Marriage has been identified as an important factor associated with health and mortality. Married people generally enjoy better health (Duncan, Wilkerson, & England, 2006; Wu & Hart, 2002) and lower mortality (Boyle, Feng, & Raab, 2011; Ikeda et al., 2007; Va et al., 2011) than their nonmarried counterparts. Widowed people typically experience the effects of bereavement, with spousal death affecting health and mortality, particularly in older people. It is well known that the effects of chronic disease and deterioration of physical function increase with age. The older people have a poorer health status and are more likely to die than relatively younger people. Thus, what are the impacts of widowhood, chronic disease, and function limitation among the older people? How great is this impact? Population aging and the accompanying increase in the number of widowed people in Taiwan are important and timely issues to study.
This study follows three lines of inquiry in the literature that developed in relative isolation from each other. First, Wu, Lan, Chen, Chiu, and Lan (2011) pointed out that older people impose a huge health cost on society. They noted survival rates correlated negatively with chronic disease and different physical function domains. Of these, current guidelines identify older people with chronic kidney disease at a heightened risk of cardiovascular and all-cause mortality in the United States (Tonelli et al., 2006). Wang, Mi, Shan, Wang, and Ge (2007) highlighted chronic disease risks and noted that number of total deaths increased from 128 to 145 and that such risks contributed to the rise in China’s mortality rate from 27% to 32% between 1990 and 2003. In addition, some studies addressing various disease situations found a 1.5-fold increase in gout mortality risk (Cohen, Kimmel, Neff, Agodoa, & Abbott, 2008), a 2.9- to 3.3-fold increase in coronary heart disease mortality risk (Lakka et al., 2002), and a 1.34-fold increase in ischemic stroke mortality risk (The Emerging Risk Factors Collaboration, 2010). Finally, Strong, Mathers, Leeder, and Beaglehole (2005) presented the mortality and burden of disease projections for chronic diseases. Their goal was to reduce chronic disease death rates by an additional 2% annually, which would, if achieved, extend the lives of some 36 million people by 2015.
Deterioration in physical function is common in elderly people. Physical function disabilities refer to conditions that make an elderly person unable to care for himself or herself and participate in activities of community and daily living (Lin, Hwang, Liu, & Lin, 2012). Yeh et al. (2012) assessed 2,729 community-dwelling persons older than 50 years old with no initial disability and found that incident disability (e.g., mild level: running, carrying weight, and squatting; moderate level: climbing stairs, walking, and standing; severe level: grasping and raising arms up) occurred after 10 years. They further indicated that women were at greater risk of developing more severe disabilities. Considering various indicators of physical function, Cooper, Kuh, and Hardy (2010) found that the hazard ratio for mortality comparing the weakest with the strongest grip strength quartile is 1.67 and comparing the slowest with the fastest walking speed quartile is 2.87. Therefore, physical function disability was not only a predictive index of health status but also had a direct impact on elderly survival rates (Ostir, Markides, Black, & Goodwin, 1998). In a similar spirit, we examined whether the impacts of chronic disease and physical function on mortality were stronger or weaker among the older people in Taiwan.
In Taiwan, the results from studies of mortality have identified several associated factors. Those who were older, were men, had a lower level of education, engaged in fewer social activities, and had cardiovascular symptoms faced a higher mortality risk (Chen et al., 2007; Ho, Li, & Liu, 2009; Liu, Tsou, & Hammitt, 2007; Zimmer, Martin, & Lin, 2005). In addition, although many studies have improved our understanding of the relationship between chronic disease, physical function, and mortality, they typically focus on a specific category of living facility or living arrangement (Chang, & Chueh, 2011; Lin, Chang Yeh, Chen, & Hung, 2010; Lin et al., 2012). Little attention has been paid to widowhood, particularly in the older people. To address this gap, this article examined mortality risk for widowed older persons. We controlled socioeconomic characteristics and investigated whether widowhood, chronic disease, and physical function were significantly associated with elderly survival rates. Moreover, although it is widely accepted that women as a group have a longer life expectancy than men, we examined whether this relative advantage holds true even after spousal death.
Data were derived from the Survey of Health and Living Status (SHLS) of the Middle Aged and Elderly in Taiwan, a joint survey conducted by the Bureau of Health Promotion, Department of Health (DOH), and the Population Studies Center at the University of Michigan. Data were generated by a random survey conducted in 331 cities, townships, and rural areas (Taiwan’s relatively remote mountain districts were not included in this survey). These data included three panels and six waves of data between 1989 and 2007 (1989, 1993, 1996, 1999, 2003, and 2007). In its analysis of mortality among the older people, this study only focused on longitudinal data collected between 2003 and 2007.
The SHLS panel used in this study took place between 2003 and 2007, with response rates for the two years about 91.5% and 91.1%, respectively. Initial interviews were held in 2003 for married groups1 with 3,768 effective respondents aged 57 years and older. This pool of effective respondents, reduced by participant deaths to 3,098, was then followed up in 2007. Therefore, 670 effective respondents had died between 2003 and 2007. This article studied the survival of the 3,768 respondents in 2003 and analyzed the relative determinants of mortality between the two married groups.
Samples included in the SHLS data were segregated into two groups, that is, those still alive and those who had died during the observation period. The former, comprising those who did not die during the sample period (January 2003–December 2007), was designated as “right-censored” spells. The latter, comprising those who had died during the sample period, was known as “uncensored” spells. For uncensored spells, the study duration began with the dates when individuals were first interviewed and ended with their respective dates of death. Conversely, right-censored spells were alive throughout the study period. This variable can be categorized as dependent, with the uncensored variable coded 1 for deceased and 0 for still alive.
Main independent variables
This study was designed to explore concurrently the relationships between mortality hazard and widowhood, chronic disease, and physical function. Hence, variables included widowhood, chronic diseases, and physical function based on data from 2003. First, for widowhood from SHLS data, original marital statuses were categorized into five groups: married with a currently living spouse and married with a spouse who had died, divorced, separated, or unmarried. This study considers the former two groups only and defines participants whose spouses had died as in “widowhood.” In addition, because of gender difference in health behavior and chronic diseases (Bauer, Göhlmann, & Sinning, 2007; Lin et al., 2010), this article further divided widowhood groups by gender, namely widowers (men) and widows (women).
Because chronic diseases and physical function disabilities directly affect survival rates, particularly in the older people, this study considered these two variables. On the basis of categories used in previous studies (Chang & Chueh, 2011; Ho et al., 2009; Lin et al., 2012), chronic diseases identified in this study included hypertension, diabetes, heart disease, stroke, tumor, bronchitis, pneumonia, and other respiratory diseases. A “1” was assigned to respondents reporting one or more of these items and a “0” was assigned to those reporting none. This study defined physical function as the ability to carry out daily activities and tasks such as those included in the list of activities of daily living (ADLs; e.g., bathing; dressing/undressing; eating; getting out of bed, standing up, sitting in a chair; moving about the house; using the lavatory) and those included on the index of instrumental ADLs (IADLs; e.g., buying personal use items, managing money, riding the bus or train independently, doing physical work at home or around the house, sweeping, washing dishes, taking out the garbage; other light tasks). All index variables were deemed necessary for maintaining a comfortable living environment. For each activity, a value of 0 indicated “no problem performing,” 1 indicated “some difficulty,” 2 indicated “significant difficulty,” and 3 indicated “unable to perform.” Scores for all were summed and then divided and converted to multicategorical variables (no problem performing, score of 0; some difficulty, score of 1∼6; significant difficulty, score of 7∼12; unable to perform, score of 13∼18) to describe physical functions2. This study hypothesized that those subjects with ailments or functional limitations should experience a higher mortality hazard than experienced by those without such ailments or other limitations (Chen et al., 2007; Lin et al., 2012).
Because different socioeconomic backgrounds impact mortality differently, this article controlled for certain socioeconomic factors to better elicit the influences of widowhood, chronic diseases, and physical function on mortality. Socioeconomic characteristics controlled for included age, ethnic group, education, and 2003 household income. Variable details are described below.
Following the categorizing protocols suggested by Zimmer et al. (2005) and Chen et al. (2007), the variables of age and ethnicity placed subjects into four different groups, namely Age 1 (57–66 years), Age 2 (67–76 years), Age 3 (77–86 years), and Age 4 (87 years and over) and Hokkien, Hakka, Mainlander, and other. Age 1 (aged 57–66 years) and Hokkien, respectively, were used as reference variables. On the basis of the results of previous studies, we hypothesized that relatively older persons and those who were Hakka would be more likely to die than relatively younger persons and those who were Hokkien (Buckley, Denton, Robb, & Spencer, 2004; Chen et al., 2007; Lin & Lin, 2006; Zimmer et al., 2005).
Following Zimmer, Martin, and Chang (2002) and Zimmer et al. (2005), we classified educational level into the three groups of no formal education (0 years), primary education (1–6 years), and higher education (greater than 6 years). On the basis of previous research (Chen et al., 2007; Lin & Lin, 2006; Zimmer et al., 2005), we hypothesized that older persons with higher educational levels would be less likely to die than older persons with lower educational levels.
Only 1,895 (50.3%) of participants in the 2003 survey data provided data on household income. Therefore, this article followed Buckley et al. (2004) and performed an ordinary least squares regression relationship between participant household income as a dependent variable and the individual’s age and other significant characteristics as independent variables. The resulting relationship was used to estimate household income of 3,768 participants3. We then used the following process to set a standardized relative-income quartile. A set of dummy variables squares was defined, Rij (j = 1, 2, 3, 4), to represent the quartile group in which an individual was located based on his or her relative household income. Reliability was quite high (Cronbach’s alpha = .90). Finally, based on the results of Buckley et al. (2004), Chen et al. (2007), and Zimmer et al. (2005), we hypothesized that older married persons with higher household incomes should face a lower mortality risk than older married persons with lower household incomes.
Proportional Hazard Model
The Cox proportional hazard model was used to estimate mortality risk as a way to examine determinants of mortality for participants. Following Cox (1972), we compared mortality hazard in two or more groups with duration exposure. Individuals faced certain additional characteristics that might affect mortality hazard. In particular, the Cox proportional hazard model did not need to make any assumptions about the shape of the baseline hazard function4. Therefore, the proportional hazard function could be defined as
where hi (t; xi) represented the mortality hazard at exact time t, h 0 (t) was the baseline hazard function depending on t and not xi, β’ was the parameter vector (the prime mark [’] denotes transposition), and xi was the covariate vector (e.g., widowers, widows, chronic diseases, physical function, and socioeconomic factors).
Figures 1 and 2 illustrate that the SHLS sample included 3,768 married elderly participants (including 1,941 men [315 widowers and 1,626 nonwidowers] and 1,827 women [863 widows and 964 nonwidows]) and survival rate data from 2003 to 2007 (60 months). Survival rates for the widowed were lower than for those whose spouses were still alive regardless of gender and time period. Widower survival rates decreased significantly from 1 at the outset to 0.85 at the 30th month and to 0.64 at the 60th month. Nonwidower survival rates decreased slightly from 1 to 0.92 and 0.82 during the same observed periods. Survival rates for married female widows and nonwidows decreased from 1 to 0.90 and 0.78 and from 1 to 0.95 and 0.90 during the same observed periods. The gap in survival rates for widowers and nonwidowers was larger than for widows and nonwidows.
Table 1 further summarizes descriptive statistics for the initial (2003) data collection period. For married men, mean survival duration was 55 months, with 392 elderly men dying between 2003 and 2007. In the first study data set (2003), the proportion of widowers among married men was 16.2%. Mean ages were 77.91 years (SD = 7.63 years) for widowers and 75.65 years (SD = 7.88 years) for widows. Less than 20% of men and over half of women reported informal education as their highest level of education. Estimated household income was NT$952,000 and NT$794,000 per year for male and female participants, respectively. About 40% of participants (37.4% for men, n = 726; 42.3% for women, n = 772) reported hypertension problems, heart disease was more prevalent than diabetes among both genders, fewer than 10% of married men (8.2%, n = 159) and women (7.6%, n = 139) had previously experienced stroke, and 3% of all participants reported having had a tumor disease. In terms of physical function disabilities, a majority of men reported no difficulties in ADL task performance, whereas over half of married women (56.2%, n = 1,026) reported IADL difficulties.
This study further subdivided the widowhood group by gender to examine mortality risk for widowers and widows. As shown in Table 2, when socioeconomic factors were controlled, the hazard ratio of married male widowers was 1.308. This meant that, after a wife died, the husband had a 30.8% greater risk of dying compared with those men whose spouses were still alive. In addition, hazard ratios increased with age (p < .000) and decreased with education (p < .05). This reflected the fact that older people and those with lower levels of education faced a higher mortality hazard and were more likely to die than those relatively younger in age and with better education. Finally, the hazard ratio for Mainlanders was less than 1 and significant (p < .05). This meant that Mainlander participants had a lower mortality hazard and were less likely to die than their Hokkien counterparts.
In terms of chronic disease, the results found most chronic diseases to be significant factors (with the exception of hypertension, diabetes, and heart disease), suggesting that, in most cases, there were statistically significant associations between chronic disease and mortality (p < .05). In addition, similar significant and positive effects on mortality were found for ADL and IADL variables (p < .05, with the exception of some ADL performance difficulties). This evidence implied that ailments and ADL/IADL variables significantly affected the mortality risk of elderly male participants.
Turning to the married women, the results were similar to those for men. Widows also faced a higher mortality hazard and were more likely to die than those whose spouses were still alive. In addition, most results of socioeconomic factors were similar to men with the exception of ethnic and income variables. In terms of ethnicity, being a Mainlander showed significance for men but not for women. Income (fourth) indicated significance for elderly women but was insignificant for elderly men. In terms of chronic disease, the results indicated stroke and tumor to be significant factors, suggesting strong associations with mortality in married women. In addition, ADL and IADL also showed similar results to those of the elderly men.
To compare different rates of death hazard probability between widowers and widows as categorized by sociodemographic factors, chronic diseases, and physical functions, this study used four different models to describe mortality hazard. The initial model (Model 1) controlled for widowhood peers and socioeconomic variables; Models 2 and 3 described the singular influence of chronic disease and physical function, respectively, on mortality; and Model 4 considered the combined influences of socioeconomic factors, chronic diseases, and physical function on mortality. Test results illustrated in the last rows of Table 3 show significance (p = <.001) for all four models. This suggests a firm rejection of the null hypothesis and that the Cox proportional hazards model is a fitted object. The log likelihood of Model 4 was the smallest (−1868.047) of the four, suggesting that this model may be most appropriate to describe widowhood mortality hazard.
For the widowhood group, Table 3 shows the hazard ratio for widowers to be greater than 1 and significant among all four models. This means that widowers in the study faced a higher mortality risk and were more likely to die than their widow peers. We further observed the hazard ratios of widowers among these four models. The hazard ratios began at 1.818 and rose to 2.695. This finding provides suggestive evidence that widowhood was strongly correlated with observed factors (sociodemographic factors, chronic disease, and physical function) among participants.
The main finding of this study was that widowhood had significant and positive effects on mortality. As expected, widowers and widows faced greater mortality risks and were more likely to die than those whose spouses were still alive. As previous research noted, marriage discouraged risky behavior (e.g., smoking, heavy alcohol use, illicit drug use) and encouraged healthy behavior (e.g., exercise, visiting the doctor; Duncan et al., 2006). Moreover, family functions were very important, particularly in Asian countries. Most Asian countries lack a comprehensive social welfare system, which encourages people to depend on their family (e.g., spouse, children) in later life. Likewise, marriage in Taiwan is greatly valued and viewed as an essential life event for men and women. Those who have lost their spouses may experience greater social isolation than their nonwidowed peers. Hence, elderly individuals who had experienced the death of their spouses would show a higher mortality risk than those who had not.
In terms of chronic diseases, tumor and stroke were identified as having a significant influence on mortality for both married men and women. However, bronchitis, pneumonia, and other respiratory diseases significantly affected mortality risk only for married men. This is consistent with Taiwan government statistics (DOH, Executive Yuan, Taiwan, ROC, 2012) that show tumors as the leading causes of death in recent years in Taiwan. The tumor factor is an independent prognostic marker for mortality, so tumor necrosis factor has specific biological effects and is a marker of frailty in the very old people (Bruunsgaard, Andersen-Ranberg, Hjelmborg, Pedersen, & Jeune, 2003). In addition, this study further found that stroke also menaced survival rates significantly. Stroke was the second leading cause of death in the United States in the early 2000s and remains a major and growing health problem as the population ages (Ingall, 2004). Therefore, we should pay more attention to the impacts of tumor and stroke on mortality risk in the older people.
Another notable finding was that ADL and IADL had significant and positive effects on mortality for both married men and women. Our results provided evidence that ADL and IADL may provide relatively reliable indicators of mortality (Lin & Lin, 2006; Malmstrom et al., 2007; Zimmer et al., 2005). Specific medical conditions may be asymptomatic or subject to differential diagnosis and thus subject to considerable measurement error. ADL and IADL were objective indicators. Respondents were clear about their health status when it was affected by functional limitations. Therefore, it may be that ADL and IADL are better barometers and more sensitive to qualitative, preclinical changes, thus providing a better reflection of mortality hazard. Personal physical functions, particularly IADL, are of primary concern to the older people.
Although physical function showed a significant effect for married men and women, this article echoed the opinion of Yeh et al. (2012) that women were at greater risk of physical function disability than men. Reasons for this may be, in large part, because of greater pessimism and East Asian social mores. Women often assume additional responsibilities for household management and financial matters, which are considered rewarding in Taiwan and other Chinese social settings (Chan, Brownridge, Tiwari, Fong, & Leung, 2008; Li, 2005). Most of women consider their families the center of their lives both in the present and the future. They invest their entire selves in and find self-worth through their housekeeping role (Shu, Chuang, Lin, & Liu, 2008). Once they begin experiencing physical function disabilities, they believe themselves less able to contribute to the family. Therefore, they are likely at higher risk of depression and face a higher mortality hazard ratio. In contrast, men are generally more optimistic, which would justify the relatively weak influence of physical function on mortality for our study’s male participants.
Finally, in Table 3, the estimated hazard ratio of mortality for widowers in all four models was greater than 1 and significantly higher than that for widows, indicating that widowers faced a higher mortality risk than widows. Thus, elderly men were more likely to die than elderly women, even after spousal death. This finding was consistent with previous studies (Buckley et al., 2004; Chen et al., 2007; Lin & Lin, 2006; Zimmer et al., 2005) and Taiwan government statistics (Ministry of the Interior, Department of Statistics, Taiwan, ROC, 2012). Many elderly men rely heavily on their spouses for activities central to maintaining health such as scheduling doctor visits, reminding to take medication, exercise, and cooking (Rendall, Weden, Favreault, & Waldron, 2011). Under such conditions, married men can be expected to suffer more from the loss of spouses than married women. Likewise, women typically have broader social networks than men, which can help alleviate the physical and mental stresses after the death of a husband (Espinosa & Evans, 2008). Lee, DeMaris, Bavin, and Sullivan (2001) mentioned that the impact of widowhood on the incidence of depression was greater for men than women. Many widowers adopt strategies such as remarriage to combat the increased risks to health after spousal death. Elderly widowers have been reported as more likely to remarry than elderly widows (Smith, Zick, & Duncan, 1991).
Survival analysis indicated a relatively higher mortality risk among widowed participants in comparison with their nonwidowed peers as well as a higher mortality risk for widowers than widows. We further found that this relative advantage held true even after spousal death. This facilitated the longer life expectancy of women over men in this study. Physical function was found to be a significant predictive factor related to mortality hazard for both married elderly men and married elderly women. Therefore, survival analysis of older people should take this factor into account.
Strengths and Limitations
This longitudinal study allowed us to capture a large amount of quantifiable data for empirical analysis. Limitations of this study include the use of secondary data and our reliance on self-reported questionnaires. Selection bias is a further limitation because those in divorced, separated, and never-married categories were not considered in data analysis.
The authors would like to thank the Bureau of Health Promotion at the DOH for providing the SHLS data of the middle aged and elderly in Taiwan. Thanks also to the National Science Council of Taiwan for the fellowship grant (NSC-100-2410-H-275-008) that made this research collaboration possible.
Bauer T., Göhlmann S., Sinning M. (2007). Gender differences in smoking behavior. Health Economics, 16 (9), 895–909. doi:10.1002/hec.1259
Boyle P. J., Feng Z. Q., Raab G. M. (2011). Does widowhood increase mortality risk? Testing for selection effects by comparing causes of spousal death. Epidemiology, 22 (1), 1–5. doi:10.1097/EDE.0b013e3181fdcc0b
Bruunsgaard H., Andersen-Ranberg K., Hjelmborg J. B., Pedersen B. K., Jeune B. (2003). Elevated levels of tumor necrosis factor alpha and mortality in centenarians. The American Journal of Medicine, 115 (4), 278–283. doi:10.1016/S0002-9343(03)00329-2
Buckley N. J., Denton F. T., Robb A. L., Spencer B. G. (2004). The transition from good to poor health: An econometric study of the older population. Journal of Health Economics, 23 (5), 1013–1034. doi:10.1016/j.jhealeco.2004.03.001
Chan K. L., Brownridge D. A., Tiwari A., Fong D. Y. T., Leung W. C. (2008). Understanding violence against Chinese women in Hong Kong: An analysis of risk factors with a special emphasis on the role of in-law conflict. Hong Kong, China: University of Hong Kong. doi:10.1177/1077801208325088
Chang T. Y., Chueh K. H. (2011). Relationship between elderly depression and health status in male veterans. The Journal of Nursing Research, 19 (4), 298–303. doi:10.1097/JNR.0b013e318236cf89
Chen L., Yip W., Chang M. C., Lin H. S., Lee S. D., Chiu Y. L., Lin Y. H. (2007). The effects of Taiwan’s National Health Insurance on access and health status of the elderly. Health Economics, 16 (3), 223–242. doi:10.1002/hec.1160
Cohen S. D., Kimmel P. L., Neff R., Agodoa L., Abbott K. C. (2008). Association of incident gout and mortality in dialysis patients. Journal of the American Society of Nephrology, 19 (1), 2204–2210. doi:10.1681/ASN.2007111256
Cooper R., Kuh D., Hardy R. (2010). Objectively measured physical capability levels and mortality: Systematic review and meta-analysis. British Medical Journal, 341, c4667. doi:10.1136/bmj.c4467
Cox D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society, Series B (Methodological), 34 (2), 187–220.
Duncan G. J., Wilkerson B., England P. (2006). Cleaning up their act: The effects of marriage and cohabitation on licit and illicit drug use. Demography, 43 (4), 691–710. doi:10.1353/dem.2006.0032
Espinosa J., Evans W. N. (2008). Heightened mortality after the death of a spouse: Marriage protection or marriage selection? Journal of health Economics, 27 (5), 1326–1342. doi:org/10.1016/j.jhealeco.2008.04.001
Ho S. H., Li C. S., Liu C. C. (2009). The influence of chronic disease
, physical function
and lifestyle on health transition among the middle-aged and older persons in Taiwan. The Journal of Nursing Research, 17 (2), 136–143. doi:10.1097/JNR.0b013e3181a53f94
Ikeda A., Iso H., Toyoshima H., Fujino Y., Mizoue T., Yoshimura T., Tamakosi A. (2007). Marital status and mortality among Japanese men and women: The Japan collaborative cohort study. BioMed Central (BMC) Public Health, 7 (73), 1–7. doi:10.1186/1471-2458-7-73
Ingall T. (2004). Stroke—Incidence, mortality, morbidity and risk. Journal of Insurance Medicine, 36, 143–152.
Lakka H.-M., Laaksonen D. E., Lakka T. A., Niskanen L. K., Kumpusalo E., Tuomilehto J., Salonen J. T. (2002). The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men. The Journal of the American Medical Association, 288 (21), 2709–2716. doi:10.1001/jama.288.21.2709
Lee G. R., DeMaris A., Bavin S., Sullivan R. (2001). Gender difference in the depressive effect of widowhood in later life. The Journal of Gerontology: Series B. Psychological Sciences and Social Sciences, 56 (1), S56–S61. doi:10.1093/geronb/56.1.S56
Li X. H. (2005). Empirical research on the marriage and family rights and interests protection of laid-off workers. Journal of China Women’s University, 17 (6), 10–15. (Original work published in Chinese)
Lin C. H., Lin H. S. (2006). Factors associated with survival status of older Taiwanese adults. Taiwan Journal of Public Health, 25 (5), 351–362. (Original work published in Chinese)
Lin S. J., Hwang S. J., Liu C. Y., Lin H. R. (2012). The relationship between nutritional status and physical function
, admission frequency, length of hospital stay, and mortality in old people living in long-term care facilities. The Journal of Nursing Research, 20 (2), 110–121. doi:10.1097/jnr.0b013e318254eac9
Lin Y. C., Chang Yeh M., Chen Y. M., Huang L. H. (2010). Physical activity status and gender differences in community-dwelling older adults with chronic diseases. The Journal of Nursing Research, 18 (2), 88–97. doi:10.1097/JNR.0b013e3181dda6d8
Liu J. T., Tsou M. W., Hammitt J. K. (2007). Health information and subjective survival probability: Evidence from Taiwan. Journal of Risk Research, 10 (2), 149–175. doi:10.1080/13669870701191802
Malmstrom T. K., Andresen E. M., Wolinsky F. D., Miller J. P., Stamps K., Miller D. K. (2007). Mortality risk in older inner-city African Americans. Journal of the American Geriatrics Society, 55 (7), 1049–1055. doi:10.1111/j.1532-5415.2007.01204.x
Ostir G. V., Markides K. S, Black S. A., Goodwin J. S. (1998). Lower body functioning as a predictor of subsequent disability among older Mexican Americans. The Journal of gerontology: Series A, Biological Science and Medical Science, 53A (6), M491–M495. doi:10.1093/gerona/53A.6.M491
Rendall M. S., Weden M. M., Favreault M. M., Waldron H. (2011). The protective effect of marriage for survival: A review and update. Demography, 48 (2), 481–506. doi:10.1007/s13524-011-0032-5
Shu B. C., Chuang L. Y., Lin L. L., Liu K. L. (2008). The mothering experience: Perspectives of women from Indonesia married and living in Taiwan. The Journal of NursingResearch, 16 (3), 169–176. doi:10.1097/01.JNR.0000387303.11870.fe
Smith K. R., Zick C. D., Duncan G. J. (1991). Remarriage patterns among recent widows and widowers. Demography, 28 (3), 361–374. doi:10.2307/2061462
Strong K., Mathers C., Leeder S., Beaglehole R. (2005). Preventing chronic diseases: How many lives can we save? The Lancet, 366 (9496), 1578–1582. doi:10.1016/S0140-6736(05)67341-2
The Emerging Risk Factors Collaboration. (2010). C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: An individual participant meta-analysis. The Lancet, 375 (9709), 132–140. doi:10.1016/S0140-6736(09)61717-7
Tonelli M., Wiebe N., Culleton B., House A., Rabbat C., Fok M., Garg A. X. (2006). Chronic kidney disease and mortality risk: A systematic review. Journal of the American Society of Nephrology, 17 (7), 2034–2047. doi:10.1681/ASN.2005101085
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), 1–10. doi:10.1371/journal.pone.0026600
Wang Y., Mi J., Shan X. Y., Wang Q. J., Ge K.Y. (2007). Is China facing an obesity epidemic and the consequences? The trends in obesity and chronic disease
in China. International Journal of Obesity, 31, 177–188. doi:10.1038/sj.ijo.0803354
Wu M. S., Lan T. H., Chen C. M., Chiu H. C., Lan T. Y. (2011). Socio-demographic and health-related factors associated with cognitive impairment in the elderly in Taiwan. BMC Public Health, 11 (22), 1–8. doi:10.1186/1471-2458-11-22
Wu Z., Hart R. (2002). The effects of martial and nonmarital union transition on health. Journal of Marriage and Family, 64 (2), 420–432.
Yeh C. J., Wang C. Y., Tang P. F., Lee M. C., Lin H. S., Chen H. Y. (2012). Hierarchy of higher-level physical functions: A longitudinal investigation on a nationally representative population of community-dwelling middle-aged and elderly persons. Disability Rehabilitation, 34 (15), 1271–1276. doi:10.3109/09638288.2011.641657
Zimmer Z., Martin L. G., Chang M. C. (2002). Changes in functional limitation and survival among older Taiwanese, 1993, 1996, and 1999. Population Studies, 56 (3), 265–276. doi:10.1080/00324720215931
Zimmer Z., Martin L. G., Lin H. S. (2005). Determinants of old-age mortality in Taiwan. Social Science & Medicine, 60 (3), 457–470. doi:10.1016/j.socscimed.2004.06.006
1The two married groups included participants with currently living spouses and participants whose spouses had died before 2003.
2Most SHLS data were self-reported. The measurement of ADL and IADL and multicategorical variables came directly from the SHLS data.
3The significant variables were age and education in this regression.
4Since only the baseline function is specified parametrically.
Keywords:Copyright © 2013 by the Taiwan Nurses Association.
widower; widow; chronic disease; physical function