Social isolation in older Malaysians: prevalence and risk factors
Correspondence: Dr Rahimah Ibrahim PhD, Institute of Gerontology, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia. Email: firstname.lastname@example.org
Social isolation is one of the most important emerging issues among ageing populations, as it reduces well-being, health and quality of life. The purpose of this study was to identify prevalence and risk factors of social isolation in older Malaysians.
The sample for this study was drawn from a national survey entitled ‘Patterns of Social Relationships and Psychological Well-Being among Older Persons in Peninsular Malaysia’. Social isolation was measured with the Lubben Social Network Scale.
The findings from the present study showed that 49.8% of older Malaysians are at risk for social isolation. The results of logistic regression analysis revealed that the number of sons, number of brothers, number of sisters, household size, self-rated health, place of residence, homeownership, sex and ethnicity were significantly associated with social isolation.
These findings may have some implications for social and health-care policymakers in planning and developing new and effective interventions such as educational programmes to reduce social isolation among this vulnerable population
Malaysia's population continues to undergo a demographic transition that started around 1969, when the fertility rate began to decline. In the 1980s the fertility rate was 5.6; by 1990, it had declined to 3.7, and by 2009, to 2.5. At the same time, Malaysia has enjoyed a rapid increase in life expectancy. Life expectancy at birth in Malaysia has increased from 55.8 years for men and 58.2 years for women in 1957 to 71.9 years and 76.4 years, respectively, in 2007. Furthermore, Malaysia is also experiencing epidemiological transitions. The major cause of death has changed from infectious diseases to non-communicable diseases.
One of the most important indicators of healthy ageing and healthy communities is social integration. Nowadays the number of seniors at risk for social isolation, a significant predictor of morbidity and mortality, is increasing. In the social science and health-care literature, social isolation has been found to be a prevalent problem and one of the most important emerging issues related to growing older.[6, 7]
Because social isolation is a complex concept, it has been defined differently in the social and gerontological literature. For example, it has been defined as the absence of or a decrease in the number of social interactions, contacts and relationships with other people, particularly family and friends,[9-11] a low level of integration or involvement in society[12, 13] and a low number of social contacts and interactions with others. Given these definitions, it can be concluded that social isolation is conceptualized as objective characteristics of personal relationships that can be quantified in terms of composition and size of social network and by the frequency of interactions with those that a person can share meaningful and supportive relations.
Social isolation has been discussed not only objectively but also subjectively. For instance, Cohen and Syme objectively consider social isolation as having low levels of social contact. Subjectively, their definition of social isolation is the feeling of being separate from others and being an outsider, isolated and suffering from loneliness. Social isolation refers to a lack of a sizable network of social relationships, and loneliness has been defined as an unpleasant feeling of lacking certain relationships. As such, socially isolated persons are not necessarily lonely; loneliness depends on whether the isolation was voluntary or whether a person enjoys solitude. With regard to this distinction, loneliness therefore represents the more qualitative aspect of personal relationships, and its intensity can only be described by the person who experienced it in terms of a deficit between actual and desired quality and quantity of engagement or ‘embeddedness’ within the social context.[10, 13, 17] It means that people's subjective evaluation or perception of their network of social relationships can be an unpleasant experience when it is perceived as deficient in emotional attachments and important social connections. In addition to being a dimension of qualitative emotional isolation among the elderly, social isolation was seen as a distinct quantitative dimension of loneliness. Therefore, social isolation and loneliness are seen as interrelated constructs that can be differentiated by quantitative/objective and qualitative/subjective domains.
In sum, social isolation has been defined both objectively (the lack of contact and interaction with other people) and subjectively (the feeling of loneliness or lack of companionship or close, genuine communication with others).[11, 19] Thus, on a spectrum, we would have social isolation or being disconnected, excluded or marginalized from society at one end, and social participation or being engaged or integrated with others in the social setting on the other end; however, each circumstance may have varying effects on a person's subjective experience. It should be noted that loneliness and social isolation are distinct but interrelated concepts, although they have often been used interchangeably. Nonetheless, both objective and perceived social isolation (i.e. loneliness) have been found to negatively impact on older persons' health and well-being.[21, 22]
A growing body of literature provides strong evidence that social isolation can increase in advanced ages, significantly affect psychological well-being, and increase the risk of suicide among elderly people. Similarly, perceived social isolation or loneliness was identified as a ‘geriatric giant’ because of its adverse consequences on physical health, such as elevated blood pressure, lower immune response, reduced cognition and sleep problems,[13, 26] and on psychological well-being, as it is closely linked to depression.[21, 22, 26-28] Most importantly, in the review by Cacioppo and Hawkley, perceived social isolation, which is related to the quality of interaction rather than quantity, was found to be a stronger predictor of negative health outcomes, negative appraisal of safety in one's environment and lower quality of life than objective social isolation. Thus, the importance of dealing with social isolation to improve older people's well-being and quality of life is increasingly recognized in international policy.
For a number reasons the study of social isolation among elderly people is very important. First, there is some evidence that social isolation may become more common with increasing age. Second, older adults are more likely to experience bereavement and chronic health problems, both of which may increase their need for social support and companionship. Third, research indicates that older adults who experience social isolation are at greater risk for mortality, increased morbidity, depression and cognitive decline. Finally, the World Health Organization has emphasized that for good health, the prevention of social isolation is necessary.
In light of this consideration and the potential threat of social isolation to the physical and mental health of older persons, social isolation is receiving increased attention from gerontological researchers. A review of literature related to ageing shows a prevalence of social isolation and loneliness ranging from 2% to 49% and from 3% to 90%, respectively.[32-35] The great variety in research findings may be the result of differing research designs, divers measuring instruments, and definitions and estimations of social isolation.
A series of gerontological studies has identified several risk factors that can lead to social isolation, such as, poor health, mental illness, geographic location, communication difficulties, place of residence, being a single man, and transport difficulties. It has been found that predictors of social isolation can differ based on health-care opportunities and social programmes offered within an individual's country of residence. In order to prevent and alleviate social isolation among older people, identifying risk factors is an important area for policy and practice. Because social isolation among older Malaysians is not well understood, this study was conducted.
The purpose of the present study is twofold: (i) to describe the prevalence of social isolation and (ii) to identify sociodemographic and health characteristics that predict social isolation amongst older Malaysians living in the community.
Materials and Methods
The data for this study were derived from a national survey entitled ‘Patterns of Social Relationships and Psychological Well-Being among Older Persons in Peninsular Malaysia’, which involved cross-sectional research from 2007 to 2009. The details of methodology have been published in full elsewhere. Briefly, a multistage cluster sampling design was employed to obtain a representative sample of community-dwelling elderly persons living in Peninsular Malaysia. The overall sample size of the survey was approximately 1880 people aged 60 years and older. Data were collected through face-to-face interviews in respondents' homes conducting by trained interviewers using standardized questionnaires.
Social isolation was measured with the Lubben Social Network Scale-6 (LSNS-6). LSNS-6 is a brief, validated instrument developed to gauge social isolation in older adults by measuring size, closeness and frequency of contact with friends and family members. The total scale score is obtained by summing the six items, with scores ranging from 0 to 30. A score below 12 suggests that an individual is at risk of social isolation. LSNS-6 has been validated in several countries,[8, 39-43] including Malaysia. In this study, the Cronbach's alpha for the total LSNS-6 was 0.81.
Sociodemographic variables included sex, age, number of siblings, number of children, household size, household income, homeownership, place of residence, employment status, marital status and ethnicity. Self-rated health was measured on a Likert scale with responses ranging from poor to excellent.
SPSS v19.0 (IBM, Amonk, NY, USA) was used for data analysis. Ranges, frequency distributions, percentages, means and standard deviations were utilized to describe preliminary data. χ2 analyses and a series of t-tests were employed to compare sociodemographic, socioeconomic and health status of socially isolated and non-socially isolated respondents. A binary logistic regression analysis was used to identify significant factors of social isolation. In the first step, the problems of multicollinearity and fitness of the model were assessed. Problems of multicollinearity in the regression model were evaluated with the bivariate correlation matrices of independent variables. Exploratory data analysis revealed that assumptions for binary logistic regression, including no outliers, low multicollinearity and low error in explanatory variables, had been met. Goodness-of-fit test was measured by the Hosmer–Lemeshow test. For a well-fit model, the Hosmer–Lemeshow statistic should be non-significant (P > 0.05), which would indicate that there was no difference between observed and model-predicted values and suggests that the model's estimates fit the data at an acceptable level.[45, 46] Results of the Hosmer–Lemeshow test (χ2(8) = 11.05, P = 0.199) indicated that the model was acceptable, meaning that the model predicted values significantly similar to what they should be.
In the study, all possible variables were entered into one model. The dependent variable (respondents' social isolation) was dichotomized into 1 = socially isolated and 0 = non-socially isolated. The sociodemographic variables that were tested included age, number of siblings, number of children, sex (man = 1, woman = 0), marital status (married = 1, not married = 0), place of residence (urban = 1, rural = 0), ethnicity (Malay = 1, non-Malay = 0), level of education attained, employment (employed = 1, unemployed = 0), total household income, and homeownership (homeownership = 1, no homeownership = 0).
All respondents were 60 years of age or older (69.7 ± 7.6 years), and there were 989 women (52.7%). Almost 53.0% of the respondents examined in this study were married; the rest reported their marital status as separated/divorced, widowed or never been married. Less than one-quarter (24.3%) of the sample was employed. With regard to education, 38.0% of respondents had no formal education, slightly more than half (51.7%) had a primary school education, 9.7% had a secondary school education and 1.1% had some university education or were university graduates.
To assess the representativeness of the respondent pool, the demographic characteristics of the sample were compared with the national population aged 60 years and over based on the 2000 National Census data. Representativeness analysis was conducted using χ2 goodness-of-fit tests and one sample t-test to assess whether survey respondents were representative of the population in terms of sex, place of residence and age. Results of the representativeness analysis showed no significant differences in terms of sex (χ2 = 1.79, P = 0.181), place of residence (χ2 = 1.04, P = 0.307) or age (χ2 = 0.94, P = 0.347), indicating that respondents were closely representative of the elderly Malaysian population.
The prevalence of social isolation was based on an LSNS-6 score of less than 12; 49.8% of the respondents in this study were found to be at risk for social isolation. Prevalence of social isolation among older women was 54.8% , and 44.3% among older men.
Results of χ2 analyses and a series of t-tests are summarized in Table 1. As shown in Table 1, cross-tabular analysis revealed that two groups of respondents (socially isolated and non-socially isolated) differed significantly on the basis of employment status (χ2(1) = 6.25, P ≤ 0.05), ethnicity (χ2(1) = 9.90, P ≤ 0.01), homeownership (χ2(1) = 10.74, P ≥ 0.01), place of residence (χ2(1) = 6.98, P ≤ 0.01), marital status (χ2(1) = 15.11, P ≤ 0.01), sex (χ2(1) = 19.90, P ≤ 0.01) and level of education attained (χ2(2) = 8.23, P ≤ 0.05). Additionally, the findings from a series of t-tests showed that the two groups (socially isolated and non-socially isolated) differed significantly in terms of age (70.2 vs 69.1 years, P ≤ 0.01), number of brothers (1.7 vs 2.4, P ≤ 0.01), number of sisters (1.9 vs 2.3, P ≤ 0.01) and number of sons (2.5 vs 2.7, P ≤ 0.01). There was also a significant difference between socially isolated older adults and their non-socially isolated counterparts with regard to their health status (1.8 vs 1.9, P ≥ 0.01) based on a self-rated health scale. No significant differences were found between socially isolated and non-socially isolated older adults in terms of number of daughters, household size and household income.
Table 1. A comparison of sociodemographic factors in socially and non-socially isolated elderly
|Employment status|| || || || || || || |
|Unemployed||51.3|| || ||48.7|| || ||6.25*|
|Employed||44.5|| || ||55.5|| || || |
|Ethnicity|| || || || || || || |
|Non-Malay||56.1|| || ||43.9|| || ||9.90**|
|Malay||47.7|| || ||52.3|| || || |
|Homeownership|| || || || || || || |
|Non-homeownership||62.1|| || ||37.9|| || ||10.74**|
|Homeownership||48.6|| || ||51.4|| || || |
|Place of residence|| || || || || || || |
|Rural||47.0|| || ||53.0|| || ||6.98**|
|Urban||53.5|| || ||46.5|| || || |
|Marital status|| || || || || || || |
|Unmarried||54.8|| || ||45.2|| || ||15.11**|
|Married||45.8|| || ||54.2|| || || |
|Sex|| || || || || || || |
|Women||54.6|| || ||45.4|| || ||19.90**|
|Men||44.3|| || ||55.7|| || || |
|Education|| || || || || || || |
|Non-formal education||53.7|| || ||46.3|| || ||8.23*|
|Primary education||47.8|| || ||52.2|| || || |
|Secondary and/or tertiary education||44.3|| || ||55.7|| || || |
|Age|| ||70.2 years||7.72|| ||69.1 years||7.37||−3.16**|
|Self-rated health|| ||1.8||0.62|| ||1.9||0.61||4.00**|
|Number of sons|| ||2.5||1.76|| ||2.7||1.77||2.63**|
|Number of daughters||2.5||1.75|| ||2.6||1.70||1.04|
|Number of brothers|| ||1.7||1.79|| ||2.4||2.08||7.09**|
|Number of sisters|| ||1.9||1.77|| ||2.3||1.94||5.59**|
|Household size|| ||4.4||2.94|| ||4.2||2.52||−1.26|
|Household income|| ||$408.97||1290.36|| ||$453.71||437.8||1.01|
In the next stage, a binary logistic regression was utilized to identify what significant predictors contributed to social isolation. The finding of the binary logistic regression revealed a significant model (χ2(15) = 114.56, P ≤ 0.01). The following factors were significantly associated with social isolation: number of sons, number of brothers, number of sisters, household size, self-rated health, place of residence, homeownership, sex, marital status and ethnicity. Age, marital status, household income and level of education were not significantly associated with social isolation. Table 2 shows that the likelihood of being socially isolated was decreased in men (odds ratio (OR) = 0.75, P ≤ 0.05, 95% confidence interval (CI) = 0.59–0.96), number of sons (OR = 0.94, P ≤ 0.05, 95%CI = 0.89–1.00), number of brothers (OR = 0.88, P ≤ 0.01, 95%CI = 0.83–0.93), number of sisters (OR = 0.92, P ≤ 0.01, 95%CI = 0.87–0.97), health status (OR = 0.77, P ≤ 0.01, 95%CI = 0.65–0.91), homeownership (OR = 0.69, P ≤ 0.05, 95%CI = 0.48–1.01) and being Malay (OR = 0.68, P ≤ 0.05, 95%CI = 0.52–0.87). Further results from Table 2 show that living in an urban area (OR = 1.31, P ≤ 0.05, 95%CI = 1.06–1.61) and a larger household (OR = 1.04, P ≤ 0.05, 95%CI = 1.00–1.08) significantly increased the odds of being socially isolated.
Table 2. Results of binary logistic regression
|Number of sons||−0.06||0.03||3.96||0.94*||0.89||1.00|
|Number of daughters||−0.01||0.03||0.08||0.99||0.94||1.05|
|Number of brothers||−0.13||0.03||18.70||0.88**||0.83||0.93|
|Number of sisters||−0.08||0.03||8.00||0.92**||0.87||0.97|
|Place of residence||0.27||0.11||6.41||1.31*||1.06||1.61|
The primary purpose of the present study was to identify the prevalence rate of social isolation among a representative sample of older Malaysians aged 60 years and older. We used an objective approach, specifically, the number of interactions and social contacts with family and friends, to measure social isolation according to LSNS-6. The employed dataset did not include items on subjective or perceived social isolation, which can be identified based on existing measurements of loneliness such as the UCLA Loneliness Scale and De Jong Gierveld Loneliness Scale. Measurements were limited to the objective approach.
Our finding showed that slightly less than half (49.8%) of the respondents were at risk for social isolation, which represents a high rate of social isolation among older Malaysians. This finding supports the gerontological literature concerning the high prevalence of social isolation and loneliness among elderly people. For example, in a systematic review of intervention for social isolation, Dickens et al. reported that the prevalence of social isolation is 7%–17%, depending on how social isolation was defined and measured.  In another study by Grieves and Farbus, the prevalence of social isolation was found to range from 2% to 20% and to go as high as 35% for elderly in assisted-living situations. To compare, the prevalence of loneliness among the elderly in Finland has been reported to be 40%. Similarly, the finding from a Swedish study showed that 35% of older individuals suffered from loneliness. In contrast, Johnson et al. reported that 62% of elderly people in the USA experience loneliness. These great differences in prevalence of social isolation and loneliness among older persons in different studies may be due to varying measurement methods.
The second main aim of this study was to identify the predictors of social isolation among older Malaysians living in the community. The findings from binary logistic regression revealed that the number of sons, number of brothers, number of sisters, household size, self-rated health, place of residence, homeownership, sex and ethnicity were statistically associated with social isolation.
The sex of respondents was found to be a significant determinant of social isolation, with older women were more likely to be socially isolated than older men. This finding supports the previous studies.[49, 50] The greater likelihood that older women will be alone increases their risk of suffering from social isolation.[15, 44, 52-54] However, this finding was inconsistent with a few studies that have documented older men as being much more likely to be isolated than women.[55-57]
The next factor statistically associated with social isolation was ethnicity. Older Malays were less likely to be at risk of social isolation than other ethnic groups such as Chinese and Indians. This finding supports the results from previous studies.[31, 58] It has been suggested that ethnicity may moderate the relationship between sociodemographic factors and social isolation.
Finding of this study also showed that health status is another significant predictor of social isolation. This result parallels the findings of previous studies.[56, 59, 60] For example, Mullins et al. found that an individual's perceived health is associated with social isolation. Results of another study of elderly people also showed that residents of assisted living facilities who experienced chronic health problems, sensory impairments and long-term illness were more socially isolated from family and other residents than those residents who were healthier. Similarly, Havens et al. found that health and social factors are important predictors of social isolation.
Non-homeownership was found to be another significant risk factor for social isolation among older Malaysians. It seems that this factor may reduce the number of opportunities for social interaction with others, potentially leading to increased social isolation. A similar finding was reported among older Canadians.
Our results also revealed that the place where respondents lived (urban vs rural) was significantly related to social isolation. The present study found older persons living in urban areas were at greater risk of social isolation than those living in rural areas. This finding is inconsistent with a study conducted of elderly Finns that found that those living in a large city were less often lonely than those living in small cities or rural areas. One reason why elderly living in urban areas suffer from more social isolation is that older urban residents may have less frequent contact with friends or relatives than those living in rural areas.
Another significant risk factor for social isolation was found to be household size. Results of binary logistic regression revealed that larger household size is associated with greater risk of social isolation. This finding is inconsistent with earlier studies in which larger household size promoted intergenerational support and reduced the isolation of elderly people.[61, 62]
One possible reason for the varying result is that larger household size may reflect economic constraints rather than preference and may be associated with overcrowding, poor sanitation and abuse. In our study, the mean household size and household income of the socially isolated group were 4.4 people and $408.97, respectively, compared with 4.2 people and $453.71 for the non-socially isolated group.
Further results revealed that a greater number of siblings is significantly associated with a lower risk of social isolation among elderly people. This finding supports studies showing that older people with living siblings have higher life satisfaction, a greater sense of emotional security, fewer depressive symptoms and greater feelings of control over their lives.
The number of sons was found to be a significant determinant of social isolation among older Malaysians, but neither bivariate correlation nor binary logistic regression found a significant association between the number of daughters and social isolation. A possible explanation for non-significant association between the number of daughters and social isolation is that older Malaysians are more likely to live with their sons rather than their daughters, particularly when children are married. This occurs as a result of cultural norms and practices, which may lead elderly persons to have less contact with their daughters. This finding parallels the results of a study conducted by Lin et al. in Taiwan. They studied sex differences in adult children's support of their parents among older Taiwanese and found that sons generally provided more support to their older parents than daughters.
In our study, the bivariate analyses showed that marital status, level of education and age are significantly associated with social isolation. These finding are consistent with results of a study of elderly Finns, which indicated that advancing age, widowhood, a low level of education and poor income increase risk for loneliness. Additionally, research shows that unmarried elderly people are specifically at risk of social isolation because an important element of their social network is missing. An interesting observation based on comparing the results of the bivariate and binary logistic regression is that social isolation is a function of respondents' sex and marital status, but only sex (i.e. being a woman) independently increased the likelihood for social isolation. Marital status did not emerge as a significant predictor of social isolation probably because marital status differed significantly by sex (χ2 (3) = 392.53, P ≤ 0.001); 77.8% of older women were unmarried, divorced or separated compared to 22.2% of older men. This result may also imply that being a woman, rather than being unmarried, puts an individual at a higher risk for social isolation given the fact that in almost every country women normally outlive men, and the chances that older women remarry following a divorce or widowhood are lower than for older men.[67, 68]
The findings of this study provided additional support for the gerontological literature concerning the high prevalence of social isolation among elderly people and related factors. These findings may have some implications for social and health-care policymakers to plan new and effective interventions, such as educational programmes, to reduce social isolation among this vulnerable population. The emerging field of educational gerontology can have a leading role in promoting awareness and developing protocols to help isolated older persons. Currently, this field focuses on three educational goals: promoting positive attitudinal and behavioural change among adults, raising public awareness on ageing and improving attitudes towards the elderly, and training professionals working with older people.
Awareness creation, as a core goal of educational gerontology, contributes to removing stereotypes and creating an anti-discriminatory environment to prepare for more extensive social integration programmes to reduce social isolation in the community. As isolated elderly people are not easily detected and discovered, except in the event of crisis, public awareness about ageing and older persons is crucial in making the problem of social isolation more visible. Next, supportive protocols for the elderly include access to adult education programmes (e.g. University for the Third Age), as these programmes have the potential to improve the well-being of the elderly. As a case in point, participating in the University of the Third Age was found to improve the health and well-being of elderly people through the mediation of social support.
Another supportive protocol is in the form of support services such as health and nutrition, recreational and educational services, which enable socially isolated older adults to maintain their health and achieve the desired social competency. Professionals working in these areas with socially isolated older adults should be trained in using the critical educational gerontology approach to empower the elderly by helping them develop autonomy, self-assertiveness, and leadership skills.
This study is a part of the national survey ‘Patterns of Social Relationships and Psychological Well-Being among Older Persons in Peninsular Malaysia’ (PSRPWO). This project was funded by Ministry of Science, Technology and Innovation (Intensification of Research Priority Areas Project No.: 04-01-04-SF0479) (Kuala Lumpur, Malaysia). We are grateful to the older persons who participated in this study.