The effect of having a spouse on the mortality risk in later life --Empirical evidence from China

doi:10.21203/rs.3.rs-2599005/v1

Download PDF

Research Article

The effect of having a spouse on the mortality risk in later life --Empirical evidence from China

https://doi.org/10.21203/rs.3.rs-2599005/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

Examines the impact of the presence or absence of a spouse on the risk of death among older Chinese people and explores issues such as differences in mortality risk among different populations.

Method

Using the five-period longitudinal data of the China Health Impact Survey on the Elderly (CLHLS) from 2005 to 2018 as the basis, a Cox risk regression model was constructed using the Chinese elderly population aged 65 to 105 years as the study population, with the risk of death among the elderly as the dependent variable and the presence of a spouse among the elderly as the core independent variable.

Results

Controlling for variables such as individual characteristics of the elderly, socio-economic characteristics and health status, the presence or absence of a spouse had a significant effect on the risk of death in older people, with older people with a spouse having a 45.2% lower risk of death than older people without a spouse, an effect that was more prevalent in older people, older people who were male, older people in rural areas, and older people without financial support from children or pension support. Further analysis of the risk of death without a spouse was widowed > divorced > unmarried.

Conclusion

Based on a sample of older people in China, we found that having a partner in later life is a worthwhile life goal, i.e. having a partner in old age is effective in reducing the risk of death in later life and is an important factor in coping with ageing and enjoying a high quality of life in later life. Being divorced, widowed and single is detrimental to health, especially in older, male, rural and insecure older people. To reduce the negative impact of spouselessness in later life on the risk of death in old age, we need to focus on the elderly, male elderly, rural elderly and the insecure elderly.

Spouse

older adults

The Cox model

Chinese Longitudinal Healthy Longevity Survey

In reality, we often find that people with intimate partners have a lower mortality risk than those who lack close friends and spouses, it means people who have a partner are happier, healthier and live longer than those who live alone. When we are old, our ability is declining, the spouse, as the closest person in the old age, plays a decisive role in the life and health of the elderly, and the support provided by the spouse is particularly important. It is generally accepted that marriage, as an important social factor, has a significant impact on the level of death among the elderly people.

Regarding the effect of have a spouse on the mortality risk in the older people, both cross-sectional and panel data demonstrate the protective effect of spouses on human lifespan, which is consistent across countries and cultures; There is a "protective effect" of spouses on human life and health, and the absence of spouses has a "damaging effect" on human life (Lerman, 2002). Manzoli et al. (2007) reviewed the relationship between marriage and mortality risk since the mid-19th century, and compared and predicted mortality in the two groups with and without a spouse, resulting in a 18% lower mortality risk in the spouse population than in the non-spouse population, and the protective effect is stronger in Europe and North America than in Asia; Scanfato et al. (2008) analyzed Italian 10-year follow-up data and found that elderly people with spouse lived longer; GoldmanN et al. (1993) found that marriage can reduce the mortality risk of the couple elderly by 5%-15% compared with the elderly without a couple. Due to China's unique family traditional culture, the effect of marriage on the risk of death in the elderly in China may be stronger (Huang Qingbo, 2014), among which Li Jianxin (2022), Hao Hongsheng (1995), You Yunzhong (2009) and Liu Zheng (1986) conducted a comparative analysis of mortality under different marital conditions, the first three were based on Chinese census data, the latter was based on Beijing mortality survey data, all came to the nearly identical conclusion that the mortality rate was significantly higher for unmarried people than for married ones.

On the issue of gender heterogeneity, scholars from all walks of life have reached inconsistent conclusions. Zick and Smith(1991) believed that the survival benefits of spouses may only benefit men, but Lillard and Waite(1995) found the protective effect on women. Manzoli et al. (2007) found no difference in the protective effect of spouses for men and women in a meta-analysis of mortality in old age. With regard to age heterogeneity, Zeng Yi (2004) found that the death probability of the elderly without spouses was higher than that of the elderly with spouses, while the death probability of the younger elderly was not significantly affected by the marital status, which may indicate that the "damaging effect" of marriage has a long-term accumulation. However, Zhao Xiaohang et al. (2019) believed that the absence of spouses has a serious impact on the younger elderly. With the passage of time, the influence on the elderly is relatively weakened. The reason for the opposite conclusion may be that the tracking of different generations of the elderly has a great impact on people's perceptions.With regard to urban and rural heterogeneity, Li Anqi et al. (2019) argued that having a spouse can significantly improve the mental health status of the elderly in rural areas, and men can gain more benefits. In addition, other studies have shown that a higher marital history also increases the risk of death for both men and women, excluding the influence of age (G Tatangelo, 2017). Based on the results of the multi-level model and SAS hybrid program, Heejeong Choi (2013) concluded that the health benefits of having or not having a spouse may be amplified in individuals with higher levels of education and income.

According to the current research results, most scholars' studies on the influence of spouse on death risk have reached similar conclusions. First, few of these studies specifically discuss the influence of spouse or not on death risk of elderly Chinese. Although Li Jianxin (2022), Hao Hongsheng (1995), You Yunzhong (2009) and Liu Zheng (1986) have discussed, they are not based on the more representative data of the Chinese Longitudinal Healthy Longevity Survey(CLHLS). Although Zeng Yi (2004) also discussed, he did not carry out in-depth analysis on the types with or without spouses, such as the differences in the impact of widowed spouse, divorce and singledom on the risk of death, such as the impact of having or not having a spouse on the risk of death under different genders, types of household registration and children's economic support.

Second, because Chinese people's cultural views on family comfort may be different from those in the West, there are few studies on the death risk of this elderly group in China. Third, some scholars use the data of the Chinese Longitudinal Healthy Longevity Survey to study family issues, but the samples do not strictly track the same group of individuals, so that the differences in individual socioeconomic environment have an impact on the research results. Based on this, this paper, based on the data of the Chinese Longitudinal Healthy Longevity Survey (hereinafter referred to as CLHLS), explores the influence of having or not having a spouse on the death risk of the elderly in China, and analyzes the heterogeneity of gender, age, household registration type and economic support with or without children. At the same time, no spouse is divided into widows, divorces, unmarried, and unmarried. To judge the risk of death and analyze the causes.

2.1 Data sources and samples

In this study, the Chinese Longitudinal Healthy Longevity Survey (CLHLS) was used as the data source. The database included social science surveys on the health of the elderly in 23 provinces and autonomous regions, and the respondents were aged 65 and above and adult children aged 35–64. The data is a comprehensive and systematic statistics of the health survey of the elderly population in China. The questionnaires are divided into two types: the questionnaire of the surviving respondents and the questionnaire of the bereaved family members. The data of the living conditions and death date of the deceased and survivors are collected.

This paper selected five longitudinal data of CLHLS database in 2005, 2008, 2011, 2014 and 2018. In terms of specific data processing, first of all, the deletion process was carried out for the respondents without specific time of death during the investigation period. Secondly, some interviewees with more missing values were excluded and were not suitable for inclusion in the model. Finally, the sample of respondents older than 105 years was deleted, and the sample of respondents between 65 and 105 years old was retained as the research object. The base sample size included in the statistical model was 10409. During the observation period from 2005 to 2018, a total of 8581 elderly people died and 1828 elderly people were lost to follow-up.

2.2 Variables

2.2.1 Dependent variable: death risk

The dependent variable in this paper is death risk, which refers to the possibility of death faced by disabled elderly people from 2005 (base period) to 2018 (end period), including both survival state and survival time. The survival time of the respondents was calculated by subtracting the 2005 survey date from the date of death for the respondents who died during the survey period, and subtracting the 2005 survey date from the survey termination date for the respondents who were lost to follow-up during the survey period (except before the 2008 survey). Finally, the average survival time of the elderly is 50.22 months.

2.2.2 Independent variable: marital status

The core independent variable of this paper is marital status. According to the data questionnaire of the Chinese Longitudinal Healthy Longevity Survey, when respondents were asked about their current marital status, their responses were divided into five states: married, unmarried, remarried, widowed and divorced. We classified married and remarried respondents as having a spouse; widowed, divorced and unmarried respondents as having no spouse. Finally, the results of the sample marital status obtained in this paper are as follows: 3,058 samples have spouses and 7,363 samples have no spouses.

2.2.3 Control Variables

In this paper, the factors affecting marital status and death risk were controlled from the aspects of individual characteristics, economic characteristics and health characteristics, so as to reduce the estimation deviation caused by missing variables as much as possible. Personal characteristics include gender, age, residence, and Educational attainment. Economic characteristics include career level before the age of 60, Pension entitlement, financial support for children, and adequate medical care in childhood, chronic disease, and Health self-assessment.

The individual characteristics of the elderly often directly affect their mortality risk. The differences of male and female physiological and psychological characteristics, age, residence conditions and education level all have a certain degree of influence on the mortality risk of the elderly. In the gender variable, we divided males into 1 and females into 0; In terms of age variables, we divided 60–69 into 1, 70–79 into 2, 80–105 into 3 according to the Chinese elderly population of the young, middle and old division. In the residence variable, we divided towns into 1 and rural areas into 0. In the education variable, we divided the uneducated into 1, the elementary school education into 2, and the junior high school and above into 3.

The economic characteristics of the elderly often determine whether they are able to enjoy a high quality of life in later life, indirectly affecting their mortality risk. The occupational level before the age of 60 is one of the important indicators to measure the income level, which directly determines the economic level of the elderly. We divided the senior level into 1 and the lower level into 0; Pension is the welfare of the elderly and the exclusive economic source of the elderly group. We divided those with pension into 1 and those without pension into 0. Elderly people may also have children to provide them with a certain source of money in their later years, so we divided those who have children's economic support into 1, otherwise 0.

The health characteristics of the elderly indicate their current physical condition and are animportant factor influencing their risk of death. The presence or absence of adequate health care in childhood tends to influence whether an elderly person has a healthy physical foundation; we scored the presence of adequate health care in childhood as 1 and the absence as 0. We used chronic diseases to measure the health status of the elderly before death, where having any of the chronic diseases in the questionnaire is 1 and 0 otherwise; health self-assessment is the elderly person's perception of their own health and we scored poor as 1, average as 2 and good as 3. The descriptive statistics for each variable are shown in Table 1:

Table 1

Statistical table of variables
Variables		Frequency	Percentage (%)	Average Value	Standard Deviation
Gender	Famele = 0	5885	56.47	0.44	0.50
Gender	Male = 1	4536	43.53	0.44	0.50
Age	(60–69)=1	773	7.42	2.67	0.61
	(70–79)=2	1934	18.56
	(80–105)=3	7714	74.02
residence	Rural areas = 0	5965	57.24	0.43	0.49
residence	Towns = 1	4456	42.76	0.43	0.49
Educational attainment	Uneducated = 1	6540	62.76	1.47	0.66
	primary school = 2	2902	27.85
	Junior high school and above = 3	979	9.39
Career level before the age of 60	High-level = 1	804	7.72	0.07	0.27
Career level before the age of 60	Low-level = 0	9617	92.28	0.07	0.27
Pension entitlement	Yes = 1	2068	19.84	0.18	0.39
Pension entitlement	No = 0	8353	80.16	0.18	0.39
Financial support from children	Yes = 1	6839	65.63	0.66	0.47
Financial support from children	No = 0	3582	34.37	0.66	0.47
Adequate medical care in childhood	Yes = 1	3561	34.17	0.34		0.47
Adequate medical care in childhood	No = 0	6860	65.83	0.34		0.47
Chronic disease	Yes = 1	6970	66.88	0.67		0.47
Chronic disease	No = 0	3451	33.12	0.67		0.47
Health self-assessment	Bad = 1	2576	24.72	2.20	0.81
	Normal = 2	3227	30.97
	Well = 3	4618	44.31

2.3 Model Setting

Cox risk regression model does not require the assumption of the distribution of baseline risk function, and can effectively correlate the event occurrence time and its influencing factors, which is an important method for event-oriented data survival analysis. Since Cox is able to discuss the relationship between multiple factors and survival, analyze the truncated data involved in the research process, and use Kaplan-Meier method to draw the survival curve to verify that the risk ratio of each single factor satisfies the assumption of equal proportion, this paper adopts Cox risk regression model to analyze the covariable of the influence of the elderly death risk. In this paper, the covariable affecting the death risk of the elderly is defined as $\text{X}=({\text{x}}_{\text{j}1},\dots {\text{x}}_{\text{j}2}{)}^{\text{t}}$, $\text{ℎ}\left(\text{t}\right|\text{X})$ is the risk factors for X, the basic formula of the regression model is:

$$\begin{array}{c}ℎ\left(\text{t}|\text{X}\right)={\text{ℎ}}_{0}\left(\text{t}\right)c\left({{\beta }}^{\text{t}}\text{X}\right)\#\left(1\right)\end{array}$$

Among, ${\text{ℎ}}_{0}\left(\text{t}\right)$ is the benchmark risk function,${\beta }=({{\beta }}_{1},\dots {{\beta }}_{\text{p}}{)}^{\text{t}}$ i, the coefficient vector of the covariable༌$\text{c}\left({{\beta }}^{\text{t}}\text{X}\right)$ is the param, ter model of the covariable。Due to $\text{ℎ}\left(\text{t}|\text{X}\right)$ is positive༌$\text{c}\left({{\beta }}^{\text{t}}\text{X}\right)$ generally in exponent, al form: $\text{c}\left({{\beta }}^{\text{t}}\text{X}\right)=\text{exp}\left({{\beta }}^{\text{t}}\text{X}\right)=\text{e}\text{x}\text{p}（\sum _{\text{k}=1}^{\text{p}}{{\beta }}_{\text{k}}{\text{X}}_{\text{k}}）$。Thus $\text{ℎ}\left(\text{t}|\text{X}\right)={\text{ℎ}}_{0}\left(\text{t}\right)\text{exp}\left({{\beta }}^{\text{t}}\text{X}\right)={\text{ℎ}}_{0}\left(\text{t}\right) \text{e}\text{x}\text{p}（\sum _{\text{k}=1}^{\text{p}}{{\beta }}_{\text{k}}{\text{X}}_{\text{k}}）$。

If the covariable vectors of the two old people samples are respectively $\text{X}$ and ${\text{X}}^{\ast }$,so, the risk ratio between the two is:

$$\frac{\text{ℎ}\left(\text{t}|\text{X}\right)}{\text{ℎ}\left(\text{t}|{\text{X}}^{\ast }\right)}=\frac{{\text{ℎ}}_{0}\left(\text{t}\right) \text{e}\text{x}\text{p}（\sum _{\text{k}=1}^{\text{p}}{{\beta }}_{\text{k}}{\text{X}}_{\text{k}}）}{{\text{ℎ}}_{0}\left(\text{t}\right) \text{e}\text{x}\text{p}（\sum _{\text{k}=1}^{\text{p}}{{\beta }}_{\text{k}}{{\text{X}}^{\ast }}_{\text{k}}）}=\text{e}\text{x}\text{p}[{\sum }_{\text{k}=1}^{\text{p}}({\text{X}}_{\text{k}}-{\text{X}}_{\text{k}}^{\ast })\left] \right(2)$$

After the baseline risk rate is eliminated, the time pattern of risk rate changes is eliminated, and the above risk ratio becomes a constant that does not change over time. Covariates affect the improvement or reduction of the whole benchmark risk rate in proportion, so Cox model is also called proportional model, and risk ratio is also called relative risk, which is the relative risk faced by individuals with certain characteristics relative to the reference class (such as gender variable, one of which is set as the reference class in this paper). If the covariable is univariate and the value is 0–1, the risk ratio can be simplified as:

$$\begin{array}{c}\frac{\text{ℎ}\left(\text{t}|\text{X}\right)}{\text{ℎ}\left(\text{t}|{\text{X}}^{\ast }\right)}=ex\text{p}\left({{\beta }}_{1}\right)\#\left(3\right)\end{array}$$

Cox analysis can effectively deal with the uncertain distribution of survival time, and has the ability to investigate the effect of multiple influencing factors on survival time. Therefore, in the end, Cox was selected to gradually control the influence of multiple factor analysis on the elderly death risk, and the analysis tool was Stata17.0 statistical analysis software.

3.1 cumulative hazard function on spouse status

In order to gain a preliminary understanding of the relationship between with or without a spouse and the mortality risk in the elderly, using the Nelson-Aalen survival analysis method, cumulative hazard function curves were plotted to visualise the mechanism by which with or without a spouse affects the probability of mortality risk in the elderly. As illustrated in Fig. 1, the cumulative instantaneous mortality rate in the curve rises gradually with analysis time, the trends are almost identical, and having a spouse could reduce the mortality risk for the elderly while having no spouse increases the mortality risk for the elderly.

3.2 cox proportional-hazards model regression results

As shown in Table 2, The model analysis found that with or without a spouse had a significant effect on the mortality risk for the elderly, the elderly who have a spouse having 45.2% lower mortality risk than these who have no spouse. Models 2 to 4 are stepwise additions of control variables such as individual characteristics, socio-economic characteristics, health status variables, three regression models are obtained for the mortality risk for the elderly. After controlling for all variables, have spouse still had a significant effect on the mortality risk, with the direction and level of significance remaining unchanged, the inclusion of individual characteristics, socio-economic characteristics and health status variables also resulted in a greater reduction in the mortality risk, suggesting that control variables such as individual characteristics have a protective effect on the mortality risk in the elderly. Among individual characteristics, the mortality risk is significantly higher for men than for women by 14%, for middle-aged elderly by 30.1% compared to younger elderly, and for older elderly by about twice as much as younger elderly. The poeple who have higher education level, have the lower mortality risk, however the significance of education level decreases as the socioeconomic characteristics, health status variables are controlled, suggesting that the effect of education level on the mortality risk is generated by the correlation between socioeconomic characteristics and health status. Among the socio-economic characteristics, occupation before age 60 and pension have no significant effect on the mortality risk among the elderly, while financial support from children have a significant effect on the mortality risk among the elderly, with those with financial support from children having an 18.3% lower mortality risk than those without financial support from children, and significant at the 1% level. The reason for this maybe that elderly people without pensions are more likely to have financially supported by their children's children, and this perceptual tendency weakens the impact of pensions on mortality risk, which in China is somewhat compensated by intergenerational financial support from the family, which weakens the impact of pensions on mortality risk due to the incomplete coverage of pensions in China. For the health conditions, health self-assessment also has a significant effect on the mortality risk in the elderly people, with the better the health self-assessment, the lower the mortality risk in the elderly, confirming that health self-assessment (ADL) is a good measure of health status for elderly people.

Table 2

Basic regression results
		Model 1	Model 2	Model 3	Model 4
Marital status	Have a spouse (have no spouse)	0.548*** (0.014)	0.770*** (0.022)	0.792*** (0.023)	0.782*** (0.022)
Gender	Male (Female)		1.140*** (0.030)	1.171*** (0.031)	1.205*** (0.032)
age	Medium age(Lower age)		1.323*** (0.078)	1.301*** (0.077)	1.303*** (0.077)
age	High age (Lower age)		3.186*** (0.176)	3.056*** (0.170)	3.059*** (0.170)
Residence	City (rural)		0.989 (0.023)	0.932** (0.026)	0.934** (0.026)
Educational attainment	primary school (uneducated) Junior high school and above (uneducated)		0.912* (0.026) 0.784* (0.035)	0.933 (0.027) 0.851* (0.042)	0.944 (0.027) 0.862 (0.043)
Career level before the age of 60	High-level occupations (low)			0.980 (0.053)	0.984 (0.053)
Pension entitlement	Yes (No)			0.947 (0.038)	0.970 (0.034)
Whether there is financial support for children	Yes (No)			0.817*** (0.037)	0.819*** (0.037)
Whether there is adequate medical care in childhood	Yes (No)				0.989 (0.023)
Whether there is a chronic disease	Yes (No)				1.047* (0.025)
Health self-assessment	Well (bad) Normal (bad)				0.712* (0.021) 0.620* (0.017)

Notes: 1. *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level.

2. Reference groups are shown in brackets for variables and standard errors are shown in brackets below the data.

3.3 Heterogeneity test results

As shown in Table 3 the analysis of age heterogeneity found that having a spouse significantly reduced the mortality risk in the oldest old people compared to the middle-aged and younger elderly, meaning that marriage have a stronger protective effect on the oldest old people. For educational attainment, the protective effect of spouses on the mortality risk in old age becomes weaker as educational attainment increases. In terms of gender, men have a higher mortality risk without a spouse compared to women. This is because under traditional Chinese beliefs, men and women play different roles in the marital home, with women taking on more of the burden of care and men gaining more from marriage, and once a man leaves his spouse, it means the beginning of self-care. About urban-rural heterogeneity shows that the mortality risk is higher for elderly people without a spouse in rural areas than in towns, suggesting that elderly people in rural China are more dependent on their spouses than in towns. Spouses are able to reduce the mortality risk for elderly people without financial support from children to a greater extent than for those with financial support from children, this is an important benefit of having a spouse in later life. Having a spouse reduces the mortality risk for elderly people at lower levels to a greater extent than for those whose occupation was at the upper levels before the age of 60. Having a spouse reduces the mortality risk for non-pensioners to a greater extent than for pensioners, but overall difference is not significant, as most of the elderly people in the sample were born in the last century in China and there is little difference in their level of work when they were younger, so the difference in their pensions is not very large.

Table 3

Heterogeneity analysis results
Variables A: age	Lower age	Medium age		High age
Have a spouse (have no spouse)	1.07	0.960		0.732***
Control variables	YES	YES		YES
Sample sizes	773	1934		7714
Chi2	0.34	0.51		98.83***
VariablesB: educational attainment	uneducated	primary school		Junior high school and above
Have a spouse (have no spouse)	0.556***	0.571***		0.580***
Control variables	YES	YES		YES
Sample sizes	6540	2902		979
Chi2	54.96***	170.29***		267.87***
Variables C: gender	Male		Female
Have a spouse (have no spouse)	0.454***		0.579***
Control variables	YES		YES
Sample sizes	4533		5876
Chi2	372.06***		268.55***
Variables D: residence	City		rural
Have a spouse (have no spouse)	0.584***		0.495***
Control variables	YES		YES
Sample sizes	4456		5965
Chi2	301.35***		335.42***
Variables E: inancial support for children	Yes		No
Have a spouse (have no spouse)	0.645***		0.552***
Control variables	YES		YES
Sample sizes	6839		3582
Chi2	185.22***		215.52***
Variables F: Career level before the age of 60	High-level		Low-level
Have a spouse (have no spouse)	0.632***		0.552***
Control variables	YES		YES
Sample sizes	804		9617
Chi2	27.03***		547.76***
Variables G: Pension entitlement	Yes		No
Have a spouse (have no spouse)	0.575***		0.467***
Control variables	YES		YES
Sample sizes	8353		2068
Chi2	102.50***		425.32***

Notes: 1. *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level.

2. Reference groups are shown in brackets for variables and standard errors are shown in brackets below the data.

3.4 Further analysis of results

As shown in Table 4 for those have no spouse, we are divided into widowed, divorced and unmarried so that we could be compared those with have a spouse, and it is found that widowhood brought the highest mortality risk to the old people, followed by divorce and finally unmarried.

Table 4

Further analysis of the results
Numerical values Marital status		Cox risk value	Chi2
Group 1	Have a spouse (Widowed)	0.545***	635.30***
Group 2	Have a spouse (Divorce)	0.721*	2.57*
Group 3	Have a spouse (Unmarried)	0.763**	6.72**

Notes: 1. *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level.

2. Reference groups are shown in brackets for variables and standard errors are shown in brackets below the data.

Widowhood has the highest impact on the mortality risk in elderly people The explanation for the highest impact of widowhood on mortality risk in elderly people relies on two main mechanisms: selection and protection mechanisms(Williams&Umberson, 2004). Those widows who are good in all respects are more likely to remarry, while those who remain widowed are more likely to be in poor health or socio-economically disadvantaged, and it is these that increase their mortality risk. On the other hand, the common lifestyle and life circumstances of couples make the mortality risk of couples more similar, and the death of one partner is followed by the death of the other partner soon afterwards for similar reasons. On the other hand, marriage has a protective effect according to the protective mechanism, which is a protective effect on health through social and economic, and the effect of widowhood on the risk of death stems mainly from the loss of these social and material supports.

About divorce, although it is consensual for people to end a marriage relationship, it is painful for the majority of the elderly, and when a long-standing committed relationship ends abruptly, people often react with anger, anguish and resignation, and this outburst of negative emotions greatly deteriorates people's physical health, leading to a decline in immune function, cardiovascular disease and hormonal disorders, which may also translate into chronic illness and affect the mortality risk Secondly, marriage provides a variety of benefits including companionship, caregiving and emotional support (Peters, 1986), and its dissolution means the loss of a number of important resources and the rapid erosion of all the benefits normally associated with marriage. Compared to the married elderly, the divorced elderly need to spend more energy to adapt to the losses and the old living conditions caused by the divorce, resulting in negative effects on their mental health, which in turn impairs their physical and cognitive health (Eaker ED., 1992), ultimately increases their mortality risk. Thirdly, due to their weakened social relationships, divorced elderly are more likely to feel lonely and depressed, and are more likely to smoke, drink alcohol and abuse drugs for long periods of time, leading to life-threatening conditions (Elwert & Chrietakis, 2008).

For unmarried people, married elderly share a variety of resources with their spouse, who is the primary provider of social support, providing strong support in terms of care needs and emotional needs of older people, all of which form a favourable protection for the health status of older people, while unmarried people have fewer social relationships to build intimate relationships with others, resulting in a less fulfilling life than married people and a sense of well-being in life The lack of social interaction with close people is essential to satisfy our need to belong, and without this satisfaction, a variety of problems can occur and affect our risk of dying.

This paper examines the effect of having a spouse on the risk of death for a specific group of elderly people in China, and finds that the effect of having a spouse on the risk of death for elderly people is significantly lower than the risk of death for elderly people without a spouse by 45.2%, controlling for individual characteristics, socio-economic characteristics and health status; In terms of individual characteristics, the mortality risk for men is 14% higher than that for women, and the mortality risk for middle-aged elderly is 30.1% higher than the younger elderly, and twice as high for older people as for younger people; the higher the level of education, the lower mortality risk; among socio-economic characteristics, financial support from children reduces the mortality risk by 18.3%; in terms of health status, the better self-assessment, the lower mortality risk. In addition, having a spouse significantly reduces the mortality risk in old age, and marriage has a stronger protective effect on the elderly; due to traditional Chinese beliefs, Chinese women bear more of the burden of care in marriage, so the risk of death without a spouse is higher for men; the elderly in rural China are more dependent on their spouse than in urban areas; pensions and occupation before the age of 60 have an effect on the risk of death in old age, but the overall difference is not significant. Further analysis concluded that the mortality risk without a spouse is widowed > divorced > unmarried. Widowhood leads to the highest mortality risk mainly due to selection and protection mechanisms, while divorce causes short-term outbreaks of negative emotions in both partners deteriorating the body and inducing diseases and chronic illnesses affecting themortality risk among the elderly, while intimate relationships are beneficial to our physical health and unmarried people have fewer intimate relationships with others, making their lives less happy.

Based on a sample of elderly in China, we have found that having a partner in later life is a worthwhile life goal, having a partner in old age is effective in reducing the mortality risk in later life and is an important factor in coping with ageing and enjoying a high quality of life in later life. Being divorced, widowed and single is detrimental to health, especially for older, male, rural and insecure older people. In order to reduce the negative impact of spouselessness in later life on the mortality risk in old age, we need to focus on older people, older men, older people in rural areas and older people who are insecure.

Ethics approval and consent to participate

No ethical implications

Consent for publication

Not applicable

Availability of data and materials

The datasets for this study can be found in the CLHLS [https://doi.org/10.18170/DVN/UWS2LR].

Competing interests

The authors declare that they have no competing interests

Funding

Not applicable

Authors' contributions

Shanshan Zhang and Shan Guo conceived this research. Hualei Yang was responsible for the methodology. Shan Guo conducted software analyses. Shanshan Zhang conducted the necessary validations.Hualei Yang conducted a formal analysis and managed the investigation. Shanshan Zhang and Shan Guo gathered resources, curated all the data, wrote/prepared the original draft, and were responsible for project administration. Shanshan Zhang and Shan Guoreviewed and edited the manuscript, were responsible for visualization, supervised the project. All the authors contributed to the article and approved the submitted version.

Acknowledgement

No Acknowledgement

Abellan van Kan, G., Rolland, Y., Berhman, H., Morley, J. E., Kritchevsky, S. B., & Vellas, B. (2008). The I.A.N.A. task force on frailty assessment of older people in clinical practice. Journal of Nutrition and Healthy Aging, 12, 29–37.
Barban N. Family trajectories and health: A life course perspective[J]. European Journal of Population/Revue européenne de Démographie, 2013, 29(4): 357-385.
Berntsen, K. N., & Kravdal, Ø. (2012). The relationship between mortality and time since divorce, widowhood or remarriage in Norway. Social Science & Medicine, 75(12), 2267–2274.
Bergman, H., Ferrucci, L., Guralnik, J., Hogan, D. B., Hummel, S., Karunananthan, S., et al. (2007). Frailty: an emerging research and clinical paradigm—issues and controversies. Journal of Gerontology: Biological Sciences and Medicine Sciences, 62, 731–737.
Brockmann H, Klein T. Love and death in Germany: The marital biography and its effect on mortality[J]. Journal of Marriage and Family, 2004, 66(3): 567-581.
Brockmann, H., Müller, R., & Helmert, U. (2009). Time to retire – Time to die? A prospective cohort study of the effects of early retirement on long-term survival. Social Science & Medicine, 69(2), 160–164.
Cheung, Y. B. (2000). Marital status and mortality in British women: a longitudinal study. International Journal of Epidemiology, 29(1), 93–99.
Choi, H., & Marks, N. F. (2013). Marital Quality, Socioeconomic Status, and Physical Health. Journal of Marriage and Family, 75(4), 903–919.
Donoho, C. J., Crimmins, E. M., & Seeman, T. E. (2013). Marital Quality, Gender, and Markers of Inflammation in the MIDUS Cohort. Journal of Marriage and Family, 75(1), 127–141.
Donrovich R, Drefahl S, Koupil I. Early life conditions, partnership histories, and mortality risk for Swedish men and women born 1915–1929[J]. Social science & medicine, 2014, 108: 60-67.
DYKSTRA, P. A., & KEIZER, R. (2009). The wellbeing of childless men and fathers in mid-life. Ageing and Society, 29(08), 1227–1242.
Green, K. M., Doherty, E. E., Fothergill, K. E., & Ensminger, M. E. (2012). Marriage Trajectories and Health Risk Behaviors Throughout Adulthood Among Urban African Americans. Journal of Family Issues, 33(12), 1595–1618.
Gnambs T, Staufenbiel T. The structure of the General Health Questionnaire (GHQ-12): two meta-analytic factor analyses. Health Psychol Rev. 2018;12(2):179–94.
Guimaraes, R. M. (2007). Health capital, life course and ageing. Gerontology, 53, 96–101.
Hemström Ö. Is marriage dissolution linked to differences in mortality risks for men and women?[J]. Journal of Marriage and the Family, 1996: 366-378.
Hughes M E, Waite L J. Marital biography and health at mid-life[J]. Journal of health and social behavior, 2009, 50(3): 344-358.
Kaplan, R. M. (2006). Marital status and longevity in the United States population. Journal of Epidemiology & Community Health, 60(9), 760–765.
Koball, H. L., Moiduddin, E., Henderson, J., Goesling, B., & Besculides, M. (2010). What Do We Know About the Link Between Marriage and Health? Journal of Family Issues, 31(8), 1019–1040.
Lawrence E M, Rogers R G, Zajacova A, et al. Marital happiness, marital status, health, and longevity[J]. Journal of Happiness Studies, 2019, 20(5): 1539-1561.
Lerchl A. Month of birth and life expectancy: role of gender and age in a comparative approach[J]. Naturwissenschaften, 2004, 91(9): 422-425.
Liu H, Waite L. Bad marriage, broken heart? Age and gender differences in the link between marital quality and cardiovascular risks among older adults[J]. Journal of health and social behavior, 2014, 55(4): 403-423.
Li, M. (2008). Investigation on China’s increasing divorce rate. Government Legality, 14, 2–5 [in Chinese].
Metsä-Simola, N., & Martikainen, P. (2013). The short-term and long-term effects of divorce on mortality risk in a large Finnish cohort, 1990–2003. Population Studies, 67(1), 97–110.
Paul Downward, Simona Rasciute, Harish Kumar, Mental health and satisfaction with partners: a longitudinal analysis in the UK, BMC Psychology, 10.1186/s40359-022-00723-w, 10, 1, (2022).
Perkins J M, Lee H, James K S, et al. Marital status, widowhood duration, gender and health outcomes: a cross-sectional study among older adults in India[J]. BMC public health, 2016, 16(1): 1-12.[21] Richmond P, Roehner B M. Impact of marital status on health[J]. Physica A: Statistical Mechanics and Its Applications, 2017, 486: 542-553.
Robards, J., Evandrou, M., Falkingham, J., & Vlachantoni, A. (2012). Marital status, health and mortality. Maturitas, 73(4), 295–299.
Ruggeri K, Garcia-Garzon E, Maguire Á, Matz S, Huppert FA. Well-being is more than happiness and life satisfaction: a multidimensional analysis of 21 countries. Health Qual Life Outcomes. 2020;18(1):1–16.
Sbarra D A, Hasselmo K, Nojopranoto W. Divorce and death: A case study for health psychology[J]. Social and personality psychology compass, 2012, 6(12): 905-919.
Tatangelo G, McCabe M, Campbell S, et al. Gender, marital status and longevity[J]. Maturitas, 2017, 100: 64-69.
Va, P., Yang, W.-S., Nechuta, S., Chow, W.-H., Cai, H., Yang, G., … Xiang, Y.-B. (2011). Marital Status and Mortality among Middle Age and Elderly Men and Women in Urban Shanghai. PLoS ONE, 6(11), e26600.
Zhang Z, Hayward M D. Gender, the marital life course, and cardiovascular disease in late midlife[J]. Journal of Marriage and Family, 2006, 68(3): 639-657.
Zhu, H., & Xie, Y. (2007). Socioeconomic differentials in mortality among the oldest old in China. Research on Aging, 29(2), 125–143.
Zimmer, Z. (2005). Health and living arrangement transitions among China’s oldest-old. Research on Aging, 27(5), 526–555.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The effect of having a spouse on the mortality risk in later life --Empirical evidence from China

Status:

Version 1

Abstract

Purpose

Method

Results

Conclusion

Figures

1 Introduction

2 Model And Method

2.1 Data sources and samples

2.2 Variables

2.2.1 Dependent variable: death risk

2.2.2 Independent variable: marital status

2.2.3 Control Variables

2.3 Model Setting

3 Rseults

3.1 cumulative hazard function on spouse status

3.2 cox proportional-hazards model regression results

3.3 Heterogeneity test results

3.4 Further analysis of results

4 Conclusion

Declarations

References

Additional Declarations

Status:

Version 1