Participants were recruited using a stratified random-sampling scheme. To be specific, we first obtained the most up-to-date list of residents in all of the 11 districts in Kaohsiung City, Taiwan, after the study protocol had been approved by the Ethics Committee of the University. A total of 106,697 residents aged 65 or above were identified. We then randomly chose and scheduled a face-to-face interview for 6,367 senior citizens from this pool, with the number chosen in each district proportional to the population size of that district.

Of the 6,367 scheduled interviews, 860 were not conducted: 42 residents died prior to the interview, 742 were unable to be located, 76 declined an interview. Although 5,507 interviews were completed by October 2000, 82 interviews were held with proxies (family member or close acquaintance); 46 candidates had mental illness; 4 were newly diagnosed with dementia during the study; 386 had over half of the data missing; 131 had the key variable, that is, living alone, missing. They were therefore excluded from the final analysis. As a result, data analysis was carried out on 4,858 people.

Sociodemographic variables collected included age, gender, marital status, religion, occupation, education level, and main source of income.

Living alone. A binary variable measured by asking the respondents whether they were currently living alone or not.

Social support. Perceived positive support from friends, that is, whether they had at least one good/close friend to talk to.

Loneliness. Sense of loneliness was a subjective measure on a 3-point Likert scale: strong, some, and little.

Chronic conditions. Participants were asked whether they had to visit a hospital at least once during the last 12 months due to stroke, hypertension, diabetes, cardiac heart disease, cancer, or Parkinson’s disease; or regularly for vision, hearing, or teeth problems.

Functional status. Functional status is the ability to perform ADL and instrumental activities of daily living (IADL) that reflect a person’s capacity to live independently. The Katz ADL index (1976) that addresses bathing, dressing, toileting, transferring, continence, and feeding was used. The instrument measures ADL of the respondents on a three-point scale: independent, semi-independent (needs a part-time assistant) or dependent. On the other hand, respondents were asked to make a self-evaluation of whether or not they could perform each IADL on their own: preparing meals, shopping, taking medicine, traveling out of walking distance, managing money, and using the telephone. Internal consistency for ADL and IADL, measured using Cronbach’s alpha, in our sample was α = 0.94 and α = 0.90, respectively.

Short Portable Mental Status Questionnaire (SPMSQ). The instrument uses a 10-item scale to measure orientation, personal history, remote memory and calculation ability. One point is given for each correct answer. A summary score is constructed by summing up the 10 items, yielding a range from 0 to 10 (Pfeiffer, 1975). Internal consistency of the SPMSQ scale for this study was 0.98.

A comparison between those living alone and those living with someone are presented in Table 1. Those of the male gender and not married (including never married, divorced, separated, and widowed) were more likely to be living alone. In addition, source of income, occupation, and religion were significantly associated with living alone or not. The living alone group had a higher percentage with no income or receiving government assistance, without a job or retired, and more of them were Christians. Those who were living alone also had lower ADL and IADL scores, meaning that they were less dependent on assistants or instruments to perform their daily life. Moreover, they were, on average, 1.5 years older than elderly living with someone. However, there were no differences in baseline chronic conditions and SPMSQ score between the two groups; nor were the two groups significantly different in any of the health problems.

The adjusted odds ratios derived from the multivariate analysis are presented in Table 2. All the significant factors shown in Table 1, except age and ADL, remained significant in the logistic regression. The nonsignificance of ADL was mainly attributable to the fact that it was highly correlated with IADL (r = 0.79, p < 0.001). Age became nonsignificant because the difference between the two groups was very small (mean difference of 1.5 years), and it was significant in the univariate analysis purely because of the large sample size. Interestingly, we found significant interaction between gender and marital status. Not only were males 1.9 times – and those who were not married 19.6 times – more likely to be living alone, but also males who did not have a partner at the time of the interview were at an even higher risk of living alone.

The degree of feeling lonely and whether the participants had close friends to talk to were compared. As shown in Table 3, the living alone group had lower perceived positive supports, and more of them had a strong sense of loneliness.

The effect of living alone on perceived positive supports was still significant, by fitting another logistic regression with perceived support as the dependent variable and living alone an independent variable, after adjustment for all significant covariates identified in Table 2. Those living alone were less likely (odds ratio = 0.6, 95% CI = 0.5 to 0.8) to have close friends to talk to. Similarly, using a multinomial logistic regression, we found that the living alone group, after removing the effects of the significant covariates, was more likely to have a strong sense of loneliness than having little feeling of loneliness (OR = 3.6, 95% CI = 2.4 to 5.3) or some feeling of loneliness (OR = 2.5, 95% CI = 1.7 to 3.7).

Of the 4,858 seniors interviewed, 3.3% were over 85 years old, 33% were illiterate, and 42% received the major part of their income from their children. These percentages are representative of the overall elderly population in Kaohsiung. Note also that the distributions of gender, age, marital status, and education level of senior citizens living in Kaohsiung are similar to those of the entire Taiwan elderly population (Wu & Chang, 1997). Hence, our study sample is quite representative of the Taiwanese population who are aged 65 or above.

The concept of social support is difficult to measure meaningfully, and can be a cause or effect. Proxy indicators (e.g., occupational status) have been used to make inferences about the effect of social activities on health care problems (Fabrigoule et al., 1995). We have used marital status as a measure of support for predicting living alone. On the other hand, social support can be a result of living alone. Like Potts (1997), we have used perceived support from friends as proxies for social support for this purpose.

Marital status was a significant predictor of living alone in our study. Having a partner keeps the elderly from living alone, and feeling lonely can be expected whenever the relationships of attachment are lost (Kim, 1999). Seeman and Berkman (1988) examined the relationship between structural characteristics of social networks and two types of support, namely, emotional and instrumental support. They found that the spouse could be an important source of support. In addition, research has found that marriage affects the well-being of both men and women (Spitz & Ward, 2000), it provides individuals with a network of assistance and support (Barrett, 1999).

As gender was a strong predictor of living alone, which in turn affected level of perceived social support, one might argue that the observed difference in support received between the living alone and living with other groups could be explained by socialization history. We appreciate the fact that the population sampled in this study grew up in a peasant (traditional) Taiwanese society. The peasant society emphasizes the social customs and rules by identifying men with instrumental roles, that is, men should work hard and meet family expenses, while women should take good care of family members (Merton, 1968). In Chinese society, men are supposed to be fearless and independent. Consequently, men may be more reluctant to ask for support. Moreover, many elderly men in Taiwan may have devoted a significant portion of their lives to their career, and so may not have developed supportive social networks or skills to marshal social support. Nevertheless, living alone was still a significant predictor of perceived support even after adjusting for the effect of gender and other covariates confirmed that those living alone did feel they had less social support.

The benefit of providing social support to different sex groups of elderly populations has been well studied. While many found that improved social support could lead to better health for men (Unger et al. 1999); others reported that it was more beneficial for women (Tower & Kasl, 1996); and some failed to find any gender differences (Umberson, 1992). Further research is needed to clarify not only the benefit of support for males and females, but also to investigate the effect of providing social support on the health of elderly individuals living alone. This will have implications for targeting the most appropriate group for interventions.

Elders in our sample who were supported by their own savings or retirement funds were more likely to live alone than were those who were supported by children. They are financially independent and can afford to have health care expenditure, assistive devices, nutritious food, and entertainment. This group may have sufficient income to compensate for a lack of social support by paying for household help or travel to visit distant relatives or friends. However, elderly people who had no income or who were receiving governmental assistance were even more overrepresented in the living alone group. This is similar to findings that living alone is a major risk factor for poverty in older American women (Lewis, 1997). This, perhaps, is the group towards which appropriate interventions should be targeted.

We found that living alone had a negative effect on loneliness and perceived social support. Other studies also have reported a negative association between social support and loneliness (Kim & Baik, 2002; Steverink, Westerhof, Bode, & Dittmann-Kohli, 2001). Loneliness is a subjective and unpleasant experience that occurs when an individual’s network of social relationships is significantly deficient in either quality or quantity (Peplau, 1985). Weiss (1989) emphasized the importance of distinguishing between emotional isolation and social isolation when studying loneliness in older people. The former results from the lack of a close intimate attachment to another person, while the latter results from the lack of a network of social relationships (Bondevik & Skogstad, 1998). Apart from partners, people also need friends to enhance social support. Perceiving emotional support is especially important for very old adults who face a variety of age-related challenges to their functional ability and health (Patrick, Cottrell, & Barnes, 2001). Older people are more likely to lose family members and friends and to be more vulnerable to loneliness and social isolation. Social engagement with friends probably challenges them to communicate effectively and to participate in interpersonal exchanges.

The results of our study do not necessarily imply that a policy or program that encourages social support for senior citizens living alone will lead to a reduction in associated health concerns. The justification for such a program is not limited to its potential effects on health issues but on the accumulated evidence that social activities and engagement are generally helpful to the elderly population. It may be impractical that all chronically ill and elderly individuals should live with their family members. In fact, many who live alone cherish their independence and want to remain in their own homes. It is our recommendation that health authorities should provide more incentives for elderly to interact more closely with their families and friends to increase their social activity levels.

Andersson, L. (1992). Loneliness and perceived responsibility and control in elderly community residents. Journal of Social Behavior and Personality, 3, 431-443.

Avlund, K., & Holstein, B. E. (1998). Functional ability among elderly people in three service settings: The discriminatory power of a new functional ability scale. European Journal of Epidemiology, 14, 783-790.

Barrett, A. E. (1999). Social support and life satisfaction among the never married: Examining the effects of age. Research on Aging, 21, 46-72.

Bondevik, M., & Skogstad, A. (1998). The oldest old, ADL, social network, and loneliness. Western Journal of Nursing Research, 20, 325-343.

Campion, E. W. (1996). Home alone, and in danger. The New England Journal of Medicine, 334, 1738-1739.

Fabrigoule, C., Letenneur, L., Dartigues, J. F., Zarrouk, M., Commenges, D., & Barberger-Gateau, P. (1995). Social and leisure activities and risk of dementia: A prospective longitudinal study. Journal of the American Geriatrics Society, 43, 485-490.

Fratiglioni, L., Wang, H., Ericsson, K., Maytan, M., & Winblad. B. (2000). Influence of social network on occurrence of dementia: A community-based longitudinal study. The Lancet, 355, 1315-1319.

Fried, T., Bradley, E. H., Williams, C. S., & Tinetti. M. E. (2001). Functional disability and health care expenditures for older persons. Archives of Internal Medicine, 161, 2602-2607.

Gurley, R. J., Lum, N., Sande, M., Lo, B., & Katz. M. H. (1996). Persons found in their homes helpless or dead. New England Journal of Medicine, 334, 1710-1716.

Hays, J. C., Steffens, D. C., Flint, E. P., Bosworth, H., & George, L. (2001). Does social support buffer functional decline in elderly patients with unipolar depression? American Journal of Psychiatry, 158, 1850- 1855.

Hicks, T. J. (2000). What is your life like now? Loneliness and elderly individuals residing in nursing homes. Journal of Gerontological Nursing, 26, 15-19.

Holmen, K., Ericsson, K., Andersson, L., & Winblad, B. (1992). Loneliness among elderly people living in Stockholm: A population study. Journal of Advanced Nursing, 17, 43-51.

Katz, S., & Akpom, C. A. (1976). 12. Index of ADL. Medical Care, 14, 116-118.

Kim, O. (1999). Predictors of loneliness in elderly Korean immigrant women living in the United States of America. Journal of Advanced Nursing, 29, 1082-1088.

Kim, O. S., & Baik, S. H. (2002). Loneliness, social support and family function among elderly Korean women. Nursing Health Science 4, A7-8.

King, M. B., & Tinetti, M. E. (1995). Falls in community-dwelling older persons. Journal of the American Geriatric Society, 43, 1146-1154.

Lewis, M. I. (1997). An economic profile of American older women. Journal of American Medical Women’s’ Association, 52, 107-112.

Lichtenberg, P. A., MacNeill, S. E., & Mast. B. T. (2000). Environmental press and adaptation to disability in hospitalized live-alone older adults. The Gerontologist, 40, 549-556.

Merton, R. K. (1968). Social theory and social structure. New York: Free Press.

Patrick, J. H., Cottrell, L. E., & Barnes, K. A. (2001). Gender, emotional support, and well-being among the rural elderly. Sex Roles, 45, 15-29.

Peplau, L. A. (1985). Loneliness research: Basic concepts and findings. In I. J. Sarason & B. R. Sarason (Eds.) Social support: Theory, research, and applications. Boston, MA: Martinus Nijhoff.

Pfeiffer, E. (1975). A Short Portable Mental Status Questionnaire for the assessment of organic brain deficit in elderly patients. Journal of the American Geriatrics Society, 23, 433-441.

Potts, M. K. (1997). Social support and depression among older adults living alone: The importance of friends within and outside of a retirement community. Social Work, 42, 348-362.

Power, C., Matthews, S., & Manor. O. (1998). Inequalities in self-rated health: Explanations from different stages of life. Lancet, 351, 1009-1014.

Seeman, T. E., & Berkman, L. F. (1988). Structural characteristics of social networks and their relationship with social support in the elderly: Who provides support. Social Science & Medicine, 26, 737-749.

Spitze, G., & Ward, R. (2000). Gender, marriage, and expectations for personal care. Research On Aging, 22, 451-469.

Steverink, N., Westerhof, G. J., Bode, C., & Dittmann-Kohli, F. (2001). The personal experience of aging, individual resources, and subjective well-being. The Journals of Gerontology, Series B., Psychological Sciences and Social Sciences, 56, 364-373.

Tower, R. B., & Kasl, S. V. (1996). Gender, marital closeness, and depressive symptoms in elderly couples. The Journals of Gerontology, Series B., Psychological Sciences and Social Sciences, 51, 115-129.

Umberson, D. (1992). Gender, marital status and the social control of health behavior. Social Science and Medicine, 34, 907-917.

Unger, B., McAvay, G., Bruce, M. L., Berkman, L., & Seeman, T. (1999). Variation in the impact of social network characteristics on physical functioning in elderly persons: Macarthur studies of successful aging. The Journal of Gerontology, 54B, S245-S251.

Weiss, R. S. (1989). Reflections on the present state of loneliness research. In M. Hojat & R. Crandall (Eds.), Loneliness: Theory, research, and applications (p. 116). Newbury Park, CA: Sage.

Wu, S. C., & Chang, M. T. (1997). Elderly population in Taiwan (pp. 56-60). Institute of Health Policy & Management of National Taiwan University: Taipei, Taiwan.