Study design

This study is a prospective cohort study, where participants were recruited from cohort members of the National Integrated Project for Prospective Observation of Non-communicable Disease and its Trends in the Aged (NIPPON DATA2010) [23]. This project combined data from the National Health and Nutrition Survey (NHNS2010) [24] and the Comprehensive Survey of Living Conditions of the People on Health and Welfare (CSLC2010) [25]. These surveys were conducted by the Ministry of Health, Labour and Welfare of Japan. Briefly, in 2010, the CSLC was conducted to survey national living conditions in approximately 290,000 households from 5510 randomly selected areas throughout Japan. Of the 5510 areas, 300 areas were selected to conduct NHNS to survey lifestyles and collect blood samples in the same year, when trained interviewers obtained informed consent for NIPPON DATA2010. A total of 2898 participants (1239 men and 1659 women aged 20 years and older) in these areas agreed to participate in the baseline NIPPON DATA2010 survey. Of the participants, seven were excluded because it was impossible to merge the data. The remaining 2891 participants (1236 men and 1655 women) provided baseline data. We included 1928 participants aged 30 years and older as of 2020 in this study, after excluding participants who were lost to follow-up (n = 647) or who had data missing from the questionnaires or on marital status and household size at baseline (n = 316). The response rate was 93.4%.

Ethics statement

We obtained written informed consent from all the participants. This study was approved by the Institutional Review Board of Shiga University of Medical Science (No. 22–29, 2010) and the Fukuoka University Clinical Research and Ethics Center (U21-09-001).

Socioeconomic factors

Data on annual household income, employment status, years of education, and household size at baseline were obtained using self-administered questionnaires for NHNS2010, CSLC2010, and NIPPON DATA2010. Annual household income was classified into high (≥6 000 000 yen), middle (2 000 000–6 000 000 yen), and low (<2 000 000 yen), and employment status was defined as employed or self-employed, underemployed, or unemployed. Years of education were classified into three categories: <9, 9–12, and ≥13 years. Marital status and household size were categorized as married individuals or singles living with someone and singles living alone. The widowed or separated individuals were classified into a single group.

Other factors

At baseline, when body weight and height were measured, participants wore light clothing without shoes. Each participant’s body mass index (BMI) was calculated by dividing their body weight in kilograms by height in meters squared. Smoking and drinking habits were classified into current, past/non-smoker, and drinker groups.

Unhealthy changes in lifestyles during the COVID-19 pandemic

We distributed questionnaires on differences in body weight and lifestyle behaviors before and during the lockdown in Japan from April to May 2020, in October 2020 [2]. The topics covered by the questionnaires included changes in body weight (≥ +3 kg, +1 - +3 kg, -1 - +1 kg, -3 - -1 kg, or < - 3 kg), and frequency (increased, not changed, or decreased) of takeout food ordering, physical activity, eating between meals, and consumption of alcohol and vegetables from before to during the first COVID-19 pandemic in Japan. Physical activity included some activity indices, including sports, occupation, and housework. Unhealthy changes were defined as follows: 1) weight gain (≥ +3 kg), 2) an increase in frequency of eating between meals, or 3) alcohol consumption, 4) a decrease in frequency of vegetable consumption, or 5) physical activity, and 6) at least one unhealthy change mentioned above.

Statistical analysis

The Kruskal–Wallis test was used to identify mean differences in age and BMI as continuous variables. For categorical variables, Fisher’s exact test was used to examine whether there were any differences in the observed distributions among the marital status and household size categories. Next, we stratified the number and percentage of unhealthy changes according to marital status and household size. We then performed a multivariate logistic regression analysis to examine whether marital status and household size were associated with unhealthy lifestyle changes after adjusting for covariates (age, sex, BMI, annual household income, employment status, years of education, and smoking and drinking habits at baseline). Socioeconomic factors were included in the model because they are also regarded as determinants of health from the perspective of affordability, accessibility, and availability of healthcare information and resources [26–29]. All analyses were performed by dividing the participants into two age groups with an age cut-off of 65 years. Statistical significance was set at P<0.05. All analyses were performed using Stata SE version 16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC.) and SAS (version 9.4; SAS Institute, Inc., Cary, NC).

The unhealthy changes in lifestyles among older people living alone

Older adults living alone showed higher odds of having experienced at least one negative lifestyle change than married people. To our knowledge, no study has examined the association between unhealthy lifestyle changes during the pandemic, marital status, and household size among older adults. According to Lehtisalo et al. [22] and Salman et al. [18], older people living alone were more likely to experience decreased physical activity, decreased vegetable consumption, and increased loneliness than those living with someone during the lockdown in 2020. Although the authors did not assess the impacts of marital status on the outcomes, these findings would reflect the impacts of older adults living alone, since global data show that more than 90% of all one-person households are unmarried in developed countries [30].

Although social isolation does not necessarily parallel to emotional loneliness, Tilburg [31] documented people living alone manifested significantly more emotional loneliness than those who living with a spouse or partner. Hearne [19] also showed singles living alone were at a higher risk of psychological distress than those married and co-habiting during the lockdown. Loneliness and low levels of life satisfaction induce decreased physical activity [18], increased alcohol consumption [32] and emotional eating, which includes overeating of high-calorie foods or snacks to mitigate psychological stress [20]. Meanwhile, Poelman et al. [33] and Birditt et al. [17] documented that older people were less likely to perceive lockdown-related stress or change their lifestyles than younger people because they generally have fewer chances to face abrupt financial crisis due to the subsequent extreme psychological stress. However, these findings may have been modified by marital status and household size.

Over the decades, Japan has been facing super-aging challenges; the population aged ≥ 65 was approximately 28.8% (36 million) in 2020 and is expected to increase to 35% (39 million) by 2040 [34]. In addition, the percentage of one-person households among older people has doubled (27%) over 30 years [34, 35]. Decreased physical activity, unhealthy eating, and psychological stress predispose older people to impaired cognitive function and immune system [36, 37], sarcopenia [38], and obesity [20], which will impose a considerable burden of medical costs capacity even after convergence of COVID-19 pandemic. Furthermore, missing healthcare visits, decreased health care services, and highly recommended cocooning due to their vulnerability to infection exacerbated pre-existing chronic health conditions, including obesity, hypertension, and diabetes, which globally spread alongside the COVID-19 pandemic [39]. To tackle this problem, social cohesion and digital inclusion for older people living alone are crucial. Developing an age-friendly digital platform allows them to keep communication channels open to maintain healthcare intervention and social cohesion, including online exercise, medical appointments, and interaction with other people, keeping safety amid the infectious disease crises. Notably, Fingerman et al. [40] suggested older people living alone are more reactive than those living with someone to social contact with friends or care service providers.

Socioeconomic factors and unhealthy changes in lifestyles

We found that higher income and education were likely to associate with unhealthy lifestyle changes, especially weight gain of more than 3 kg, less physical activity, and more alcohol consumption. One plausible explanation is the prevalence of teleworking, the rate of which almost tripled among Japanese employees during the first lockdown in 2020 [41]. Larger companies are more likely to offer to telework [42], which causes prolonged stay-at-home periods, and subsequently decreases the frequency of physical activity [43, 44]. In this sense, the implementation of teleworking may also underlie the association between higher education and unhealthy changes. The association between older people with higher income and more alcohol consumption might be reasoned by more disposable money to spend on alcohol. In fact, a multi-cohort study of the UK [45] demonstrated decreased alcohol consumption in the younger generation, who had more financial difficulties than the older generation during the lockdown. However, younger people with higher incomes did not show a significant increase in alcohol consumption compared to those with lower incomes. Some epidemiological studies reported that younger people appeared to prefer on-site drinking to drinking at home; therefore a sweeping closure of pubs and bars during the pandemic might not have led to increased alcohol consumption at home among younger people [32, 45].

Weight gain during the COVID-19 pandemic

The younger singles had higher odds of weight gain (≥ 3 kg) than the married ones during the pandemic. While previous studies using online surveys showed no significant association between marital status and weight gain during lockdown [46–48], those who live in smaller households have been shown to be associated with less physical activity [22, 49], poorer sleep quality [50], and more psychological stress [19, 22, 51], leading to weight gain. Percentages of those who regularly work out in gyms or outside might have affected the association because of the closure of sports facilities [52].

The strengths of our study are the large sample size, high response rate, and wide age range of participants in the nation-wide survey. Quite a few surveys on lifestyle changes during the COVID-19 pandemic offered online questionnaires, where respondents are likely to be biased toward younger generations. Nonetheless, we have to mention several limitations. First, owing to the nature of a survey using questionnaires, we cannot eliminate the possibility of recall bias and attrition bias. Second, the percentages of essential workers, those in healthcare, delivery services, agriculture, and retail establishments might have affected our results, since they had fewer opportunities to stay at home, which would have potentially induced behavioral changes during the pandemic [53]. Third, methodological differences, including the duration of the research period, sample selection, and types of variables included in the quantitative analyses, may have caused discrepancies across studies. Fourth, there is a possibility that selection bias limited the internal validity of our study. Fifth, there might be changes in marital status and household size during the study period. However, a recent study reported both the percentages of marriage and divorce decreased by 10–20% during the pandemic possibly due to the stay-at-home recommendation in Japan [54]. Additionally, non-differential misclassification like this generally leads to a null hypothesis [55], however, we identified the significant association between marital status, household size, and unhealthy changes, as well as age effects. Although the household income of participants also may have changed from the baseline, the impacts would be trivial since over ten years the average annual change rates in household income is approximately 4%, which is hardly expected to change the categories participants belong. Lastly, we did not ask the content of foods or estimate the intake of nutrients using food frequency questionnaires to define unhealthy eating habits.

Collaborators

NIPPON DATA2010 research group. Co-principal investigators: Katsuyuki Miura (Shiga University of Medical Science, Otsu, Shiga) and Akira Okayama (Research Institute of Strategy for Prevention, Tokyo).

Research members: Hirotsugu Ueshima (Shiga University of Medical Science, Otsu, Shiga), Shigeyuki Saitoh (Sapporo Medical University, Sapporo, Hokkaido), Kiyomi Sakata (Iwate Medical University, Yahaba, Iwate), Atsushi Hozawa (Tohoku University, Sendai, Miyagi), Yosikazu Nakamura (Jichi Medical University, Shimotsuke, Tochigi), Tomonori Okamura (Keio University School of Medicine, Tokyo), Nobuo Nishi (National Institute of Health and Nutrition, National Institutes of Biomedical Innovation, Health and Nutrition, Tokyo), Takayoshi Ohkubo (Teikyo University School of Medicine, Tokyo), Yoshitaka Murakami (Toho University, Tokyo), Toshiyuki Ojima (Hamamatsu University School of Medicine, Hamamatsu, Shizuoka), Hideaki Nakagawa (Kanazawa Medical University, Uchinada, Ishikawa), Yoshikuni Kita (Tsuruga Nursing University, Tsuruga, Fukui), Aya Kadota, Yasuyuki Nakamura, Naomi Miyamatsu (Shiga University of Medical Science, Otsu, Shiga), Takehito Hayakawa (Ritsumeikan University, Kyoto), Nagako Okuda (Kyoto Prefectural University, Kyoto), Katsushi Yoshita (Osaka City University Graduate School of Human Life Science, Osaka), Yoshihiro Miyamoto, Makoto Watanabe (National Cerebral and Cardiovascular Center, Suita, Osaka), Naoyuki Takashima (Kindai University Faculty of Medicine, Osaka-Sayama, Osaka), Akira Fujiyoshi (Wakayama Medical University, Wakayama), Kazunori Kodama, Fumiyoshi Kasagi (Radiation Effects Research Foundation, Hiroshima), Yutaka Kiyohara (Hisayama Research Institute for Lifestyle Diseases, Hisayama, Fukuoka), Hisatomi Arima (Fukuoka University, Fukuoka), Toshiharu Ninomiya (Kyushu University, Fukuoka).