- Email: [email protected]
Eating Together and Health-Related Quality of Life Among Korean Adults
ARTICLE IN PRESS Research Article Eating Together and Health-Related Quality of Life Among Korean Adults Min-Jung Choi, PhD1; Yong Gyu Park, PhD2; Yang Hyun Kim, MD, PhD3; Kyung Hwan Cho, MD, PhD3; Ga Eun Nam, MD, PhD3 ABSTRACT Objective: To investigate the association between frequency of eating together and health-related quality of life (HRQoL). Design: Population-based cross-sectional study. Participants and Setting: A total of 13,920 adults (≥19 years). Data obtained from the Korea National Health and Nutrition Examination Survey conducted from 2014 to 2016. The European Quality of Life-5 Dimensions index was used to assess the HRQoL. Main Outcome Measures: Relation between the frequency of eating together (eating alone, 1 time/d, and ≥2 times/d) and HRQoL. Analysis: ANOVA, correlations, and multivariable logistic regression. Results: A significantly positive association was observed between the frequency of eating together and the European Quality of Life-5 Dimensions index (P < .001). An increase in the frequency of eating together was associated with decreased odds ratios of having problems in subdimensions (self-care, usual activities, and anxiety or depression; P for trend <.005). Conclusions and Implications: A decreased frequency of eating together is associated with low HRQoL. Additional work is warranted to confirm the association between the frequency of eating together and HRQoL in young adults, aged 19−64 years. Further studies are required to identify whether an intervention to increase the frequency of eating together would improve HRQoL in individuals who usually eat alone. Key Words: population, quality of life, mealtime (J Nutr Educ Behav. 2019;000:1−8.) Accepted November 20, 2019.
INTRODUCTION In 2017, 28.6% of households in Korea were 1-person households,1 which is expected to increase to 36.3% by 2045.2 Accordingly, the number of people who eat alone is increasing steadily because of various factors such as living away from family for education, irregular and informal mealtimes,
working long hours, lack of time, and failed relationships.3−5 Eating a meal is a social activity that helps form relationships with other people; it goes beyond the act of putting food into one’s mouth.6 Eating alone has been reported to be associated with poor physical health leading to nutritional imbalance,7 weight problems,8 and metabolic syndrome,9 as well as
1
Department of Nursing, College of Nursing, The Catholic University of Korea, Seoul, Republic of Korea 2 Department of Biostatistics, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea 3 Department of Family Medicine, Korea University Anam Hospital, Korea University College of Medicine, Seoul, Republic of Korea Conflict of Interest Disclosure: The authors have not stated any conflicts of interest. Address for correspondence: Yong Gyu Park, PhD, Department of Biostatistics, College of Medicine, The Catholic University of Korea, 222 Banpo-daero, Seocho-gu, Seoul 06591, Republic of Korea; E-mail: [email protected]. Or: Ga Eun Nam, MD, PhD, Department of Family Medicine, Korea University Anam Hospital, Korea University College of Medicine, 82 Goryeodae-ro, Seongbuk-gu, Seoul 02841, Republic of Korea; E-mail: [email protected]. Ó 2019 Society for Nutrition Education and Behavior. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jneb.2019.11.013
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
mental health problems including stress,3 loneliness,10 and depression.11 The number of people who eat all 3 meals of the day alone has increased in the Korean population and has become an important social concern.12 Health-related quality of life (HRQoL) represents a concept of subjective physical and psychological well-being, which serves as an important indicator for evaluating chronic diseases and predicting health outcomes such as mortality and morbidity.13 Among the various tools used for measuring HRQoL, the European Quality of Life-5 Dimensions (EQ-5D) questionnaire has been recognized as suitable for the general population. Although it has only a few questions, it is divided into various profiles.14 Several previous studies showed that older adults who eat alone are unhappy15 and depressed,11 and have a lower HRQoL than those who eat with other people. However, few studies examined the association between the frequency of eating together and HRQoL considering the demographic
1
ARTICLE IN PRESS 2
Choi et al
factors, socioeconomic status, health behaviors, and comorbidities in the adult population. Therefore, this study aimed to examine the association between the frequency of eating together and HRQoL considering potential confounding variables using a nationwide data set from the Korea National Health and Nutrition Examination Survey (KNHANES),16 which is representative of the South Korean population.
METHODS Survey Overview and Study Population The KNHANES has been conducted every year by the Korea Centers for Disease Control & Prevention since 1998. The KNHANES provides national statistics on health and nutritional status, which are used to calculate the health index and evaluate health programs in South Korea. It is conducted using a systematic, stratified cluster sampling design and consists of 3 parts: a health interview, a nutrition survey, and a health examination. Of the 23,080 participants surveyed in the 2014−2016 KNHANES, this study initially evaluated 18,303 individuals aged ≥19 years. Then, 4,383 individuals were excluded because they (1) did not respond to questions regarding the frequency of eating together, and had a meal less than twice a week; (2) did not have information on socioeconomic status; or (3) did not have HRQoL data. Responses of 13,920 participants were included in the current study. The Research Ethics Committee of the Korea Centers for Disease Control & Prevention approved the survey protocol, and this study was approved by the Institutional Review Board of the Catholic University of Korea, Seoul St Mary’s Hospital (Approval No. KC18ZESI0062). The researchers obtained informed consent from all participants.
Assessment of the Frequency of Eating Together Six questions related to eating together were used in this study, the first of which was, “How many times a week did you eat breakfast in the last year?” The same question was then asked with regard to lunch and dinner.
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 Participants who responded less than twice a week in each meal were excluded. The participants were then instructed to answer Yes or No to the question, “Did you usually have breakfast with other people last year?” The same question was also asked regarding lunch and dinner. Participants who responded No to all 3 questions were assigned to the 0 times/d (eating alone) group. Those who responded No to 2 of the 3 questions were assigned to the 1 time/d eating-together group, and those who responded No to none or 1 of the 3 questions were assigned to the ≥2 times/d eating-together group.
Assessment of HRQoL The HRQoL was evaluated using the Korean version of the EQ-5D scale.15 The EQ-5D records the level of selfreported problems on the basis of 5 dimensions, including mobility, selfcare, usual activities, pain or discomfort, and anxiety or depression. Each dimension was assessed on the basis of a single question with 3 response levels (no problems, some problems, and extreme problems). The responses some problems and extreme problems were recorded as 1 point each, whereas no problems was given 0 points. The EQ-5D scores were calculated using the weighted model to transform these health states into Korean population−based health states.17 The EQ-5D index ranges from 1 to 1, in which 1 stands for perfect health status, 0 stands for death, and <0 represents health status worse than death.18 The Korean version of the EQ-5D has been evaluated for validity and reliability.18
Potential Confounding Variables Household income was divided into 4 quartiles. Education level was divided into elementary school or below, middle school, high school, and university or higher. Marital status was reported as with a spouse or without a spouse, and the household type was divided into those with 1 person or ≥2 people. Participants who answered Yes to economic activities were considered as employed. Based on smoking status, participants were classified as never-smokers, former smokers (had smoked ≥100 cigarettes during their lifetime but were
not smoking currently), and current smokers (had smoked ≥100 cigarettes during their lifetime and were still smoking). Problem drinkers were identified as those who consumed alcohol at least twice a week, with ≥7 servings/sitting (61 g) for men and ≥5 servings/ (40 g) for women.19 Physical activity (occupational, leisure, and transportation) was classified as Yes or No. Participants categorized as Yes were those who performed moderateintensity physical activity for at least 150 min/wk, high-intensity physical activity for 75 min/wk, or a combination of moderate- and high-intensity activities (1 minute of high-intensity physical activity = 3 minutes of moderate-intensity physical activity).20 Stress levels were determined on the basis of the response to the question, “How much stress do you usually feel in your daily life?” The allowed responses were very low, low, high, and very high. Participants were classified as those who had low (very low or low) or high (high or very high) stress levels. Participants were also asked whether they had ever been diagnosed with hypertension, diabetes, hyperlipidemia, stroke, heart disease (myocardial infarction or angina), arthritis (osteoarthritis or rheumatoid arthritis), or cancer. Trained medical staff measured participants’ height and body weight using standardized procedures in mobile examination centers and calculated body mass index (BMI) by dividing body weight (kg) by height squared (m); a participant with BMI ≥25 kg/m2 was defined as obese.21
Statistical Analysis All statistical analyses were conducted using SAS software (version 9.3; SAS Institute, Cary, NC; 1996) to account for the complex sampling design. The KNHANES is based on a 2-stage stratified cluster sampling method across districts and households; the sampling weights were obtained from these strata. Data are presented as mean § SE for continuous variables and as a percentage (SE) for categorical variables. The researchers compared the characteristics of participants on the basis of their frequency of eating together using 1-way ANOVA for continuous variables and Pearson chi-square test for categorical variables as a test of homogeneity. This study evaluated the
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 association between the frequency of eating together and HRQoL indices (EQ-5D index) using ANCOVA. Model 1 was adjusted for age and sex, and model 2 was adjusted for the same parameters in addition to household income, education level, marital status, 1-person households, and employment. Model 3 was further adjusted for smoking status, alcohol consumption, BMI, physical activity, stress, and comorbidities. Associations between the frequency of eating together and the 5 dimensions of the EQ-5D were analyzed by multivariable logistic regression. The odds ratios (ORs) and 95% confidence intervals (CIs) of having problems in subdimensions of the EQ5D were calculated in the group eating together 1 time/d and ≥2 times/d compared with the eating-alone group. Associations between the frequency of eating together and the EQ5D index in the relevant subgroups according to age, sex, obesity, type of
Choi et al
household, and marital status were also assessed using an ANCOVA after accounting for all potential confounding variables. Statistical significance was defined by a 2-tailed P < .05. Because there was order among the frequencies of eating together, P for trend was presented using the Cochran-Armitage test. P was presented with or without Bonferroni adjustment in the subgroup analysis.
RESULTS Characteristics of Participants Based on the Frequency of Eating Together Table 1 presents the characteristics of the study participants on the basis of the daily frequency of eating together. The group that ate together ≥2 times/d accounted for the highest proportion (67.1%), whereas the group that ate alone accounted for
3
13% of the total number of participants. The eating-alone group had the highest mean age (53.3 years) and the lowest socioeconomic status; these individuals were likely to be living alone (P < .001). The proportions of current smokers, problem drinkers, and regular exercisers were highest in the eating together ≥2 times/d group. This group also had the highest proportion of participants with high levels of stress (P < .001). The prevalence of comorbidities such as hypertension, diabetes mellitus, hyperlipidemia, stroke, cardiovascular disease, arthritis, and cancer was highest in the eating-alone group (P < .001).
Associations Between the Frequency of Eating Together and HRQoL The Figure shows that a lower frequency of eating together was
Table 1. Characteristics of Study Participants Based on Frequency of Eating Together (n = 13,920)
Characteristics Age, y ≤39 40−64 ≥65 Sex (male) Household income (lowest quartile) Education level (university or higher) Marital status (with spouse) 1-person households Occupation (yes) Current smoker Problem drinker Body mass index, kg/m2 <25 ≥25 Physical activity (yes) Subjective level of stress (high) Comorbidities Hypertension Diabetes mellitus Hyperlipidemia Stroke Cardiovascular disease Arthritis Cancer
0 Times/d (n = 814)
1 Time/d (n = 2,763)
≥2 Times/d (n = 9,343)
Statisticsa
53.3 § 0.7 27.2 (1.7) 37.7 (1.5) 35.0 (1.4) 40.4 (1.6) 36.6 (1.6) 23.9 (1.4) 73.5 (1.7) 36.3 (1.6) 43.2 (1.4) 21.2 (1.3) 8.6 (0.9) 24.1 § 0.1 64.9 (1.4) 35.1 (1.4) 46.7 (1.4) 32.2 (1.4)
44.4 § 0.4 41.2 (1.2) 46.7 (1.2) 12.1 (0.7) 39.9 (1.1) 13.4 (0.9) 37.5 (1.2) 72.0 (1.3) 11.8 (0.9) 57.7 (1.1) 20.7 (1.0) 11.4 (0.8) 23.6 § 0.1 67.8 (1.1) 32.2 (1.1) 53.5 (1.2) 28.4 (1.0)
46.2 § 0.3 37.0 (0.8) 47.8 (0.7) 15.2 (0.5) 54.0 (0.5) 12.4 (0.5) 41.3 (0.9) 79.2 (0.7) 3.0 (0.3) 67.5 (0.6) 21.9 (0.6) 12.7 (0.5) 23.9 § 0.0 66.3 (0.6) 33.7 (0.6) 52.8 (0.7) 26.4 (0.6)
70.95 298.77 − − 145.33 472.96 406.86 35.98 1,257.07 268.20 36.62 14.35 6.42 3.25 − 16.41 18.31
<.001 <.001 − − <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 .002 .20 − <.001 <.001
30.4 (1.3) 12.5 (0.9) 19.1 (1.1) 3.4 (0.5) 3.4 (0.4) 18.5 (1.0) 6.0 (0.6)
15.9 (0.8) 6.7 (0.5) 11.7 (0.6) 1.4 (0.2) 1.7 (0.3) 9.6 (0.6) 3.5 (0.4)
17.5 (0.5) 6.5 (0.3) 12.2 (0.4) 1.6 (0.1) 1.9 (0.1) 8.6 (0.3) 3.7 (0.2)
136.50 58.35 53.61 24.52 17.05 139.91 18.59
<.001 <.001 <.001 <.001 <.001 <.001 <.001
P
a Continuous variables were F values and categorical variables were chi-square values for homogeneity. Note: Data are presented as mean § SE or percentage (SE). Statistical significance was defined by a 2-tailed P <.05. P was obtained using 1-way ANOVA for continuous variables and Pearson chi-square test for categorical variables as a test of homogeneity. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included.
ARTICLE IN PRESS 4
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
0.97
35
0.959 0.952
30
0.96 0.95
25
0.94 0.93
20 15
0.900 10 5 0 0 me/day
1 me/day
0.92 2
EQ-5D index
0.91
Mobility
0.9
Self-care
0.89
Usual acvies
0.88
Pain/discomfort
0.87
Anxiety/depression
≥2 mes/day
Figure. The EQ-5D index and incidence of problems in the 5 dimensions based on the frequency of eating together. The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model. It ranges from 1 to 1. The Y-axis is the EQ5D incidence of problems and the X-axis is the frequency of eating together. A total of 13,920 adults (≥19 years) who had each meal more than twice a week were included. They responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016. EQ-5D indicates European Quality of Life-5 Dimensions. associated with a lower EQ-5D index and higher proportions of problems with mobility, self-care, usual activities, pain or discomfort, and anxiety or depression (P < .001). Table 2 presents results from the ANCOVA regarding the associations between the frequency of eating together and the EQ-5D index. Significantly positive associations were observed between the frequency of eating together and the EQ-5D index irrespective of adjustment for confounding variables (all P’s for trend <.001).
Associations Between the Frequency of Eating Together and Each Dimension of the EQ-5D Table 3 lists results of the multivariable logistic regression analysis of the relation between the frequency of eating together and each dimension of the EQ-5D. In models 1 and 2, the ORs for having problems in all subdimensions significantly decreased in the 1 time/d and ≥2 times/d groups compared with the eating-alone group.
Also, the ORs showed a decreasing trend as the frequency of eating together increased (P for trend <.005). After adjusting for potential confounding variables (model 3), the eating together ≥2 times/d group had a significantly lower OR for problems in the areas of self-care (0.65; 95% CI, 0.46−0.92; P = .03) and usual activities (0.66; 95% CI, 0.52−0.84; P = .003) than the eating-alone group. The eating together 1 time/d and ≥2 times/d groups had significantly lower ORs for problems with anxiety or depression
Table 2. Analysis of Covariance Between the Frequency of Eating Together and the EQ-5D Index (n = 13,920)
Model
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P
P for Trend
Model 1 Model 2 Model 3
0.907 § 0.005 0.913 § 0.005 0.815 § 0.012
0.942 § 0.003 0.930 § 0.004 0.828 § 0.012
0.949 § 0.003 0.934 § 0.003 0.831 § 0.011
<.001 <.001 <.001
<.001 <.001 <.001
EQ-5D indicates European Quality of Life-5 Dimensions. Note: The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranges from 1 to 1. Statistical significance was defined by 2-tailed P <.05 and P for trend <.05. These data evaluated the association between the frequency of eating together and the EQ-5D index using ANCOVA. Data are presented as mean § SE. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. Model 1 was adjusted for age and sex. Model 2 was adjusted for variables in model 1 and household income, education level, marital status, 1-person households, and employment status. Model 3 was adjusted for variables in model 2 and smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities.
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
5
Table 3. Odds Ratio (95% CI) of Having Problems a in Each Dimension of the EQ-5D Based on the Frequency of Eating Together (n = 13,920) Subdimensions of the EQ-5D Mobility Model 1 Model 2 Model 3 Self-care Model 1 Model 2 Model 3 Usual activities Model 1 Model 2 Model 3 Pain or discomfort Model 1 Model 2 Model 3 Anxiety or depression Model 1 Model 2 Model 3
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P
1 1 1
0.62 (0.50−0.78) 0.79 (0.63−0.99) 0.88 (0.69−1.11)
0.53 (0.44−0.64) 0.71 (0.58−0.87) 0.82 (0.66−1.02)
<.001 .005 .21
<.001 .002 .09
1 1 1
0.62 (0.43−0.88) 0.79 (0.55−1.14) 0.87 (0.60−1.26)
0.42 (0.31−0.58) 0.58 (0.42−0.80) 0.65 (0.46−0.92)
<.001 .002 .03
<.001 .001 .009
1 1 1
0.56 (0.43−0.72) 0.69 (0.53−0.89) 0.77 (0.59−1.00)
0.43 (0.35−0.54) 0.57 (0.46−0.71) 0.66 (0.52−0.84)
<.001 <.001 .003
<.001 <.001 .001
1 1 1
0.76 (0.65−0.90) 0.91 (0.76−1.09) 0.98 (0.81−1.18)
0.65 (0.56−0.75) 0.80 (0.68−0.95) 0.88 (0.74−1.04)
<.001 .01 .14
<.001 .003 .06
1 1 1
0.52 (0.43−0.63) 0.64 (0.51−0.79) 0.72 (0.57−0.90)
0.41 (0.35−0.49) 0.53 (0.44−0.64) 0.63 (0.51−0.77)
<.001 <.001 <.001
<.001 <.001 <.001
P for Trend
EQ-5D indicates European Quality of Life-5 Dimensions. Individuals who responded that there were some problems or extreme problems in each dimension of EQ-5D were classified as having problems. The EQ-5D questionnaire included questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranged from 1 to 1. Statistical significance was defined by a 2-tailed P < .05 and P for trend <.05. Statistical test performed with multivariable logistic regression. The odds ratios and 95% confidence intervals of having problems in subdimensions of the EQ-5D were calculated on the basis of the frequency of the eating-together group compared with the eating-alone group. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. Model 1 was adjusted for age and sex. Model 2 was adjusted for variables in model 1 and household income, education level, marital status, 1-person households, and employment status. Model 3 was adjusted for variables in model 2 and smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities. a
(0.72; 95% CI, 0.57−0.90; and 0.63; 0.51−0.77, respectively; P < .001) than the eating-alone group. The ORs for having problems with self-care, usual activities, and anxiety or depression tended to be of lower frequency in the eating-together groups (P for trend = .009, .001, and <.001, respectively).
Subgroup Analyses Table 4 shows the adjusted means of the EQ-5D indices on the basis of the frequency of eating together in subgroups based on age, sex, BMI, household type, and marital status after adjusting for potential confounding variables. The associations between frequency of eating together and the EQ5D index had significant interactions with age, BMI, and marital status (P for
interaction = .02, .04, and .04, respectively). In addition, the EQ-5D indices significantly increased in groups with a higher frequency of eating together in young (19−39 years) and middleaged (40−64 years) individuals, men, women, obese and nonobese individuals, and those living with others or a spouse. However, this association was attenuated in young adults after Bonferroni adjustment.
DISCUSSION This study examined the association between the frequency of eating together and HRQoL using a nationwide data set representative of the South Korean population. After adjusting for potential confounding variables, a lower frequency of eating together was
associated with a lower EQ-5D index in the areas of self-care, usual activities, and anxiety or depression than the eating-together group. Participants who ate alone tended to be older, female, and of low socioeconomic status. They also tended to have lower levels of physical activity, higher stress levels, and more comorbidities, including obesity. These findings are consistent with the results of previous studies.3,5,7,8,12,22 Furthermore, this study found that a decreased frequency of eating together was associated with a lower EQ-5D index (Table 2) after adjusting for confounding variables. These findings are in line with those from previous studies, which reported that people who ate dinner alone had a lower EQ-5D index than those who
ARTICLE IN PRESS 6
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
Table 4. Associations Between the Frequency of Eating Together and the EQ-5D Index in Subgroups (n = 13,920)
Subgroup Age, y ≤39 (n = 3,775) 40−64 (n = 6,328) ≥65 (n = 3,817) Sex Men (n = 5,697) Women (n = 8,223) Body mass index, kg/m2 <25 (n = 9,251) ≥25 (n = 4,669) Households, n Alone (n = 1,534) ≥2 members (n = 12,386) Marital status With spouse (n = 11,844) Without spouse (n = 2,076)
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P for Trenda
P for Trendb
0.914 § 0.018 0.838 § 0.013 0.811 § 0.021
0.919 § 0.018 0.851 § 0.013 0.823 § 0.021
0.924 § 0.018 0.855 § 0.012 0.826 § 0.020
.02 .001 .13
.05 .003 .39
0.838 § 0.013 0.839 § 0.011
0.848 § 0.011 0.852 § 0.011
0.852 § 0.011 0.854 § 0.011
.006 .004
.01 .008
0.853 § 0.009 0.812 § 0.015
0.864 § 0.009 0.828 § 0.013
0.865 § 0.009 0.836 § 0.013
.02 .001
.03 .002
0.808 § 0.018 0.841 § 0.009
0.819 § 0.018 0.852 § 0.008
0.818 § 0.018 0.855 § 0.008
.23 <.001
.46 <.001
0.834 § 0.009 0.848 § 0.025
0.849 § 0.008 0.856 § 0.025
0.853 § 0.008 0.856 § 0.025
<.001 .34
<.001 .67
P for Interaction .02 − − − .53 − − .04 − − .32 − − .04 − −
EQ-5D indicates European Quality of Life-5 Dimension. Values are presented without Bonferroni adjustment; bValues are presented after Bonferroni adjustment. Note: The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranged from 1 to 1. Values are presented as mean § SE using ANCOVA adjusting for age, sex, household income, education level, marital status, 1-person households, employment status, smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities. Statistical significance was defined by a 2-tailed P < .05 and P for trend <.05. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. a
ate with others.12 In addition, elderly people who ate alone were shown to have a lower subjective sense of health and happiness.23 With a decrease in the frequency of eating together, each dimension of the EQ-5D showed increased odds for health-related problems. Specifically, adults who ate alone had more problems with self-care and usual activities than those who ate with other people more than twice a day. Previous studies showed that men between the ages of 50 and 60 years, with low self-care levels (washing and dressing), were more likely to eat dinner alone.12 Eating alone had a negative impact on usual activities such as food selection,23 regular eating habits, and food preparation.7,24 Findings from previous and current studies suggested that a decreased frequency of eating together may be closely linked to health-related problems. Participants of this study who ate alone were more likely to have anxiety or depression than those who ate with other people more than once a day.
Previous studies reported that eating alone was strongly associated with depression,12,22 and that people who ate their main meal alone ≥4 times/wk were unhappy.15 Eating together is a routine social activity,25,26 and communal eating is the norm in Korean societies. Eating alone may decrease the chance of sharing food, information, and a social support system with family, friends, or members of the community.7 Therefore, people who eat alone are also more likely to feel isolated and lonely, which could lead to anxiety or depression. In this study, a decrease in the frequency of eating together was related to a lower HRQoL (without Bonferroni adjustment) in young and middle-aged individuals. The middle-aged individuals persistently showed this association even after Bonferroni adjustment. Contrary to results from previous studies regarding eating alone, which focused on elderly individuals and young children,7,10 these results suggest that a decreased frequency of eating together is related to lower HRQoL even in
younger adults, aged 19−64 years, who have more chances to eat together.27 Elderly people who eat alone had the lowest EQ-5D index, which was consistent with previous findings that elderly individuals who eat alone are susceptible to malnutrition,7 depression,10 and a poor HRQoL.28 However, the EQ-5D index did not show a significant trend across the frequencies of eating together, contrary to findings from other studies.15,23 This could be due to the fact that in elderly people who have become accustomed to living and eating alone during the aging process, physical symptoms such as cognitive impairment and sleep disturbances might have a greater impact on their HRQoL than the frequency of eating together. The association between eating alone and low HRQoL was observed in both obese and nonobese individuals. Eating alone may be a nutritional risk factor,8 and the presence of other people may affect the amount and quality of meals consumed. The incidence of metabolic syndrome was shown to be
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 high in people who eat alone,9 which is also a factor related to low HRQoL.29 In addition, a lower frequency of eating together was associated with a lower HRQoL in both men and women. A previous study suggested that because elderly women have more opportunities for social interactions besides during mealtime, the mortality risk for this group is lower than that for elderly men.30 However, it was reported that eating alone deprives women of the emotional satisfaction of caring for those who are close, which results in a feeling of social isolation and low HRQoL.24 This study found an association between the frequency of eating alone and low HRQoL in individuals living with others and those living with a spouse. Previous research suggested a higher mortality rate in older men who live with others but eat alone, compared with those who live and eat alone.30 Eating alone despite living with others could be indicative of insufficient social ties with family and inadequate care.31 These findings suggest that even people who live with others but eat alone may have low HRQoL.
Limitations This study has the following limitations. First, because the survey was based on self-reported questionnaires, recall bias might exist. Second, although KNHANES is a large-scale survey of a highly representative population, causal relationships between the different variables cannot be determined. Third, people may have diverse eating patterns, which were not considered in this study. Finally, although this study considered the confounding variables identified in previous studies, there may have been residual confounding variables.
IMPLICATIONS FOR RESEARCH AND PRACTICE This study indicates that eating together may be an important behavior related to HRQoL. These findings suggest that family and society need to pay careful attention to individuals who eat alone. It may be helpful to consider public nutrition programs
that emphasize and implement not only nutritional intervention but also programs related to eating together to improve HRQoL in individuals who usually eat alone. Future studies are warranted to confirm the association between the frequency of eating together and HRQoL in young adults, considering dietary behavior and nutrient intake. In addition, long-term prospective studies and those from other countries could be considered to confirm the effects of eating alone on HRQoL.
Choi et al
9.
10.
11.
ACKNOWLEDGMENTS The authors thank the Korea Centers for Disease Control & Prevention for providing data.
12.
REFERENCES 1. Statistics Korea. 2017 population and housing census total count result. http:// kostat.go.kr/portal/korea/kor_nw/1/ 10/1/index.board?bmode=read&aSeq= 370326. Accessed December 30, 2018. 2. Statistics Korea. Household projections by province: 2015 » 2045. http:// kostat.go.kr/portal/korea/kor_nw/1/ 2/6/index.board?bmode=read&aSeq= 362332. Accessed December 30, 2018. 3. Danesi G. Pleasures and stress of eating alone and eating together among French and German young adults. Menu J Eat Hosp Res. 2012;1:77–91. 4. Hanna KL, Collins PF. Relationship between living alone and food and nutrient intake. Nutr Rev. 2015;73: 594–611. 5. Kim CK, Kim HJ, Chung HK, Shin D. Eating alone is differentially associated with the risk of metabolic syndrome in Korean men and women. Int J Environ Res Public Health. 2018;15:E1020. 6. Cruwys T, Bevelander KE, Hermans RCJ. Social modeling of eating: a review of when and why social influence affects food intake and choice. Appetite. 2015;86:3–18. 7. Ishikawa M, Takemi Y, Yokoyama T, et al. Eating together’ is associated with food behaviors and demographic factors of older Japanese people who live alone. J Nutr Health Aging. 2017;21:662–672. 8. Tani Y, Kondo N, Takagi D, et al. Combined effects of eating alone and living alone on unhealthy dietary behaviors, obesity and underweight in older
13.
14.
15.
16.
17.
18.
19.
7
Japanese adults: results of the JAGES. Appetite. 2015;95:1–8. Kwon AR, Yoon YS, Min KP, Lee YK, Jeon JH. Eating alone and metabolic syndrome: a population-based Korean National Health and Nutrition Examination survey 2013−2014. Obes Res Clin Pract. 2018;12:146–157. Her E, Seo S. Why not eat alone? The effect of other consumers on solo dining intentions and the mechanism. Int J Hosp Manag. 2018;70:16–24. Hong SY. Association between meal companionship and depression in older adults: data from KNHANES IV-2. Korean Soc Living Environ Syst. 2016;23: 617–627. Gyoung SE, Yoon YS, Jun YY, et al. Factors associated with eating alone in Korean adults: findings from the Sixth Korea National Health and Nutrition Examination Survey, 2014. Korean J Fam Pract. 2017;7:698–706. Mapes DL, Lopes AA, Satayathum S, et al. Health-related quality of life as a predictor of mortality and hospitalization: the Dialysis Outcomes and Practice Patterns Study (DOPPS). Kidney Int. 2003;64:339–349. The EuroQol Group. EuroQol-a new facility for the measurement of healthrelated quality of life. Health Policy. 1990;16:199–208. Yiengprugsawan V, Banwell C, Takeda W, Dixon J, Seubsman SA, Sleigh AC. Health, happiness and eating together: what can a large Thai cohort study tell us. Glob J Health Sci. 2015;7:270–277. Kweon S, Kim Y, Jang MJ, et al. Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES). Int J Epidemiol. 2014;43:69–77. Lee YK, Nam HS, Chuang LH, et al. South Korean time trade-off values for EQ-5D health states: modeling with observed values for 101 health states. Value Health. 2009;12:1187– 1193. Korea Centers for Disease Control & Prevention. Validity and reliability evaluation for EQ-5D in Korea. 2011. http:// www.ndsl.kr/ndsl/commons/util/ ndslOriginalView.do?dbt=TRKO&cn= TRKO201300000474&rn=&url=&pageCode=PG18. Accessed September 1, 2019. Ko H, Song YM, Shin JY. Factors associated with alcohol drinking behavior of cancer survivors: the Korean National
ARTICLE IN PRESS 8
20.
21.
22.
23.
Choi et al Health and Nutrition Examination Survey. Drug Alcohol Depend. 2017;171:9–15. World Health Organization. Global recommendations on physical activity for health. Geneva, Switzerland: WHO Press; 2010. https://apps.who.int/iris/ bitstream/handle/10665/44399/ 9789241599979_eng.pdf;jsessionid= 043F451E7CE49BCB9873BE554276BEBA?sequence=1. Accessed September 1, 2019. Pan WH, Yeh WT. How to define obesity? Evidence-based multiple action points for public awareness, screening, and treatment: an extension of AsianPacific recommendations. Asia Pac J Clin Nutr. 2008;17:370–374. Wang X, Shen W, Wang C, et al. Association between eating alone and depressive symptom in elders: a cross-sectional study. BMC Geriatr. 2016;16:19. Kimura Y, Wada T, Okumiya K, et al. Eating alone among community-
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
24.
25.
26. 27.
28.
dwelling Japanese elderly: association with depression and food diversity. J Nutr Health Aging. 2012;16:728–731. Vesnaver E, Keller HH, Sutherland O, Maitland SB, Locher JL. Alone at the table: food behavior and the loss of commensality in widowhood. J Gerontol B. 2016;71:1059–1069. Holm L, Lauridsen D, Lund TB, Gronow J, Niva M, M€akel€a J. Changes in the social context and conduct of eating in four Nordic countries between 1997 and 2012. Appetite. 2016;103:358–368. Herman CP. The social facilitation of eating. A review. Appetite. 2015;86:61–73. Takeda W, Melby MK. Spatial, temporal, and health associations of eating alone: a cross-cultural analysis of young adults in urban Australia and Japan. Appetite. 2017;118:149–160. Kwon KM, Lee JS, Jeon NE, Kim YH. Factors associated with health-related
quality of life in Koreans aged over 50 years: the fourth and fifth Korea National Health and Nutrition Examination Survey. Health Qual Life Outcomes. 2017;15:243. 29. Saboya PP, Bodanese LC, Zimmermann PR, Gustavo AD, Assumpcao CM, Londero F. Metabolic syndrome and quality of life: a systematic review. Rev Lat Am Enferm. 2016;24: e2848. 30. Tani Y, Kondo N, Noma H, Miyaguni Y, Saito M, Kondo K. Eating alone yet living with others is associated with mortality in older men: the JAGES Cohort Survey. J Gerontol B Psychol Sci Soc Sci. 2018;73:1330–1334. 31. Kuroda A, Tanaka T, Hirano H, et al. Eating alone as social disengagement is strongly associated with depressive symptoms in Japanese communitydwelling older adults. J Am Med Dir Assoc. 2015;16:578–585.
INTRODUCTION In 2017, 28.6% of households in Korea were 1-person households,1 which is expected to increase to 36.3% by 2045.2 Accordingly, the number of people who eat alone is increasing steadily because of various factors such as living away from family for education, irregular and informal mealtimes,
working long hours, lack of time, and failed relationships.3−5 Eating a meal is a social activity that helps form relationships with other people; it goes beyond the act of putting food into one’s mouth.6 Eating alone has been reported to be associated with poor physical health leading to nutritional imbalance,7 weight problems,8 and metabolic syndrome,9 as well as
1
Department of Nursing, College of Nursing, The Catholic University of Korea, Seoul, Republic of Korea 2 Department of Biostatistics, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea 3 Department of Family Medicine, Korea University Anam Hospital, Korea University College of Medicine, Seoul, Republic of Korea Conflict of Interest Disclosure: The authors have not stated any conflicts of interest. Address for correspondence: Yong Gyu Park, PhD, Department of Biostatistics, College of Medicine, The Catholic University of Korea, 222 Banpo-daero, Seocho-gu, Seoul 06591, Republic of Korea; E-mail: [email protected]. Or: Ga Eun Nam, MD, PhD, Department of Family Medicine, Korea University Anam Hospital, Korea University College of Medicine, 82 Goryeodae-ro, Seongbuk-gu, Seoul 02841, Republic of Korea; E-mail: [email protected]. Ó 2019 Society for Nutrition Education and Behavior. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jneb.2019.11.013
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
mental health problems including stress,3 loneliness,10 and depression.11 The number of people who eat all 3 meals of the day alone has increased in the Korean population and has become an important social concern.12 Health-related quality of life (HRQoL) represents a concept of subjective physical and psychological well-being, which serves as an important indicator for evaluating chronic diseases and predicting health outcomes such as mortality and morbidity.13 Among the various tools used for measuring HRQoL, the European Quality of Life-5 Dimensions (EQ-5D) questionnaire has been recognized as suitable for the general population. Although it has only a few questions, it is divided into various profiles.14 Several previous studies showed that older adults who eat alone are unhappy15 and depressed,11 and have a lower HRQoL than those who eat with other people. However, few studies examined the association between the frequency of eating together and HRQoL considering the demographic
1
ARTICLE IN PRESS 2
Choi et al
factors, socioeconomic status, health behaviors, and comorbidities in the adult population. Therefore, this study aimed to examine the association between the frequency of eating together and HRQoL considering potential confounding variables using a nationwide data set from the Korea National Health and Nutrition Examination Survey (KNHANES),16 which is representative of the South Korean population.
METHODS Survey Overview and Study Population The KNHANES has been conducted every year by the Korea Centers for Disease Control & Prevention since 1998. The KNHANES provides national statistics on health and nutritional status, which are used to calculate the health index and evaluate health programs in South Korea. It is conducted using a systematic, stratified cluster sampling design and consists of 3 parts: a health interview, a nutrition survey, and a health examination. Of the 23,080 participants surveyed in the 2014−2016 KNHANES, this study initially evaluated 18,303 individuals aged ≥19 years. Then, 4,383 individuals were excluded because they (1) did not respond to questions regarding the frequency of eating together, and had a meal less than twice a week; (2) did not have information on socioeconomic status; or (3) did not have HRQoL data. Responses of 13,920 participants were included in the current study. The Research Ethics Committee of the Korea Centers for Disease Control & Prevention approved the survey protocol, and this study was approved by the Institutional Review Board of the Catholic University of Korea, Seoul St Mary’s Hospital (Approval No. KC18ZESI0062). The researchers obtained informed consent from all participants.
Assessment of the Frequency of Eating Together Six questions related to eating together were used in this study, the first of which was, “How many times a week did you eat breakfast in the last year?” The same question was then asked with regard to lunch and dinner.
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 Participants who responded less than twice a week in each meal were excluded. The participants were then instructed to answer Yes or No to the question, “Did you usually have breakfast with other people last year?” The same question was also asked regarding lunch and dinner. Participants who responded No to all 3 questions were assigned to the 0 times/d (eating alone) group. Those who responded No to 2 of the 3 questions were assigned to the 1 time/d eating-together group, and those who responded No to none or 1 of the 3 questions were assigned to the ≥2 times/d eating-together group.
Assessment of HRQoL The HRQoL was evaluated using the Korean version of the EQ-5D scale.15 The EQ-5D records the level of selfreported problems on the basis of 5 dimensions, including mobility, selfcare, usual activities, pain or discomfort, and anxiety or depression. Each dimension was assessed on the basis of a single question with 3 response levels (no problems, some problems, and extreme problems). The responses some problems and extreme problems were recorded as 1 point each, whereas no problems was given 0 points. The EQ-5D scores were calculated using the weighted model to transform these health states into Korean population−based health states.17 The EQ-5D index ranges from 1 to 1, in which 1 stands for perfect health status, 0 stands for death, and <0 represents health status worse than death.18 The Korean version of the EQ-5D has been evaluated for validity and reliability.18
Potential Confounding Variables Household income was divided into 4 quartiles. Education level was divided into elementary school or below, middle school, high school, and university or higher. Marital status was reported as with a spouse or without a spouse, and the household type was divided into those with 1 person or ≥2 people. Participants who answered Yes to economic activities were considered as employed. Based on smoking status, participants were classified as never-smokers, former smokers (had smoked ≥100 cigarettes during their lifetime but were
not smoking currently), and current smokers (had smoked ≥100 cigarettes during their lifetime and were still smoking). Problem drinkers were identified as those who consumed alcohol at least twice a week, with ≥7 servings/sitting (61 g) for men and ≥5 servings/ (40 g) for women.19 Physical activity (occupational, leisure, and transportation) was classified as Yes or No. Participants categorized as Yes were those who performed moderateintensity physical activity for at least 150 min/wk, high-intensity physical activity for 75 min/wk, or a combination of moderate- and high-intensity activities (1 minute of high-intensity physical activity = 3 minutes of moderate-intensity physical activity).20 Stress levels were determined on the basis of the response to the question, “How much stress do you usually feel in your daily life?” The allowed responses were very low, low, high, and very high. Participants were classified as those who had low (very low or low) or high (high or very high) stress levels. Participants were also asked whether they had ever been diagnosed with hypertension, diabetes, hyperlipidemia, stroke, heart disease (myocardial infarction or angina), arthritis (osteoarthritis or rheumatoid arthritis), or cancer. Trained medical staff measured participants’ height and body weight using standardized procedures in mobile examination centers and calculated body mass index (BMI) by dividing body weight (kg) by height squared (m); a participant with BMI ≥25 kg/m2 was defined as obese.21
Statistical Analysis All statistical analyses were conducted using SAS software (version 9.3; SAS Institute, Cary, NC; 1996) to account for the complex sampling design. The KNHANES is based on a 2-stage stratified cluster sampling method across districts and households; the sampling weights were obtained from these strata. Data are presented as mean § SE for continuous variables and as a percentage (SE) for categorical variables. The researchers compared the characteristics of participants on the basis of their frequency of eating together using 1-way ANOVA for continuous variables and Pearson chi-square test for categorical variables as a test of homogeneity. This study evaluated the
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 association between the frequency of eating together and HRQoL indices (EQ-5D index) using ANCOVA. Model 1 was adjusted for age and sex, and model 2 was adjusted for the same parameters in addition to household income, education level, marital status, 1-person households, and employment. Model 3 was further adjusted for smoking status, alcohol consumption, BMI, physical activity, stress, and comorbidities. Associations between the frequency of eating together and the 5 dimensions of the EQ-5D were analyzed by multivariable logistic regression. The odds ratios (ORs) and 95% confidence intervals (CIs) of having problems in subdimensions of the EQ5D were calculated in the group eating together 1 time/d and ≥2 times/d compared with the eating-alone group. Associations between the frequency of eating together and the EQ5D index in the relevant subgroups according to age, sex, obesity, type of
Choi et al
household, and marital status were also assessed using an ANCOVA after accounting for all potential confounding variables. Statistical significance was defined by a 2-tailed P < .05. Because there was order among the frequencies of eating together, P for trend was presented using the Cochran-Armitage test. P was presented with or without Bonferroni adjustment in the subgroup analysis.
RESULTS Characteristics of Participants Based on the Frequency of Eating Together Table 1 presents the characteristics of the study participants on the basis of the daily frequency of eating together. The group that ate together ≥2 times/d accounted for the highest proportion (67.1%), whereas the group that ate alone accounted for
3
13% of the total number of participants. The eating-alone group had the highest mean age (53.3 years) and the lowest socioeconomic status; these individuals were likely to be living alone (P < .001). The proportions of current smokers, problem drinkers, and regular exercisers were highest in the eating together ≥2 times/d group. This group also had the highest proportion of participants with high levels of stress (P < .001). The prevalence of comorbidities such as hypertension, diabetes mellitus, hyperlipidemia, stroke, cardiovascular disease, arthritis, and cancer was highest in the eating-alone group (P < .001).
Associations Between the Frequency of Eating Together and HRQoL The Figure shows that a lower frequency of eating together was
Table 1. Characteristics of Study Participants Based on Frequency of Eating Together (n = 13,920)
Characteristics Age, y ≤39 40−64 ≥65 Sex (male) Household income (lowest quartile) Education level (university or higher) Marital status (with spouse) 1-person households Occupation (yes) Current smoker Problem drinker Body mass index, kg/m2 <25 ≥25 Physical activity (yes) Subjective level of stress (high) Comorbidities Hypertension Diabetes mellitus Hyperlipidemia Stroke Cardiovascular disease Arthritis Cancer
0 Times/d (n = 814)
1 Time/d (n = 2,763)
≥2 Times/d (n = 9,343)
Statisticsa
53.3 § 0.7 27.2 (1.7) 37.7 (1.5) 35.0 (1.4) 40.4 (1.6) 36.6 (1.6) 23.9 (1.4) 73.5 (1.7) 36.3 (1.6) 43.2 (1.4) 21.2 (1.3) 8.6 (0.9) 24.1 § 0.1 64.9 (1.4) 35.1 (1.4) 46.7 (1.4) 32.2 (1.4)
44.4 § 0.4 41.2 (1.2) 46.7 (1.2) 12.1 (0.7) 39.9 (1.1) 13.4 (0.9) 37.5 (1.2) 72.0 (1.3) 11.8 (0.9) 57.7 (1.1) 20.7 (1.0) 11.4 (0.8) 23.6 § 0.1 67.8 (1.1) 32.2 (1.1) 53.5 (1.2) 28.4 (1.0)
46.2 § 0.3 37.0 (0.8) 47.8 (0.7) 15.2 (0.5) 54.0 (0.5) 12.4 (0.5) 41.3 (0.9) 79.2 (0.7) 3.0 (0.3) 67.5 (0.6) 21.9 (0.6) 12.7 (0.5) 23.9 § 0.0 66.3 (0.6) 33.7 (0.6) 52.8 (0.7) 26.4 (0.6)
70.95 298.77 − − 145.33 472.96 406.86 35.98 1,257.07 268.20 36.62 14.35 6.42 3.25 − 16.41 18.31
<.001 <.001 − − <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 .002 .20 − <.001 <.001
30.4 (1.3) 12.5 (0.9) 19.1 (1.1) 3.4 (0.5) 3.4 (0.4) 18.5 (1.0) 6.0 (0.6)
15.9 (0.8) 6.7 (0.5) 11.7 (0.6) 1.4 (0.2) 1.7 (0.3) 9.6 (0.6) 3.5 (0.4)
17.5 (0.5) 6.5 (0.3) 12.2 (0.4) 1.6 (0.1) 1.9 (0.1) 8.6 (0.3) 3.7 (0.2)
136.50 58.35 53.61 24.52 17.05 139.91 18.59
<.001 <.001 <.001 <.001 <.001 <.001 <.001
P
a Continuous variables were F values and categorical variables were chi-square values for homogeneity. Note: Data are presented as mean § SE or percentage (SE). Statistical significance was defined by a 2-tailed P <.05. P was obtained using 1-way ANOVA for continuous variables and Pearson chi-square test for categorical variables as a test of homogeneity. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included.
ARTICLE IN PRESS 4
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
0.97
35
0.959 0.952
30
0.96 0.95
25
0.94 0.93
20 15
0.900 10 5 0 0 me/day
1 me/day
0.92 2
EQ-5D index
0.91
Mobility
0.9
Self-care
0.89
Usual acvies
0.88
Pain/discomfort
0.87
Anxiety/depression
≥2 mes/day
Figure. The EQ-5D index and incidence of problems in the 5 dimensions based on the frequency of eating together. The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model. It ranges from 1 to 1. The Y-axis is the EQ5D incidence of problems and the X-axis is the frequency of eating together. A total of 13,920 adults (≥19 years) who had each meal more than twice a week were included. They responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016. EQ-5D indicates European Quality of Life-5 Dimensions. associated with a lower EQ-5D index and higher proportions of problems with mobility, self-care, usual activities, pain or discomfort, and anxiety or depression (P < .001). Table 2 presents results from the ANCOVA regarding the associations between the frequency of eating together and the EQ-5D index. Significantly positive associations were observed between the frequency of eating together and the EQ-5D index irrespective of adjustment for confounding variables (all P’s for trend <.001).
Associations Between the Frequency of Eating Together and Each Dimension of the EQ-5D Table 3 lists results of the multivariable logistic regression analysis of the relation between the frequency of eating together and each dimension of the EQ-5D. In models 1 and 2, the ORs for having problems in all subdimensions significantly decreased in the 1 time/d and ≥2 times/d groups compared with the eating-alone group.
Also, the ORs showed a decreasing trend as the frequency of eating together increased (P for trend <.005). After adjusting for potential confounding variables (model 3), the eating together ≥2 times/d group had a significantly lower OR for problems in the areas of self-care (0.65; 95% CI, 0.46−0.92; P = .03) and usual activities (0.66; 95% CI, 0.52−0.84; P = .003) than the eating-alone group. The eating together 1 time/d and ≥2 times/d groups had significantly lower ORs for problems with anxiety or depression
Table 2. Analysis of Covariance Between the Frequency of Eating Together and the EQ-5D Index (n = 13,920)
Model
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P
P for Trend
Model 1 Model 2 Model 3
0.907 § 0.005 0.913 § 0.005 0.815 § 0.012
0.942 § 0.003 0.930 § 0.004 0.828 § 0.012
0.949 § 0.003 0.934 § 0.003 0.831 § 0.011
<.001 <.001 <.001
<.001 <.001 <.001
EQ-5D indicates European Quality of Life-5 Dimensions. Note: The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranges from 1 to 1. Statistical significance was defined by 2-tailed P <.05 and P for trend <.05. These data evaluated the association between the frequency of eating together and the EQ-5D index using ANCOVA. Data are presented as mean § SE. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. Model 1 was adjusted for age and sex. Model 2 was adjusted for variables in model 1 and household income, education level, marital status, 1-person households, and employment status. Model 3 was adjusted for variables in model 2 and smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities.
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
5
Table 3. Odds Ratio (95% CI) of Having Problems a in Each Dimension of the EQ-5D Based on the Frequency of Eating Together (n = 13,920) Subdimensions of the EQ-5D Mobility Model 1 Model 2 Model 3 Self-care Model 1 Model 2 Model 3 Usual activities Model 1 Model 2 Model 3 Pain or discomfort Model 1 Model 2 Model 3 Anxiety or depression Model 1 Model 2 Model 3
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P
1 1 1
0.62 (0.50−0.78) 0.79 (0.63−0.99) 0.88 (0.69−1.11)
0.53 (0.44−0.64) 0.71 (0.58−0.87) 0.82 (0.66−1.02)
<.001 .005 .21
<.001 .002 .09
1 1 1
0.62 (0.43−0.88) 0.79 (0.55−1.14) 0.87 (0.60−1.26)
0.42 (0.31−0.58) 0.58 (0.42−0.80) 0.65 (0.46−0.92)
<.001 .002 .03
<.001 .001 .009
1 1 1
0.56 (0.43−0.72) 0.69 (0.53−0.89) 0.77 (0.59−1.00)
0.43 (0.35−0.54) 0.57 (0.46−0.71) 0.66 (0.52−0.84)
<.001 <.001 .003
<.001 <.001 .001
1 1 1
0.76 (0.65−0.90) 0.91 (0.76−1.09) 0.98 (0.81−1.18)
0.65 (0.56−0.75) 0.80 (0.68−0.95) 0.88 (0.74−1.04)
<.001 .01 .14
<.001 .003 .06
1 1 1
0.52 (0.43−0.63) 0.64 (0.51−0.79) 0.72 (0.57−0.90)
0.41 (0.35−0.49) 0.53 (0.44−0.64) 0.63 (0.51−0.77)
<.001 <.001 <.001
<.001 <.001 <.001
P for Trend
EQ-5D indicates European Quality of Life-5 Dimensions. Individuals who responded that there were some problems or extreme problems in each dimension of EQ-5D were classified as having problems. The EQ-5D questionnaire included questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranged from 1 to 1. Statistical significance was defined by a 2-tailed P < .05 and P for trend <.05. Statistical test performed with multivariable logistic regression. The odds ratios and 95% confidence intervals of having problems in subdimensions of the EQ-5D were calculated on the basis of the frequency of the eating-together group compared with the eating-alone group. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. Model 1 was adjusted for age and sex. Model 2 was adjusted for variables in model 1 and household income, education level, marital status, 1-person households, and employment status. Model 3 was adjusted for variables in model 2 and smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities. a
(0.72; 95% CI, 0.57−0.90; and 0.63; 0.51−0.77, respectively; P < .001) than the eating-alone group. The ORs for having problems with self-care, usual activities, and anxiety or depression tended to be of lower frequency in the eating-together groups (P for trend = .009, .001, and <.001, respectively).
Subgroup Analyses Table 4 shows the adjusted means of the EQ-5D indices on the basis of the frequency of eating together in subgroups based on age, sex, BMI, household type, and marital status after adjusting for potential confounding variables. The associations between frequency of eating together and the EQ5D index had significant interactions with age, BMI, and marital status (P for
interaction = .02, .04, and .04, respectively). In addition, the EQ-5D indices significantly increased in groups with a higher frequency of eating together in young (19−39 years) and middleaged (40−64 years) individuals, men, women, obese and nonobese individuals, and those living with others or a spouse. However, this association was attenuated in young adults after Bonferroni adjustment.
DISCUSSION This study examined the association between the frequency of eating together and HRQoL using a nationwide data set representative of the South Korean population. After adjusting for potential confounding variables, a lower frequency of eating together was
associated with a lower EQ-5D index in the areas of self-care, usual activities, and anxiety or depression than the eating-together group. Participants who ate alone tended to be older, female, and of low socioeconomic status. They also tended to have lower levels of physical activity, higher stress levels, and more comorbidities, including obesity. These findings are consistent with the results of previous studies.3,5,7,8,12,22 Furthermore, this study found that a decreased frequency of eating together was associated with a lower EQ-5D index (Table 2) after adjusting for confounding variables. These findings are in line with those from previous studies, which reported that people who ate dinner alone had a lower EQ-5D index than those who
ARTICLE IN PRESS 6
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
Choi et al
Table 4. Associations Between the Frequency of Eating Together and the EQ-5D Index in Subgroups (n = 13,920)
Subgroup Age, y ≤39 (n = 3,775) 40−64 (n = 6,328) ≥65 (n = 3,817) Sex Men (n = 5,697) Women (n = 8,223) Body mass index, kg/m2 <25 (n = 9,251) ≥25 (n = 4,669) Households, n Alone (n = 1,534) ≥2 members (n = 12,386) Marital status With spouse (n = 11,844) Without spouse (n = 2,076)
0 Times/d (n = 1,814)
1 Times/d (n = 2,763)
≥2 Times/d (n = 9,343)
P for Trenda
P for Trendb
0.914 § 0.018 0.838 § 0.013 0.811 § 0.021
0.919 § 0.018 0.851 § 0.013 0.823 § 0.021
0.924 § 0.018 0.855 § 0.012 0.826 § 0.020
.02 .001 .13
.05 .003 .39
0.838 § 0.013 0.839 § 0.011
0.848 § 0.011 0.852 § 0.011
0.852 § 0.011 0.854 § 0.011
.006 .004
.01 .008
0.853 § 0.009 0.812 § 0.015
0.864 § 0.009 0.828 § 0.013
0.865 § 0.009 0.836 § 0.013
.02 .001
.03 .002
0.808 § 0.018 0.841 § 0.009
0.819 § 0.018 0.852 § 0.008
0.818 § 0.018 0.855 § 0.008
.23 <.001
.46 <.001
0.834 § 0.009 0.848 § 0.025
0.849 § 0.008 0.856 § 0.025
0.853 § 0.008 0.856 § 0.025
<.001 .34
<.001 .67
P for Interaction .02 − − − .53 − − .04 − − .32 − − .04 − −
EQ-5D indicates European Quality of Life-5 Dimension. Values are presented without Bonferroni adjustment; bValues are presented after Bonferroni adjustment. Note: The EQ-5D questionnaire includes questions on mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The EQ-5D index was calculated using the weighted model; it ranged from 1 to 1. Values are presented as mean § SE using ANCOVA adjusting for age, sex, household income, education level, marital status, 1-person households, employment status, smoking status, alcohol drinking, body mass index, physical activity, stress, and comorbidities. Statistical significance was defined by a 2-tailed P < .05 and P for trend <.05. A total of 13,920 adults (≥19 years) who had each meal more than twice a week and who responded regarding their socioeconomic status and health-related quality of life in the Korea National Health and Nutrition Examination Survey in 2014−2016 were included. a
ate with others.12 In addition, elderly people who ate alone were shown to have a lower subjective sense of health and happiness.23 With a decrease in the frequency of eating together, each dimension of the EQ-5D showed increased odds for health-related problems. Specifically, adults who ate alone had more problems with self-care and usual activities than those who ate with other people more than twice a day. Previous studies showed that men between the ages of 50 and 60 years, with low self-care levels (washing and dressing), were more likely to eat dinner alone.12 Eating alone had a negative impact on usual activities such as food selection,23 regular eating habits, and food preparation.7,24 Findings from previous and current studies suggested that a decreased frequency of eating together may be closely linked to health-related problems. Participants of this study who ate alone were more likely to have anxiety or depression than those who ate with other people more than once a day.
Previous studies reported that eating alone was strongly associated with depression,12,22 and that people who ate their main meal alone ≥4 times/wk were unhappy.15 Eating together is a routine social activity,25,26 and communal eating is the norm in Korean societies. Eating alone may decrease the chance of sharing food, information, and a social support system with family, friends, or members of the community.7 Therefore, people who eat alone are also more likely to feel isolated and lonely, which could lead to anxiety or depression. In this study, a decrease in the frequency of eating together was related to a lower HRQoL (without Bonferroni adjustment) in young and middle-aged individuals. The middle-aged individuals persistently showed this association even after Bonferroni adjustment. Contrary to results from previous studies regarding eating alone, which focused on elderly individuals and young children,7,10 these results suggest that a decreased frequency of eating together is related to lower HRQoL even in
younger adults, aged 19−64 years, who have more chances to eat together.27 Elderly people who eat alone had the lowest EQ-5D index, which was consistent with previous findings that elderly individuals who eat alone are susceptible to malnutrition,7 depression,10 and a poor HRQoL.28 However, the EQ-5D index did not show a significant trend across the frequencies of eating together, contrary to findings from other studies.15,23 This could be due to the fact that in elderly people who have become accustomed to living and eating alone during the aging process, physical symptoms such as cognitive impairment and sleep disturbances might have a greater impact on their HRQoL than the frequency of eating together. The association between eating alone and low HRQoL was observed in both obese and nonobese individuals. Eating alone may be a nutritional risk factor,8 and the presence of other people may affect the amount and quality of meals consumed. The incidence of metabolic syndrome was shown to be
ARTICLE IN PRESS Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019 high in people who eat alone,9 which is also a factor related to low HRQoL.29 In addition, a lower frequency of eating together was associated with a lower HRQoL in both men and women. A previous study suggested that because elderly women have more opportunities for social interactions besides during mealtime, the mortality risk for this group is lower than that for elderly men.30 However, it was reported that eating alone deprives women of the emotional satisfaction of caring for those who are close, which results in a feeling of social isolation and low HRQoL.24 This study found an association between the frequency of eating alone and low HRQoL in individuals living with others and those living with a spouse. Previous research suggested a higher mortality rate in older men who live with others but eat alone, compared with those who live and eat alone.30 Eating alone despite living with others could be indicative of insufficient social ties with family and inadequate care.31 These findings suggest that even people who live with others but eat alone may have low HRQoL.
Limitations This study has the following limitations. First, because the survey was based on self-reported questionnaires, recall bias might exist. Second, although KNHANES is a large-scale survey of a highly representative population, causal relationships between the different variables cannot be determined. Third, people may have diverse eating patterns, which were not considered in this study. Finally, although this study considered the confounding variables identified in previous studies, there may have been residual confounding variables.
IMPLICATIONS FOR RESEARCH AND PRACTICE This study indicates that eating together may be an important behavior related to HRQoL. These findings suggest that family and society need to pay careful attention to individuals who eat alone. It may be helpful to consider public nutrition programs
that emphasize and implement not only nutritional intervention but also programs related to eating together to improve HRQoL in individuals who usually eat alone. Future studies are warranted to confirm the association between the frequency of eating together and HRQoL in young adults, considering dietary behavior and nutrient intake. In addition, long-term prospective studies and those from other countries could be considered to confirm the effects of eating alone on HRQoL.
Choi et al
9.
10.
11.
ACKNOWLEDGMENTS The authors thank the Korea Centers for Disease Control & Prevention for providing data.
12.
REFERENCES 1. Statistics Korea. 2017 population and housing census total count result. http:// kostat.go.kr/portal/korea/kor_nw/1/ 10/1/index.board?bmode=read&aSeq= 370326. Accessed December 30, 2018. 2. Statistics Korea. Household projections by province: 2015 » 2045. http:// kostat.go.kr/portal/korea/kor_nw/1/ 2/6/index.board?bmode=read&aSeq= 362332. Accessed December 30, 2018. 3. Danesi G. Pleasures and stress of eating alone and eating together among French and German young adults. Menu J Eat Hosp Res. 2012;1:77–91. 4. Hanna KL, Collins PF. Relationship between living alone and food and nutrient intake. Nutr Rev. 2015;73: 594–611. 5. Kim CK, Kim HJ, Chung HK, Shin D. Eating alone is differentially associated with the risk of metabolic syndrome in Korean men and women. Int J Environ Res Public Health. 2018;15:E1020. 6. Cruwys T, Bevelander KE, Hermans RCJ. Social modeling of eating: a review of when and why social influence affects food intake and choice. Appetite. 2015;86:3–18. 7. Ishikawa M, Takemi Y, Yokoyama T, et al. Eating together’ is associated with food behaviors and demographic factors of older Japanese people who live alone. J Nutr Health Aging. 2017;21:662–672. 8. Tani Y, Kondo N, Takagi D, et al. Combined effects of eating alone and living alone on unhealthy dietary behaviors, obesity and underweight in older
13.
14.
15.
16.
17.
18.
19.
7
Japanese adults: results of the JAGES. Appetite. 2015;95:1–8. Kwon AR, Yoon YS, Min KP, Lee YK, Jeon JH. Eating alone and metabolic syndrome: a population-based Korean National Health and Nutrition Examination survey 2013−2014. Obes Res Clin Pract. 2018;12:146–157. Her E, Seo S. Why not eat alone? The effect of other consumers on solo dining intentions and the mechanism. Int J Hosp Manag. 2018;70:16–24. Hong SY. Association between meal companionship and depression in older adults: data from KNHANES IV-2. Korean Soc Living Environ Syst. 2016;23: 617–627. Gyoung SE, Yoon YS, Jun YY, et al. Factors associated with eating alone in Korean adults: findings from the Sixth Korea National Health and Nutrition Examination Survey, 2014. Korean J Fam Pract. 2017;7:698–706. Mapes DL, Lopes AA, Satayathum S, et al. Health-related quality of life as a predictor of mortality and hospitalization: the Dialysis Outcomes and Practice Patterns Study (DOPPS). Kidney Int. 2003;64:339–349. The EuroQol Group. EuroQol-a new facility for the measurement of healthrelated quality of life. Health Policy. 1990;16:199–208. Yiengprugsawan V, Banwell C, Takeda W, Dixon J, Seubsman SA, Sleigh AC. Health, happiness and eating together: what can a large Thai cohort study tell us. Glob J Health Sci. 2015;7:270–277. Kweon S, Kim Y, Jang MJ, et al. Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES). Int J Epidemiol. 2014;43:69–77. Lee YK, Nam HS, Chuang LH, et al. South Korean time trade-off values for EQ-5D health states: modeling with observed values for 101 health states. Value Health. 2009;12:1187– 1193. Korea Centers for Disease Control & Prevention. Validity and reliability evaluation for EQ-5D in Korea. 2011. http:// www.ndsl.kr/ndsl/commons/util/ ndslOriginalView.do?dbt=TRKO&cn= TRKO201300000474&rn=&url=&pageCode=PG18. Accessed September 1, 2019. Ko H, Song YM, Shin JY. Factors associated with alcohol drinking behavior of cancer survivors: the Korean National
ARTICLE IN PRESS 8
20.
21.
22.
23.
Choi et al Health and Nutrition Examination Survey. Drug Alcohol Depend. 2017;171:9–15. World Health Organization. Global recommendations on physical activity for health. Geneva, Switzerland: WHO Press; 2010. https://apps.who.int/iris/ bitstream/handle/10665/44399/ 9789241599979_eng.pdf;jsessionid= 043F451E7CE49BCB9873BE554276BEBA?sequence=1. Accessed September 1, 2019. Pan WH, Yeh WT. How to define obesity? Evidence-based multiple action points for public awareness, screening, and treatment: an extension of AsianPacific recommendations. Asia Pac J Clin Nutr. 2008;17:370–374. Wang X, Shen W, Wang C, et al. Association between eating alone and depressive symptom in elders: a cross-sectional study. BMC Geriatr. 2016;16:19. Kimura Y, Wada T, Okumiya K, et al. Eating alone among community-
Journal of Nutrition Education and Behavior Volume 000, Number 000, 2019
24.
25.
26. 27.
28.
dwelling Japanese elderly: association with depression and food diversity. J Nutr Health Aging. 2012;16:728–731. Vesnaver E, Keller HH, Sutherland O, Maitland SB, Locher JL. Alone at the table: food behavior and the loss of commensality in widowhood. J Gerontol B. 2016;71:1059–1069. Holm L, Lauridsen D, Lund TB, Gronow J, Niva M, M€akel€a J. Changes in the social context and conduct of eating in four Nordic countries between 1997 and 2012. Appetite. 2016;103:358–368. Herman CP. The social facilitation of eating. A review. Appetite. 2015;86:61–73. Takeda W, Melby MK. Spatial, temporal, and health associations of eating alone: a cross-cultural analysis of young adults in urban Australia and Japan. Appetite. 2017;118:149–160. Kwon KM, Lee JS, Jeon NE, Kim YH. Factors associated with health-related
quality of life in Koreans aged over 50 years: the fourth and fifth Korea National Health and Nutrition Examination Survey. Health Qual Life Outcomes. 2017;15:243. 29. Saboya PP, Bodanese LC, Zimmermann PR, Gustavo AD, Assumpcao CM, Londero F. Metabolic syndrome and quality of life: a systematic review. Rev Lat Am Enferm. 2016;24: e2848. 30. Tani Y, Kondo N, Noma H, Miyaguni Y, Saito M, Kondo K. Eating alone yet living with others is associated with mortality in older men: the JAGES Cohort Survey. J Gerontol B Psychol Sci Soc Sci. 2018;73:1330–1334. 31. Kuroda A, Tanaka T, Hirano H, et al. Eating alone as social disengagement is strongly associated with depressive symptoms in Japanese communitydwelling older adults. J Am Med Dir Assoc. 2015;16:578–585.