Background: Previous studies have suggested that living in a multi-generational household (a type of family structure prevalent in Japan) confers mixed health benefits and stresses, especially for women who report such living arrangements.
Objective: To examine, in a prospective cohort study, the impact of living arrangements on the incidence of coronary heart disease (CHD) and mortality as well as all-cause mortality in a large prospective cohort of the Japanese population.
Methods: The association between living arrangements and risk of CHD and mortality was examined prospectively within a cohort of 90 987 Japanese women and men aged 40–69 years, free of prior diagnosis of cancer and cardiovascular disease. A total of 671 cases of newly diagnosed CHD, 339 CHD deaths and 6255 all-cause deaths occurred between the baseline questionnaire (1990–4) and the end of follow-up in January 2004.
Results: After adjustment for potentially confounding variables, women living in multi-generational households (ie, with spouse–children–parents; or spouse–parents) had a two- to threefold higher risk of CHD than women living with spouses only. Women living with spouses and children also had a 2.1-fold higher risk of CHD incidence compared with married women living without children.
Conclusions: Women in a multi-generational family had a higher risk of CHD, probably due to stress from multiple family roles.
Statistics from Altmetric.com
Family structure and living arrangements are viewed as important determinants of health. For example, married people have been consistently found to have better health status than unmarried people living alone—an association that is attributed not solely to health selection but also to the protective benefits of marital ties on health (eg, through financial and social support).1–5 Other studies have also examined the deleterious impact of belonging to female-headed households on both the mother’s and children’s health, mediated by the fact that such households are more likely to experience poverty-related stress.6–16
In contrast to studies of marriage and single-female-headed households, there have been fewer studies of the health consequences of other family structures. In particular, many Asian societies are marked by a high prevalence of multi-generational family structures—that is, households in which three generations (grandparents, parents and children) co-reside. Such household structures are believed to confer a mix of benefits (in the form of inter-generational transfer of care-giving and social support) and stresses. For example, Takeda et al found that women belonging to multi-generational households in Japan were simultaneously at higher risk of sedentary behaviour and not receiving regular health check-ups, but also had a lower prevalence of cigarette smoking as well as heavy drinking compared with women living alone or within nuclear families.17 In other words, women belonging to multi-generational households exhibit a mixed pattern of damaging as well as protective health behaviours; yet to our knowledge, no study has examined the net consequence of these patterns for major health outcomes such as disease incidence or mortality.
At the same time, Japanese society has been undergoing major shifts, including rapid population ageing (and with it, rising demand for care for the elderly), combined with declining marriage and fertility rates, as well as an increase in the proportion of people living alone.18 The proportion of multi-generational households in Japan decreased from 16% in 1970 to 9% in 2000, while that of married couples co-residing with their parents and/or parents-in-law also decreased from 18% in 1970 to 11% in 2000.19 Meanwhile, the proportion of people living alone increased from 20% in 1980 to 28% in 2000,19 and is estimated to rise further to about 35% by 2025.20 The population health implications of these changes in family structure and living arrangements remain unclear.
Gender differences also need to be considered in projecting the health consequences of changing family structure. Owing to a general shortage of labour (which is itself attributed to the steep decline in the fertility rate), as well as the rise in women’s autonomy achieved through higher educational attainment, more Japanese middle-aged women are employed full time than ever before.20 Yet the burden of domestic labour (including child care and care of ageing relatives) continues to fall primarily on women, even as their workforce participation has increased.21 Since many Japanese middle-aged women continue to feel a strong sense of filial responsibility towards their ageing parents and/or parents-in-law,17 the rate of co-residing with ageing parents among women aged 45–49 years has even increased slightly (29.5% in 1993, 30.4% in 1998, 36.7% in 2003), although the corresponding rate among married couples overall has decreased.17 22
A growing number of studies have reported that men and women who live alone are unmarried and socially isolated and are at increased risk of the onset and progression of cardiovascular disease.23–27 However, much less has been documented about the health consequences of family structure and living arrangements in relation to cardiovascular morbidity and mortality. In this study, we sought to examine the gender-specific relationships between living arrangement and coronary heart disease (CHD) incidence and as well as all-cause mortality in a large prospective cohort of the Japanese population.
MATERIALS AND METHODS
The first cohort of the Japan Public Health Centre-based Prospective study (JPHC study) was initiated in 1990 (cohort I) while the second cohort was initiated in 1993 (cohort II) within 11 public health centre areas throughout the country.28 The study population of this study was defined as all residents (n = 116 896) aged 40–59 years for cohort I and 40–69 years for cohort II at baseline. Of these residents, 220 residents were excluded owing to non-Japanese nationality (n = 51), late reports of emigration occurring before the start of the follow-up period (n = 166) and incorrect birth date (n = 3). Therefore, 116 676 residents remained eligible for the study. A baseline self-administered questionnaire on various lifestyles was given to residents in 1990, 1993 and 1994; 95 374 residents responded to the questionnaire and were included in the study cohort. The overall response rate was 82%. The JPHC study was approved by the institutional review board of the National Cancer Centre, Tokyo Japan.
Baseline questionnaire survey
The questionnaire included personal and family medical history, psychosocial factors such as perceived stress, occupation, personality and lifestyle factors (such as smoking and drinking, diet and physical activity). Each participant was also asked a question about their current living arrangement: “Are you living with someone (alone, spouse, children, parents, others)?” Those who answered that they were living with a spouse, children and/or parents were further categorised into the following mutually exclusive categories: “living with spouse only”, “living with spouse and children”, “living with spouse and parents”, “living with parents and children”, “living with child only”, “living with child only” and “living with spouse, children and parents”. A total of 94 616 participants provided valid response to the question. We also excluded 3629 participants from the analysis owing to a previous history of myocardial infarction, angina pectoris, stroke, or cancer at study baseline. Therefore, a total of 43 393 men and 47 594 women were available for the analysis.
Confirmation of CHD incidence
A total of 78 hospitals were registered within the sampling area of the JPHC cohort. All were major hospitals with the capability of treating patients with acute CHD. Doctors blinded to the patent’s lifestyle data reviewed the medical records at each hospital. Acute coronary events were included in the study if they occurred after the date of return of the baseline questionnaire and before 1 January 2004.
The details of the surveillance for CHD were described previously.29 Briefly, myocardial infarction was confirmed in the medical records according to the criteria of the MONICA (Monitoring Trends and Determinants of Cardiovascular Disease) project, which requires evidence from ECGs, cardiac enzymes and/or autopsy. When such an investigation was not performed but typical chest pain (⩾20 min) was reported, a probable diagnosis of myocardial infarction was made. In the absence of such a diagnosis, deaths that occurred within 1 h from onset of symptoms were regarded as sudden cardiac deaths.
Confirmation of CHD mortality
All death certificates were returned centrally to the Ministry of Health, Welfare and Labour and coded for the National Vital Statistics. Registration of death is required by the Family Registration Law and is believed to be complete in Japan. The underlying causes of death were defined according to the International Classification of Diseases, 10th Revision (ICD-10 as follows: deaths from CHD (I20–I25)). Coronary deaths were included in the analyses if they occurred after the date of return of the baseline questionnaire and before 1 January 2004.
Among the study subjects, 6112 died, 5824 moved out of the study areas and 85 were lost to follow-up within the 10-year follow-up period.
Person-years in the follow-up period were counted from the date of the return of the baseline survey until one of the following end points. For the analysis of CHD incidence, person-years were censored at the date of disease diagnosis, the date of emigration from the study area, the date of death, or the end of study period (31 December 2003), whichever came first. For the analysis of CHD deaths, person-years were censored at the date of emigration from the study area, the date of death, or the end of the study period, whichever came first. CHD incidence and death were analysed separately, and we only counted incidence or mortality event once in each analysis. For people who were lost to follow-up, the last confirmed date of their presence in the study area was used as the date of censoring.
Analysis of covariance and χ2 tests were used to compare sex-specific, age-adjusted mean values and proportions of cardiovascular risk factors. The outcomes for this study were defined as newly occurring CHD and deaths during the study period. Hazard ratios (HRs) and their 95% confidence intervals (95% CIs) were calculated after adjustment for age and other potential confounding factors using Cox proportional hazards models. Confounding variables included age (years), public health centre area, smoking status (never, former, current), ethanol intake (non- and ex-drinkers, less than weekly, <150 g/week, 150–299 g/week, 300–449 g/week, or ⩾450 g/week), body mass index (kg/m2 in quartiles), sports at leisure time (<1 day/month, 1–3 days/month, 1–2 days/week, 3–4 days/week, or almost every day), and perceived psychological stress (less, moderate, or high). We tested the assumption of proportional hazards by using both the time-dependent covariate method and a linear correlation test, and found no violation of proportionality. All analyses were conducted using the SAS statistical package, version 9.1.
During the 1 037 893 person-years of the follow-up period (median follow-up period 11.0 years) for the 90 987 subjects (43 393 men and 47 594 women), a total of 671 cases of newly diagnosed CHD (506 men and 165 women) were included in the analyses for the incidence. For the analyses of mortality, 339 CHD deaths (242 men and 97 women), and 6255 all cause deaths (4182 men and 2073 women) were included during the 1 092 161 person-years of the follow-up period (median follow-up period 11.0 years) for the 90 987 subjects.
We separately examined living arrangements in relation to cardiovascular risk factors among men and women (table 1). We found that women residing in multi-generational households (ie, “with spouse and parent”, or “with spouse, parent and child(ren)” were much less likely to engage in risk behaviours such as cigarette smoking or heavy drinking compared with women living alone or with their spouses only (table 1). The same patterns were not seen in men residing in multi-generational families. Women residing in multi-generational households were also more likely to report a higher prevalence of stress than those living alone or with spouses only. Again, among men, we did not see such differences. Both men and women residing in multi-generational households were less likely to report being physically active than their counterparts who were living alone or living with spouses only.
Table 2 shows the hazard ratios of CHD incidence according to categories of living arrangement. In model 1, we adjusted the hazard ratios for age and public health centre area only. In model 2, we additionally adjusted for stress. Finally, in model 3, we adjusted for the full range of covariates, including some (eg, smoking, drinking, physical activity) that may be considered on the way between family structure, stress and CHD incidence.
Among men, we found little indication of differences in risk of CHD incidence according to living arrangement. By contrast, we found suggestive patterns among women indicating an increased risk of CHD incidence among those belonging to multi-generational families. For example, the multivariable-adjusted hazard ratio among women living in three-generation households (with parents, spouse, and children) was 2.00 (95% CI 1.01 to 3.94) compared with women living with spouses only. Similarly the multivariate HRs among women living with “spouses and parents” and “spouses and children” were 3.03 and 2.11, respectively (both of them statistically significant) (table 2). Women living only with children (ie, in female-headed households) were also at twice the risk of CHD incidence (HR = 2.00, 95% CI 1.16 to 3.43) compared with women living with spouses.
Table 3 presents the hazard ratios of CHD mortality according to categories of living arrangement. The patterns in this table suggest that living arrangements have a different relationship with CHD mortality than with CHD incidence. Among men, two kinds of family structure stood out as being associated with increased risks of CHD mortality: men living with their parents (HR = 2.02, 95% CI 1.03 to 3.98), and men living with “others” (HR = 3.78, 95% CI 1.95 to 7.32). A similar excess risk of CHD mortality was also seen among women who lived with their parents (HR = 4.94, 95% CI 1.81 to 13.5), although the estimate was only based upon five cases.
Table 4 presents the hazard ratios of all-cause mortality according to categories of living arrangement. The results of these models echo the patterns described for CHD mortality. Men who lived with their parents, with their children (but not spouses), or with “others” appeared to be at 28–58% increased risk of total mortality. A similar pattern was also seen in women living with their parents or with “others”. We did not find an association in either men or women between multi-generational family structure and all-cause mortality.
In this large prospective analysis in a Japanese population, we found that after adjustment for potentially confounding variables, women living in multi-generational households (living with spouse–children–parents; or spouse–parents) were at double to triple the risk of CHD incidence compared with women living with spouses only. Women living with spouses and children were also at double the risk of CHD incidence compared with married women living without children. Women belonging to multi-generational households did not appear to be at increased risk of CHD mortality compared with women living with just their spouses. This suggests that although living in multi-generational households may increase the risk of incident disease, it does not affect prognosis after established disease.
Two contrasting theories have been offered about the influence of living arrangement on women’s health outcomes. Role conflict and role strain theory posits that women are more likely to experience stress (and to adopt risky health behaviours such as smoking and heavy drinking) when they are forced to take on additional roles (ie, full-time worker and informal care-giver). In contrast to role strain theory, role enhancement theory posits that added roles may benefit women’s health through increased access to resources (such as emotional and financial support).
The hypothesised mechanisms underlying the association between the stress from multiple family roles and CHD prognosis include an unhealthy profile of behaviours (such as smoking and heavy alcohol consumption), as well as exaggerated cardiovascular reactivity to stress through neuroendocrine mechanisms.30 Stress activates neuroendocrine pathways such as the hypothalamic–pituitary–adrenal (HPA) axis and the sympathetic adrenal–medullary (SAM) axis.30 The long-term activation of these axes in turn induces a sustained increase in cortisol,30 norepinephrine secretion,30 inflammatory proteins,31 platelet abnormalities32 and endothelial dysfunction,33 which may ultimately exacerbate other cardiovascular risk factors such as hypertension,30 heart rate,34 hyperlipidaemia,30 diabetes30 and the progression of atherosclerosis.30
A previous study among Japanese subjects found that women living within multi-generational households (ie, living with spouse, children and parents) showed a healthier pattern of behaviour with respect to avoidance of smoking and heavy drinking than women living with a spouse alone.17 Consistent with that study, we found that women living in multi-generational households were less likely to be current smokers or heavy alcohol drinkers. The presence of family members, especially parents, seems to exercise some measure of social control—that is, Japanese traditional norms about women smoking and having excessive alcohol consumption, which might help to prevent risk-promoting health behaviours. However, these women were also more likely to report high perceived stress. The net impact of belonging to a multi-generational household—at least for women—appeared to be deleterious with respect to the risk of CHD. By comparison, men’s CHD risk did not appear to be affected by their living arrangement.
We note several limitations to this study. First, reverse causation is possible in the association between men and women who were living with their children, parents, or “others” (but not their spouses) and CHD mortality. Poor health status is a disadvantage in the marriage market, thereby leading to a preponderance of healthy subjects in the married population.35 Thus, men and women living only with parents or others may have been doing so because they were already sick, or were being cared for by family members. The increased risk of CHD mortality among both men and women living only with parents or others in this study might be due to such reverse causation. We note that hazard estimates were not raised for CHD incidence in these same categories. Second, living arrangements were assessed in our study through a simple question relating to current living arrangements. We did not inquire further about actual exchanges of social support or other resources. Finally, it is possible that there might be over- and/or underdiagnoses of CHD in this study. However, because the criteria we used for CHD included myocardial infarction with typical chest pain (⩾20 min) and sudden death occurring within 1 h from onset of symptoms, and because everyone had easy access to emergency care under the Japanese health insurance system, it is unlikely that heart disease was overdiagnosed among women living in multi-generational households or underdiagnosed among women living with a spouse only. And also, as far as we know, no study has shown that heart disease is overdiagnosed or underdiagnosed in Japanese people according to living arrangements.
The strengths of our study are its prospective design and large sample size, enabling us to uncover the gender-specific effects of family structure and living arrangements on disease incidence and mortality. To our knowledge, this is the first prospective study to examine the association between living arrangements and disease incidence within people residing in multi-generational families.
In conclusion, living in a multi-generational family was associated with a higher risk of CHD incidence in women, probably owing to role stress. Our findings have relevance for health policy given the current increase in workforce participation among women in Japan, the decline in the marriage rate and the rapid ageing of the population in Japanese society.
We thank all staff members in each study area and in the central office for their painstaking efforts to conduct the baseline survey and follow-up.
Study group members
Members of the Japan Public Health Centre-based Prospective Study (JPHC study, principal investigator: S Tsugane) Group are: S Tsugane, M Inoue, T Sobue and T Hanaoka, National Cancer Centre, Tokyo; J Ogata, S Baba, T Mannami, A Okayama and Y Kokubo, National Cardiovascular Centre, Osaka; K Miyakawa, F Saito, A Koizumi, Y Sano, I Hashimoto and T Ikuta, Iwate Prefectural Ninohe Public Health Centre, Iwate; Y Miyajima, N Suzuki, S Nagasawa, Y Furusugi and N Nagai, Akita Prefectural Yokote Public Health Centre, Akita; H Sanada, Y Hatayama, F Kobayashi, H Uchino, Y Shirai, T Kondo, R Sasaki, Y Watanabe, Y Miyagawa and Y Kobayashi, Nagano Prefectural Saku Public Health Centre, Nagano; Y Kishimoto, E Takara, T Fukuyama, M Kinjo, M Irei and H Sakiyama, Okinawa Prefectural Chubu Public Health Centre, Okinawa; K Imoto, H Yazawa, T Seo, A Seiko, F Ito and F Shoji, Katsushika Public Health Centre, Tokyo; A Murata, K Minato, K Motegi and T Fujieda, Ibaraki Prefectural Mito Public Health Centre, Ibaraki; K Matsui, T Abe, M Katagiri and M Suzuki, Niigata Prefectural Kashiwazaki and Nagaoka Public Health Centre, Niigata; M Doi, A Terao, Y Ishikawa and T Tagami, Kochi Prefectural Chuo-higashi Public Health Centre, Kochi; H Sueta, H Doi, M Urata, N Okamoto and F Ide, Nagasaki Prefectural Kamigoto Public Health Centre, Nagasaki; H Sakiyama, N Onga, H Takaesu and M Uehara, Okinawa Prefectural Miyako Public Health Centre, Okinawa; F Horii, I Asano, H Yamaguchi, K Aoki, S Maruyama, M Ichii and M Takano, Osaka Prefectural Suita Public Health Centre, Osaka; Y Tsubono, Tohoku University, Miyagi; K Suzuki, Research Institute for Brain and Blood Vessels Akita, Akita; Y Honda, K Yamagishi and S Sakurai, University of Tsukuba, Ibaraki; M Kabuto, National Institute for Environmental Studies, Ibaraki; M Yamaguchi, Y Matsumura, S Sasaki and S Watanabe, National Institute of Health and Nutrition, Tokyo; M Akabane, Tokyo University of Agriculture, Tokyo; T Kadowaki, University of Tokyo, Tokyo; M Noda, International Medical Centre of Japan, Tokyo; Y Kawaguchi, Tokyo Medical and Dental University, Tokyo; Y Takashima, Kyorin University, Tokyo; K Nakamura, Niigata University, Niigata; S Matsushima and S Natsukawa, Saku General Hospital, Nagano; H Shimizu, Sakihae Institute, Gifu; H Sugimura, Hamamatsu University, Shizuoka; S Tominaga, Aichi Cancer Centre Research Institute, Aichi; H Iso, Osaka University, Osaka; M Iida, W Ajiki and A Ioka, Osaka Medical Centre for Cancer and Cardiovascular Disease, Osaka; S Sato, Osaka Medical Centre for Health Science and Promotion, Osaka; E Maruyama, Kobe University, Hyogo; M Konishi, K Okada and I Saito, Ehime University, Ehime; N Yasuda, Kochi University, Kochi; S Kono, Kyushu University, Fukuoka
Competing interests: None declared.
Funding: This study was supported by Grants-in-aid for Cancer Research and for the Third Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare of Japan.
Ethics approval: Ethics committee approval obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.