together suggests that they have distant relationships with them. Compared with those who eat with others, a greater proportion of those who eat alone live with their children, children-in-law or grandchildren, and less with their spouse. This suggests that they may be eating alone because they lead different life styles, suffer from emotional distance, concerns that they will add burdens on their families if they eat together, or from uncomfortable relationships with family members such as children-in-law. This is supported by the fact that they have the weakest social ties with family. This may result in lower interest in their health shown by their families, as well as in lower self-interest. The fact that they show the lowest health literacy also supports this hypothesis. They also exhibit the lowest mobility and social ties with friends, suggesting that they are the most socially isolated not only at home but also outside. The fact that their gait speeds and IADL are the lowest imply that their poor physical functions play a role in limiting their social activities. GOHAI scores, number of remaining teeth and occlusal force are lowest in this group, indicating the possibility that they eat alone because they eat too slowly, require different menus, or because they have concerns about their oral appearance. The poor oral functions and nutritional/dietary status (low food variety and MNA-SF scores) may also be another manifestation of the lack of interest in their health shown by their families as well as by themselves.
      In any case, the sentiments or perceptions that lead them to eat alone despite living with their families are likely to be negative in nature and may be internally conceived by the older adults themselves, or externally imposed by families living together or the wider society. The functional decline, which may be a cause as well as a result from eating alone, may also contribute to the depressive outcomes.
      Stratification by age groups and multinomial regression analysis by different severities of depression revealed that fewer variables of social engagement were associated with depressive outcomes as the population ages or becomes more mentally frail. This suggests that social engagement is a more powerful predictor of mental health at earlier points along geriatric trajectories, and, thus, that effective social preventive measures require early interventions. Lower down the geriatric trajectories, social factors fall in their relative importance and the role of health and functional factors increase. This is suggested by the fact that the number of medications becomes a significant predictor for old-old, and the history of cerebrovascular diseases and MNA-SF scores become significant for severe depression.
      Outside the domain of social engagement, the independent risk factors for depressive symptoms in both age groups included GOHAI and health literacy, supporting the findings of previous studies.
      • Hassel A.J.
      • Danner D.
      • Schmitt M.
      • et al.
      Oral health-related quality of life is linked with subjective well-being and depression in early old age.
      • de Andrade F.B.
      • Lebrao M.L.
      • Santos J.L.F.
      • et al.
      Relationship between oral health-related quality of life, oral health, socioeconomic, and general health factors in elderly Brazilians.
      • Gazmararian J.
      • Baker D.
      • Parker R.
      • Blazer D.G.
      A multivariate analysis of factors associated with depression: Evaluating the role of health literacy as a potential contributor.
      Uniquely for young-old, mobility, MNA-SF, and income were associated. For old-old only, the number of medications remained a predictor of depressive symptoms.
      This study elucidates that reducing the risk of depression requires much more than medical care and that preventive measures need to be introduced early on in the geriatric trajectories, before frailty sets in. The present study shows that social factors such as eating alone pose substantial risk for mental health. Comprehensive assessment that covers a wide range of health-related domains including physical health, oral functions, nutritional, and dietary status as well as social relations will be necessary to identify those at risk effectively.
      The limitations of our study are mainly 4-fold. First, the cross-sectional nature of the study prevents it from making any conclusive comments about the causality between independent variables and the outcome. Second, data on household income were not available, and instead, individual income was used. Given that the older adults in the present study grew up in a period when it was rare for women to work after marriage, household income would have been a better indicator of the economic environment for women. Third, depressive symptoms were measured using self-administered GDS questionnaire rather than diagnosis by physicians. Fourth, the participants inevitably comprised those who had greater degrees of interest in health and lower barriers to participation in the study. This may have skewed the nature of participants, to those who were more socially active and interested in health, missing out those who were most socially disengaged.

      Conclusions

      For community-dwelling Japanese older adults, depressive symptoms were significantly associated with social engagement, with greater associations in younger and less mentally frail populations. Eating alone was identified as a key risk factor for depressive symptoms, and those who live with their families yet eat their meals alone were at highest risk. Mental health management for older adults, therefore, requires comprehensive assessment of their social relations, taking into account their companionship during mealtimes. Social preventive measures need to involve early interventions in order to augment their effectiveness against mental frailty.
      Given that depression can lay the ground for further frailty and various detrimental health outcomes, further study with a longitudinal design, with more detailed data collection on social predictors of depression, may play a pivotal role in identifying possible intervention opportunities to prevent not only mental but also physical frailties.

      Acknowledgments

      The authors thank the staff members and participants of the Kashiwa study and the following individuals for helping with the acquisition of data: Dr Takashi Higashiguchi, Fujita Health University School of Medicine; Dr Kazuko Ishikawa-Takata RD, National Institute of Health and Nutrition; Dr Yoshiya Oishi PhD DDS, Oishi Dental Clinic; Dr Noriaki Takahashi, The Nippon Dental University; Seigo Mitsutake, Tokyo Metropolitan Institute of Gerontology; and staff members of The Institute of Healthcare Innovation Project, The University of Tokyo.

      References

        • Djernes J.K.
        Prevalence and predictors of depression in populations of elderly: A review.
        Acta Psychiatr Scand. 2006; 113: 372-387
        • Schwarzbach M.
        • Luppa M.
        • Forstmeier S.
        • et al.
        Social relations and depression in late life-A systematic review.
        Int J Geriatr Psychiatry. 2014; 29: 1-21
        • Muramatsu N.
        • Akiyama H.
        Japan: Super-aging society preparing for the future.
        Gerontologist. 2011; 51: 425-432
        • Wada T.
        • Ishine M.
        • Sakagami T.
        • et al.
        Depression in Japanese community-dwelling elderly—Prevalence and association with ADL and QOL.
        Arch Gerontol Geriatr. 2004; 39: 15-23
        • Blazer D.G.
        • Hybels C.F.
        Origins of depression in later life.
        Psychol Med. 2005; 35: 1241-1252
        • Park N.S.
        • Jang Y.
        • Lee B.S.
        • et al.
        The mediating role of loneliness in the relation between social engagement and depressive symptoms among older Korean Americans: Do men and women differ?.
        J Gerontol B Psychol Sci Soc Sci. 2013; 68: 193-201
        • Beekman A.T.F.
        • Deeg D.J.H.
        • vanTilburg T.
        • et al.
        Major and minor depression in later life: A study of prevalence and risk factors.
        J Affect Disord. 1995; 36: 65-75
        • Fukunaga R.
        • Abe Y.
        • Nakagawa Y.
        • et al.
        Living alone is associated with depression among the elderly in a rural community in Japan.
        Psychogeriatrics. 2012; 12: 179-185
        • Kawachi I.
        • Berkman L.F.
        Social ties and mental health.
        J Urban Health. 2001; 78: 458-467
        • Heun R.
        • Hein S.
        Risk factors of major depression in the elderly.
        Eur Psychiatry. 2005; 20: 199-204
        • Mendes de Leon C.F.
        Social engagement and successful aging.
        Eur J Ageing. 2005; 2: 64-66
        • Park N.S.
        The relationship of social engagement to psychological well-being of older adults in assisted living facilities.
        J Appl Gerontol. 2009; 28: 461-481
        • Tilvis R.S.
        • Routasalo P.
        • Karppinen H.
        • et al.
        Social isolation, social activity and loneliness as survival indicators in old age; A nationwide survey with a 7-year follow-up.
        Eur Geriatr Med. 2012; 3: 18-22
        • Tomaka J.
        • Thompson S.
        • Palacios R.
        The relation of social isolation, loneliness, and social support to disease outcomes among the elderly.
        J Aging Health. 2006; 18: 359-384
        • Avlund K.
        • Lund R.
        • Holstein B.E.
        • Due P.
        Social relations as determinant of onset of disability in aging.
        Arch Gerontol Geriatr. 2004; 38: 85-99
        • Bassuk S.S.
        • Glass T.A.
        • Berkman L.F.
        Social disengagement and incident cognitive decline in community-dwelling elderly persons.
        Ann Intern Med. 1999; 131: 165-173
        • Bath P.A.
        • Deeg D.
        Social engagement and health outcomes among older people: Introduction to a special section.
        Eur J Ageing. 2005; 2: 24-30
        • Vesnaver E.
        • Keller H.H.
        Social influences and eating behavior in later life: A review.
        J Nutr Gerontol Geriatr. 2011; 30: 2-23
        • Kimura Y.
        • Wada T.
        • Okumiya K.
        • et al.
        Eating alone among community-dwelling Japanese elderly: Association with depression and food diversity.
        J Nutr Health Aging. 2012; 16: 728-731
        • Markle-Reid M.
        • Browne G.
        Conceptualizations of frailty in relation to older adults.
        J Adv Nurs. 2003; 44: 58-68
        • Schreiner A.S.
        • Hayakawa H.
        • Morimoto T.
        • Kakuma T.
        Screening for late life depression: Cut-off scores for the Geriatric Depression Scale and the Cornell Scale for Depression in Dementia among Japanese subjects.
        Int J Geriatr Psychiatry. 2003; 18: 498-505
        • Crooks V.C.
        • Lubben J.
        • Petitti D.B.
        • et al.
        Social network, cognitive function, and dementia incidence among elderly women.
        Am J Public Health. 2008; 98: 1221-1227
        • Lubben J.
        • Blozik E.
        • Gillmann G.
        • et al.
        Performance of an abbreviated version of the Lubben Social Network Scale among three European community-dwelling older adult populations.
        Gerontologist. 2006; 46: 503-513
        • Ishikawa H.
        • Nomura K.
        • Sato M.
        • Yano E.
        Developing a measure of communicative and critical health literacy: A pilot study of Japanese office workers.
        Health Promot Int. 2008; 23: 269-274
        • Liu C.P.
        • Leung D.S.
        • Chi I.
        Social functioning, polypharmacy and depression in older Chinese primary care patients.
        Aging Ment Health. 2011; 15: 732-741
        • Koyano W.
        • Shibata H.
        • Nakazato K.
        • et al.
        Measurement of competence—Reliability and validity of the TMIG index of competence.
        Arch Gerontol Geriatr. 1991; 13: 103-116
        • Baker P.S.
        • Bodner E.V.
        • Allman R.M.
        Measuring life-space mobility in community-dwelling older adults.
        J Am Geriatr Soc. 2003; 51: 1610-1614
        • Peel C.
        • Baker P.S.
        • Roth D.L.
        • et al.
        Assessing mobility in older adults: The UAB Study of Aging Life-Space Assessment.
        Phys Ther. 2005; 85: 1008-1019
        • Shimada H.
        • Sawyer P.
        • Harada K.
        • et al.
        Predictive validity of the classification schema for functional mobility tests in instrumental activities of daily living decline among older adults.
        Arch Phys Med Rehabil. 2010; 91: 241-246
        • Shimada H.
        • Ishizaki T.
        • Kato M.
        • et al.
        How often and how far do frail elderly people need to go outdoors to maintain functional capacity?.
        Arch Gerontol Geriatr. 2010; 50: 140-146
        • Ishii S.
        • Tanaka T.
        • Shibasaki K.
        • et al.
        Development of a simple screening test for sarcopenia in older adults.
        Geriatr Gerontol Int. 2014; 14: 93-101
        • Atchison K.A.
        • Dolan T.A.
        Development of the geriatric oral health assessment index.
        J Dent Educ. 1990; 54: 680-687
        • Naito M.
        • Suzukamo Y.
        • Nakayama T.
        • et al.
        Linguistic adaptation and validation of the General Oral Health Assessment Index (GOHAI) in an elderly Japanese population.
        J Public Health Dent. 2006; 66: 273-275
        • Kumagai S.
        • Watanabe S.
        • Shibata H.
        • et al.
        Effects of dietary variety on declines in high-level functional capacity in elderly people living in a community.
        Nihon Koshu Eisei Zasshi. 2003; 50: 1117-1124
        • Rubenstein L.Z.
        • Harker J.O.
        • Salva A.
        • et al.
        Screening for undernutrition in geriatric practice: Developing the Short-Form Mini-Nutritional Assessment (MNA-SF).
        J Gerontol A Biol Sci Med Sci. 2001; 56: M366-M372
        • Hassel A.J.
        • Danner D.
        • Schmitt M.
        • et al.
        Oral health-related quality of life is linked with subjective well-being and depression in early old age.
        Clin Oral Investig. 2011; 15: 691-697
        • de Andrade F.B.
        • Lebrao M.L.
        • Santos J.L.F.
        • et al.
        Relationship between oral health-related quality of life, oral health, socioeconomic, and general health factors in elderly Brazilians.
        J Am Geriatr Soc. 2012; 60: 1755-1760
        • Gazmararian J.
        • Baker D.
        • Parker R.
        • Blazer D.G.
        A multivariate analysis of factors associated with depression: Evaluating the role of health literacy as a potential contributor.
        Arch Intern Med. 2000; 160: 3307-3314