Social support in maintaining mental health and quality of life among community-dwelling older people with functional limitations in Malaysia: a population-based cross-sectional study

Abstract

Objective

This study aimed to examine the mediation role of perceived social support in the relationship between functional limitations, depressive symptoms and quality of life among older people in Malaysia.

Setting

The Malaysian National Health Morbidity Survey 2018: Elderly Health was a cross-sectional health community survey among adults aged 50 and above.

Participants

3977 community-dwelling older Malaysians aged 60 and above.

Outcome measures

Functional limitations were defined as personal activities of daily living (PADL) and instrumental activities of daily living (IADL), tested in separate paths in all analyses. PADL was measured using the Barthel Index, while IADL was measured using the Lawton and Brody scale. Perceived social support, depressive symptoms and quality of life were measured using the Duke Social Support Index, Geriatric Depression Scale-14 and Control, Autonomy, Self-Realisation and Pleasure-19 tools. We used mediation analysis through structural equation modelling to explore the role of perceived social support.

Results

Perceived social support mediated the relationship between PADL and IADL with depressive symptoms, with the indirect effects at −0.079 and −0.103, respectively (p<0.001). Similarly, perceived social support mediated the relationship between PADL and IADL with quality of life, with the indirect effects at 0.238 and 0.301, respectively (p<0.001). We performed serial multiple mediation analysis and found that perceived social support and depressive symptoms mediated the path between PADL and IADL with quality of life, with the indirect effects at 0.598 and 0.747, respectively (p<0.001). The relationship between functional limitations and all outcomes remained significant in all mediation analyses.

Conclusion

The present study provides evidence that perceived social support relieves the influence of functional limitations on depressive symptoms and declining quality of life among older people. Therefore, it is imperative to establish a social support system to improve the overall well-being of older people.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The National Health Morbidity Survey 2018 was a large and nationally representative health survey of older Malaysians, with 3977 respondents included in this study.

• The primary factors and outcomes were measured using validated questionnaires and scales, allowing comparison with other studies.

• Data was captured among older people living in the community, thus excluding older people residing in care facilities due to more severe functional limitations or depressive symptoms.

• The functional limitations, social support, depression and quality of life measurement tools were rated based on respondents’ perceptions and interpretations, thus subject to personal bias.

• This study analysed the perceived social support among the respondents while acknowledging that various other forms of social support exist.

Introduction

Ageing is a global challenge, driven by rising life expectancy, attributed to the improvement and higher quality of healthcare and falling fertility rates.^{1 2} Globally, the proportion of the population aged 65 and above increased from 6% in 1990 to 9% in 2019 and is expected to increase further to 16% by 2050.³ Similarly, the percentage of older people in Malaysia is projected to rise from merely 5% in 2010 to 14.5% in 2040.⁴ This demographic shift has posed a significant challenge to healthcare systems, demanding greater adaptability to address the diverse needs of older people.⁵

Among the major concerns for older people’s health are their mental well-being and quality of life. Of the various mental disorders, depression is among the highest contributor to disabilities and mortalities.⁶ The WHO estimated that 3.8% of the global population is affected by major depression, and higher among older people (5.7%) in 2021,⁷ while suicide is almost twice as frequent in older people than in younger populations.⁸ Depression occurring at any age is debilitating as it may impair functioning and quality of life and lead to various health problems.⁹ Late-life depression, or depression that begins or is detected in old age is associated with disability, increased mortality, poorer health outcomes and poorer quality of life as compared with those without depression.^10–12 Meanwhile, older people’s impaired quality of life signified a reduced ability to be independent, have autonomy and be satisfied with their lives. ^{13 14} On the other hand, evidence shows older people with a better quality of life tend to have better overall health, enabling them to age in place and live longer.^14–16 In Malaysia, depression among older people aged 60 and above in 2018 was 11.2%, while 28.6% perceived poor quality of life.¹⁷

A significant risk factor for older people developing depression and having poor quality of life is functional limitations.^18–21 Functional limitations often refer to challenges, restrictions or reliance on others when performing personal activities of daily living (PADL) or instrumental activities of daily living (IADL).^{22 23} PADL refers to a person’s ability to perform basic physical requirements, such as dressing, feeding, toileting, grooming, mobility and incontinence.²³ In contrast, IADL refers to activities that allow people to live independently in a community and are influenced by cognitive abilities.²² The WHO reported that 38.1% of older people aged 60 and above experience some form of disability, which is higher in lower-income countries.²⁴ Such functional decline is debilitating, with the resulting functional limitation often leading to decreased independence, a reduction in quality of life, various health problems and may lead to depression.^{25 26} While the relationship between functional limitations with both depression and poor quality of life is bidirectional and reciprocal, whereby each may impact the other in a progressive cycle,^{20 27} functional limitations are more often found to be prodromal towards the deterioration in both mental health and quality of life.²⁴ Apart from the impact at the personal level, functional limitations and disability among older people impose a significant financial burden. The price of caring for a disabled, older person is tripled from a non-disabled older person.²⁸ In Malaysia, the National Health Morbidity Survey (NHMS) 2018 reports that 17.0% of older people aged 60 and above had functional limitations in PADL, while 42.9% had limitations in IADL.¹⁷

The relationship between functional limitations, depressive symptoms and quality of life is complex and inter-related. While functional limitations are known to be risk factors associated with both depressive symptoms and quality of life, depressive symptoms alone are also risk factors for the quality of life.^{29 30} Depressive symptoms impair the quality of life through physiological and behavioural mechanisms such as autonomic nervous system dysfunction, inflammation, endothelial dysfunction and decreased participation in health-promoting activities.³¹ Thus, effective treatment and management of depressive symptoms predict a better quality of life. The study attempts to explain the dynamic and complex relationship between the primary factors (functional limitations) and outcomes (depressive symptoms and quality of life).

With various evidence linking functional limitations as significant risks of depression and poor quality of life among older people, research is shifting towards explaining the mechanism and ‘processes’ behind the development of the conditions. Evidence shows that one of the protective factors preventing or explaining the development of depression and declining quality of life among older people is having excellent social support.^32–35 The studies show that in the company of good or higher perceived social support, the decline in mental health and quality of life among older people who have developed functional limitations may be prevented.^32–35 Social support is any form of help made available to a person through their social connections with other people. It encompasses the benefits people gain from one another through information or expertise, emotional support, assistance and self-sufficiency.³⁶ While various forms of social support exist, perceived social support, defined as the satisfaction of feeling understood, respected and supported by a person’s social network, was found to have the most significant impact on improving older people’s well-being.^{37 38}

The examination of social support’s role necessitates contextual consideration. Malaysia, characterised by ethnic diversity, primarily comprises Malays, Chinese and Indians, each adhering to distinct religions. Despite these variations, cultural practices concerning support for older people, defined as those aged 60 and above, are notably shaped by a foundation of respect and responsibility towards them.³⁹ The prevalent approach involves close-knit familial support, often manifested through shared living arrangements dedicated to caring for older family members. Notably, approximately 70% of older people in Malaysia reside with family members, a pattern aligning with other Asian countries but exhibiting a lower prevalence compared with western nations.^{40 41} Despite this, the NHMS 2018 reports that 30.8% of older people above 60 in Malaysia perceived having poor social support¹⁷; this highlights the pressing need to enhance awareness and improve the provision of social support. In Malaysia, the healthcare system operates on a dichotomous model. The publicly funded health sector, financed through taxation, provides nearly free access to healthcare services for the entire population. Additionally, there exists a private health sector where services are fee-based. This dual approach ensures widespread accessibility to diverse healthcare services, encompassing the management of functional limitations and depressive symptoms.⁴²

As research progressed from estimating associations between factors to delving into the underlying mechanisms that come into play between a risk factor and an outcome,²⁶ a branch of study on mediation and moderation began to proliferate. In the context of the current study, social support is viewed as a mediator due to its potential role in explaining the changes in the depressive symptoms and quality of life of older people, as shown in various evidence.^{34 35 43} One of the criteria for being a mediator is that the independent variable must explain the mediator being tested.⁴⁴ This study refers to functional limitations explaining social support. Evidence shows that those with functional limitations tend to lead towards isolation and loneliness, influencing how they perceive the social support they receive.^{34 45} They may find it harder to get the support they need, as the functional limitation impairs their ability to get help and engage in social activities. At the same time, older people with functional limitations may feel like a burden to their social network, altering how they perceive social support.³⁴ A Malaysian study among older people aged 60 and above found that the ability to perform PADL and IADL significantly predict changes in social support levels.⁴⁶ Hence, social support potentially mediates the complex relationship between functional limitations-depressive symptoms-quality of life of older people.

Thus, we aim to contribute to the growing evidence of the role of social support in mediating the path between functional limitations and depression, as well as the quality of life among community-dwelling older people. We extend the analysis by incorporating serial mediation and testing both social support and depression in the trajectory of quality of life. Furthermore, social support is constructed and perceived differently across countries and contexts. We add depth to the evidence by reporting findings from a middle-income, Southeast Asian country with a unique culture shaping its social support practice and interpretation. This finding is critical for policymakers as an input to design appropriate interventions for the improved well-being of older people, looking from the perspective of social care.

Methods

Study design and setting

We used data from the NHMS 2018: Elderly Health,¹⁷ a cross-sectional, community-based survey among adults aged 50 and above in Malaysia. It used a two-stage stratified cluster sampling design based on the Department of Statistics Malaysia’s sampling frame, from which 83 000 enumeration blocks (EB) of about 80–120 living quarters were considered. A total of 110 EB were randomly selected across all states and strata in Malaysia, giving a total of 5636 eligible living quarters to be included in the survey. The survey was conducted via face-to-face interviews at respondents’ places of residence between August and October 2018. The details of the study design and sample size determination for the NHMS 2018 can be found in the published report.¹⁷ To focus on older people’s health in Malaysia, data from older people aged 60 and above, comprising 3977 respondents, were selected and analysed through a subpopulation analysis approach. This aligns with the guidelines set by the Malaysian Public Service Department, which designates the age of 60 as indicative of old age and retirement.⁴⁷ Furthermore, it adheres to the definition of older people outlined in the NHMS 2018, referring to the population aged 60 and above. All participants provided written informed consent before interviews during the NHMS 2018’s data collection.

Data

Functional limitations

This study investigated both variables representing functional limitations, PADL and IADL, as independent variables. The original 10-item Barthel Index of PADL measured the ability to perform the activities.⁴⁸ A maximum score of 20 indicates a person’s absence of PADL functional limitation, while a decreasing score indicates a decline in functional ability.

The Lawton and Brody IADL scale measured the IADL, with a total score of 8, indicating the absence of IADL limitation. The decreasing score indicates worsening function. According to a local validation study, the Malay version of the Lawton IADL has an internal consistency of 0.838.⁴⁹

Depression

The Geriatric Depression Scale (GDS)-14, a 14-item index, was chosen by the NHMS study to determine the presence of depressive symptoms among older people in Malaysia.¹⁷ The items explored depressive symptoms in the past week before the interview. Each item was answered either ‘Yes’ or ‘No’ based on the presence of symptoms. A response of ‘Yes’ was scored 1, while a response of ‘No’ was scored 0, resulting in a maximum possible total score of 14. A higher total score indicates a greater level of depression. The 14-item GDS was derived from the GDS-15 following a validation study in Malaysia, which found that one item had no discriminatory value in differentiating depression and non-depression. Hence, it was dropped from the scale.⁵⁰ The original tool from which GDS-14 was derived, the GDS-15, was used and validated through various studies.^{51 52}

Quality of life

The perceived quality of life of older people was measured using the quality of life scale of Control, Autonomy, Self-Realisation and Pleasure (CASP-19), with the total score generated by summing all 19 items between a score of 0–3 (0=never, 3=often), yielding a range of 0–57.¹⁷ A higher score indicates a better quality of life. A local study translating the CASP-19 into two languages, Malay and Mandarin, found that the internal consistency was high for the original and translated versions (Cronbach’s alpha >0.8) and acceptable construct validity.⁵³

Social support

Perceived social support refers to the total score from the 11-item Duke Social Support Index (DSSI). The index was divided into two sections: the first on the size and structure of the social network (four items), while the second measures the older people’s satisfaction towards behavioural and emotional support received from their social network (seven items). Each item was scored from 1 to a maximum of 3 for a total maximum score of 33. A higher score indicates a higher level of perceived social support. The DSSI and its subscales had consistent patterns of low to moderate correlations. The availability of social integration had strong correlations with the overall DSSI score and its two subcomponents (0.57, 0.38, 0.53).⁵⁴ A study among 565 community-dwelling older people aged 70 and above found internal consistency using Cronbach’s alpha, with the overall index at 0.77 and test–retest ability scores ranging from 0.70 to 0.81.⁵⁴ The tool was translated into Malay in Malaysia and validated in a study among older people in Kuala Pilah in 2016 with a Cronbach’s alpha of 0.79.⁵⁵

Covariates

Various covariates were included to control for socio-demographic and economic factors, including age, sex, ethnicity, strata location, highest education level, individual monthly income, employment status and marital status.

Statistical analysis

We described the characteristics of the community-dwelling older people aged 60 and above based on their socio-demographic and economic background. The sample was weighted to adjust for varying selection probabilities, non-response rate, strata, age and sex based on the 2018 Malaysian population data by the Department of Statistics Malaysia.⁵⁶ The weighted frequency (n) and percentage (%) described categorical data.

Multiple linear regressions were conducted to examine the impact of functional limitations and social support on depressive symptoms and quality of life, controlling for various covariates. These analyses allowed a comprehensive understanding of how independent variables and potential mediators affect the outcomes. Additionally, the score for depressive symptoms was added to the regression predicting quality of life to investigate the relationship between the two factors. A post-estimation analysis employing an omnibus test was performed to discern categorical predictors with more than two levels that exhibited significant associations with the outcome. Weightage was applied to all analyses.

To investigate the potential mediating role of social support between functional limitations and depressive symptoms and quality of life, a mediation analysis was performed using the structural equation modelling method. The analysis included adjustments for population estimation through the application of appropriate weighting. In separate models, the depressive symptoms and quality of life were entered as the dependent variables, the functional limitations (PADL and IADL) as independent variables and social support as the mediator being tested, adjusting for covariates found significantly associated with the outcomes from the previous multiple linear regressions. Figures 1(a) and (b) of online supplemental appendix A depict the relationship and hypothesis tested.

A mediation effect was confirmed when functional limitations, either PADL or IADL or both, affect the dependent outcomes (depressive symptoms or quality of life, in separate models) through the potential intervening mediator (social support). In this analysis, the total effect (path c) refers to the impact of the independent variables on the dependent variable. It consisted of the independent variables’ direct effects (path c) on the dependent variables and the independent variables’ indirect effects (path a×b) on the dependent variables through the mediator. The coefficients and corresponding robust SEs were used to estimate the indirect effects, which indicated the magnitude of changes in the outcomes when the independent variables (functional limitations) were held constant and the mediator (social support) changed by the amount it would have changed if the independent variable had increased by one unit.⁵⁷ Bootstrapping was employed to assess the indirect effect of the mediation. A 95% CI level for the bias-corrected results was determined based on 5000 bootstrapped iterations.⁵⁸ In addition, the mediation effect was considered significant when the indirect effect mediated by social support did not contain 0 between the lower and upper confidence limits.⁵⁹ The proportion mediated by the indirect effect over the total effect and the standardised coefficients of the indirect effects were reported to represent the effect sizes.⁶⁰

Subsequently, a serial mediation analysis was performed, testing social support and depression as potential mediators affecting the trajectory between functional limitations and quality of life (figure 1(c) of online supplemental appendix A). This model generated three indirect effects, with an additional path d indicating the effect between social support and depressive symptoms. The three indirect effects were calculated based on a1×b1, a2×b2 and a1×db2 (serial mediation).

To assess the robustness of the mediation models in estimating both direct and indirect effects when assumptions regarding confounding were potentially violated, we conducted additional mediations excluding the covariates. Subsequently, we conducted a sensitivity analysis on these models to quantify the strength of a confounder required to alter the conclusions drawn. This sensitivity analysis involved measuring the correlation parameter (ρ), which indicates the presence of omitted variables associated with the mediator and outcome. This parameter was incorporated into the average causal mediation effect calculations.⁶¹ All analyses were conducted through Stata V.16.0 (StataCorp, College Station, Texas, USA), with the significant level set at 0.05.

Patient and public involvement

Patients and the public were not involved in this study’s design, conduct, reporting or dissemination plans.

Results

The characteristics of community-dwelling older people above 60 involved in the NHMS 2018, comprising 3977 respondents, representing 3 230 340 older Malaysians, are shown in table 1. The majority of respondents fell within the 60–69 age category, comprising 66.5% (95% CI: 63.7% to 69.1%) older people. Most respondents had up to the primary level of education (14.5% no formal education (95% CI: 12.5% to 16.9%), 43.6% primary education (95% CI: 39.4% to 47.9%)), were unemployed, 75.7% (95% CI: 73.6% to 77.7%), had less than Malaysian ringgit (MYR)1000, equivalent to US$216 of monthly income (exchange rate US$1=MYR4.62⁶²), 57.3% (95% CI: 53.4% to 61.1%) and were married, 67.9% (95% CI: 65.2% to 70.5%), at the time of the study. Respondents of Malay ethnicity were the most prevalent at 57.7% (95% CI: 48.7% to 66.2%).

Table 1.

Socio-demographic and economic characteristics of community-dwelling older people above 60 (n=3977)

Variables	Count	Estimated population	Weighted %	95% CI
				LL	UL
Overall	3977	3 230 340	100.0	–	–
Sex
Male	1872	1 580 226	48.9	47.2	50.6
Female	2105	1 650 114	51.1	49.4	52.8
Age group
60–69 years	2563	2 148 278	66.5	63.7	69.1
70–79 years	1104	838 860	26.0	23.7	28.3
≥80 years	310	243 202	7.5	6.3	8.9
Education
No formal education	806	469 794	14.5	12.5	16.9
Primary education	1939	1 408 624	43.6	39.4	47.9
Secondary education	967	1 040 544	32.2	28.8	35.8
Tertiary education	265	311 378	9.6	7.4	12.5
Occupation
Employed	1050	784 812	24.3	22.3	26.4
Unemployed	2927	2 445 528	75.7	73.6	77.7
Monthly individual income (MYR)
<1000	2519	1 851 033	57.3	53.4	61.1
1000–2000	845	682 569	21.1	18.8	23.7
>2000	567	645 096	20.0	16.9	23.5
Missing	46	51 641	1.6	0.9	2.7
Marital status
Never married	87	98 087	3.0	2.3	4.0
Married	2623	2 193 327	67.9	65.2	70.5
Widow/widower/divorcee	1264	937 454	29.0	26.4	31.8
Missing	3	1472	0.0	0.0	0.1
Strata
Urban	1689	2 362 123	73.1	69.0	76.8
Rural	2288	868 217	26.9	23.2	31.0
Ethnicity
Malay	2591	1 863 766	57.7	48.7	66.2
Chinese	710	855 542	26.5	19.8	34.5
Indian	126	209 635	6.5	4.1	10.2
Bumiputera Sabah/Sarawak	436	242 033	7.5	4.3	12.8
Others	114	59 364	1.8	1.0	3.5

%, percentage; LL, lower limit; MYR, Malaysian ringgit; UL, upper limit.

The mean scores of PADL, IADL, perceived social support, depressive symptoms and quality of life are shown in table 2. In the multiple linear regressions, PADL, IADL and social support were significant predictors for both depressive symptoms and the quality of life of older people, adjusting for other covariates (table 3). Depressive symptom scores decreased, indicating lesser depression, when functioning and social support increased. In contrast, the quality-of-life score increased with higher functioning and social support. Quality of life was found to be decreasing with higher depressive symptoms scores. All socio-demographic and economic factors were found to be significant covariates predicting depressive symptoms, except for age, marital status and strata. In contrast, only education and marital status were found to be significant covariates predicting quality of life. Each subsequent mediation analysis included the covariates that predicted the outcomes significantly.

Table 2.

Weighted means, SD and range of scores of the study variables (n=3977)

Variables	Weighted mean score	SD	Min–Max score
PADL	19.37	2.30	0–20
IADL	6.75	2.00	0–8
Social support	27.65	3.56	11–33
Depressive symptoms	2.64	2.41	0–14
Quality of life	46.76	6.93	11–57

IADL, instrumental activities of daily living; PADL, personal activities of daily living.

Table 3.

Multiple linear regressions of depressive symptoms and quality of life

Variables	Depressive symptoms			Quality of life
	Coefficients	T value	P value	Coefficients	T value	P value
	B (95% CI)			B (95% CI)
PADL	−0.109 (−0.198 to −0.021)	−2.440	0.016	0.221 (0.025 to 0.417)	2.230	0.028
IADL	−0.305 (−0.400 to −0.211)	−6.390	<0.001	0.623 (0.414 to 0.831)	5.920	<0.001
Social support	−0.175 (−0.214 to −0.136)	−8.840	<0.001	0.322 (0.204 to 0.442)	5.360	<0.001
Depressive symptoms	–	–	–	−1.173 (−1.319 to −1.027)	−15.930	<0.001
Sex
Male (ref)
Female	−0.386 (−0.600 to −0.171)	−3.570	0.001	0.544 (−0.059 to 1.147)	1.790	0.077
Age
60–69 (ref)
70–79	−0.126 (−0.346 to −0.093)	−1.140	0.258	−0.539 (−1.110 to 0.031)	−1.870	0.064
≥80	−0.242 (−0.685 to 0.201)	−1.080	0.281	−0.110 (−1.175 to 0.956)	−0.200	0.838
Omnibus test (age groups)	F (0.900)	df (2,108)	0.409	F (1.75)	df (2,108)	0.179
Education
No formal education (ref)
Primary education	−0.188 (−0.513 to 0.137)	−1.150	0.255	0.268 (−0.549 to 1.086)	0.650	0.517
Secondary education	−0.434 (−0.782 to −0.085)	−2.470	0.015	1.113 (0.288 to 1.939)	2.670	0.009
Tertiary education	−0.586 (−0.987 to −0.184)	−2.890	0.005	2.109 (0.984 to 3.236)	3.710	<0.001
Omnibus test (education level)	F (3.610)	df (3,108)	0.016	F (1.75)	df (3,108)	<0.001
Occupation
Employed (ref)
Unemployed	0.277 (0.044 to 0.510)	2.350	0.020	−0.248 (−0.839 to 0.344)	−0.830	0.408
Monthly individual income (MYR)
<1000 (ref)
1000–2000	−0.110 (−0.324 to 0.104)	−1.020	0.310	0.545 (−0.128 to 1.219)	1.6	0.112
>2000	−0.452 (−0.699 to −0.205)	−3.630	<0.001	0.771 (−0.031 to 1.574)	1.9	0.059
Omnibus test (Income level)	F (7.620)	df (2,108)	<0.001	F (2.310)	df (2,108)	0.104
Marital status
Never married (ref)
Married	−0.154 (−0.722 to 0.414)	−0.540	0.592	−1.392 (−2.567 to −0.219)	−2.35	0.021
Widow/widower/divorcee	0.085 (−0.527 to 0.699)	0.280	0.782	−1.529 (−2.788 to −0.271)	−2.41	0.018
Omnibus test (marital status)	F (1.970)	df (2,108)	0.145	F (2.950)	df (2,108)	0.050
Strata
Urban (ref)
Rural	−0.021 (−0.313 to 0.271)	−0.140	0.887	−0.261 (−1.622 to 1.099)	−0.38	0.704
Ethnicity
Malay (ref)
Chinese	−0.308 (−0.655 to 0.039)	−1.760	0.081	−0.849 (-2.294 to −0.596)	−1.16	0.247
Indian	−0.441 (−0.821 to −0.061)	−2.300	0.023	0.873 (−0.485 to 2.231)	1.27	0.205
Bumiputera Sabah/Sarawak	0.411 (0.043 to 0.779)	2.220	0.029	−0.691 (−1.839 to 0.458)	−1.19	0.236
Others	0.336 (−0.152 to 0.824)	1.370	0.175	−0.775 (−1.839 to 0.458)	−0.9	0.369
Omnibus test (ethnicity)	F (3.770)	df (4,108)	0.007	F (1.420)	df (4,108)	0.232
Constant	12.327 (10.110 to 14.543)	11.020	<0.001	32.995 (27.977 to 38.013)	13.03	<0.001
Adjusted R2	0.245			0.430

Significant relationships were highlighted in bold.

df, degree of freedom; IADL, instrumental activities of daily living; MYR, Malaysian ringgit; PADL, personal activities of daily living.

Table 4 presents the results of the mediation analysis. The four paths where social support was tested as a potential mediator were shown as paths (A)—(D). In all paths, social support was found to play a significant mediator role. For example, in the path (A) between PADL and depressive symptoms, higher PADL was associated with lower depressive symptoms (β=−0.355, p<0.001). Still, the effect was altered when social support was included in the equation, with the effect toward depressive symptoms reduced (β=−0.275, p<0.001). In this mediation path, the indirect effect of perceived social support was −0.079 (−0.109 to −0.049). In this path, the proportion mediated by social support was 22.4%. A similar pattern and direction were observed in path (B) involving IADL with depressive symptoms, with 22.1% of the effect mediated by social support. The effect of functional limitations towards depressive symptoms remained significant with the mediation factor added, indicating a partial mediator role of social support. The relationships are also demonstrated in graphic figures in online supplemental appendix B. At the same time, the details of the path coefficients between all variables, including the covariates with depressive symptoms, are reported in online supplemental appendix C.

Table 4.

Weighted coefficients, SEs and CIs for mediation analysis, testing perceived social support as the potential mediator

Paths	β	SE	LLCI	ULCI
A. Independent variable: PADL, dependent variable: depressive symptoms (n=3723)
Total effect	−0.355*	0.047	−0.448	−0.263
Direct effect	−0.275*	0.039	−0.352	−0.199
Indirect effect: ADL - social support - depressive symptoms	−0.079*	0.015	−0.109	−0.049
Standardised indirect effect	−0.056
Proportion mediated (%)	22.4
B. Independent variable: IADL, dependent variable: depressive symptoms (n=3723)
Total effect	−0.466*	0.041	−0.546	−0.387
Direct effect	−0.363*	0.035	−0.431	−0.295
Indirect effect: IADL - social support - depressive symptoms	−0.103*	0.016	−0.134	−0.072
Standardised indirect effect	−0.078
Proportion mediated (%)	22.1
C. Independent variable: PADL, dependent variable: quality of life (n=3732)
Total effect	1.138*	0.142	0.861	1.416
Direct effect	0.901*	0.116	0.673	1.129
Indirect effect: functional limitations - social support - QoL	0.238*	0.049	0.143	0.332
Standardised indirect effect	0.058
Proportion mediated (%)	20.8
D. Independent variable: IADL, dependent variable: quality of life (n=3706)
Total effect	1.521*	0.141	1.244	1.798
Direct effect	1.219*	0.117	0.989	1.449
Indirect effect: functional limitations - social support - QoL	0.301*	0.045	0.212	0.390
Standardised indirect effect	0.077
Proportion mediated (%)	19.8

Mediation analysis for paths A and B includes the covariates sex, ethnicity, income level, education level and employment status. Mediation analysis for paths C and D provides the covariates for education level and marital status.

*P value<0.001.

ADL, activities of daily living; IADL, instrumental ADL; LL, lower limit; PADL, personal ADL; QoL, quality of life; UL, upper limit; β, coefficient.

Regarding the quality of life, both PADL and IADL (paths (C) and (D)) were found to be significantly associated with quality of life before adding social support (β=1.138, p<0.001 and β=1.521, p<0.001, respectively). However, social support reduced or mediated the effects (β=0.901, p<0.001 and β=1.219, p<0.001, respectively) through the indirect effects of 0.238 (0.143–0.332) in the PADL path and 0.301 (0.212–0.390) in the IADL path. Social support proportionately mediated each path by 20.8% and 19.8%, respectively. Similarly, the relationship between PADL and IADL with quality of life remained significant, with the mediation effect added into the equation, signalling the partial mediator role of social support. The relationships are also demonstrated in graphic figures in online supplemental appendix D. At the same time, the details of the path coefficients between all variables, including the covariates with quality of life, are reported in online supplemental appendix E.

The results of the serial mediation analysis are shown in table 5. In the first analysis, the relationship between PADL and quality of life was tested (path (A)), with social support and depressive symptoms entered as multiple serial mediators into the path. The relationship between PADL and quality of life was significant (β=1.134, p<0.001) before adding the multiple mediations. The effect was brought lower (mediated) when social support and depressive symptoms were added into the equation (β=0.536, p<0.001). This coefficient value was lower than the previous single mediator paths, indicating that multiple mediations (social support-depressive symptoms) had more prominent roles in explaining the quality of life of older people than social support alone. The three indirect effects were found to be significant, indicating that both social support and depressive symptoms mediated the relationship between PADL and quality of life and social support affected depressive symptoms, which eventually affected the quality of life in the multiple mediations. The proportion mediated through the indirect effect was found to be 52.7%. The relationship between PADL and quality of life remained significant after adding multiple mediators into the equation, indicating partial mediation roles by social support and depressive symptoms. Similar findings were observed for IADL, as shown in path (B) in table 5. The proportion mediated through the indirect effect was found to be 49.6%. The graphical representation is in online supplemental appendix F, while the path coefficients among all variables are reported in online supplemental appendix G.

Table 5.

Weighted coefficients, SEs and CIs in the serial mediation analysis

Paths	β	SE	LLCI	ULCI
Independent variable: PADL, dependent variable: quality of life (n=3688)
Total effect	1.134*	0.141	0.858	1.410
Direct effect	0.536*	0.098	0.344	0.727
Total indirect effects	0.598*	0.095	0.413	0.783
Indirect effect 1: functional limitations - social support - QoL	0.145*	0.032	0.082	0.209
Indirect effect 2: functional limitations - depressive symptoms - QoL	0.359*	0.061	0.242	0.476
Indirect effect 3: functional limitations - social support - depressive symptoms - QoL	0.093*	0.021	0.052	0.135
Proportion mediated (%)	52.7
Independent variable: IADL, dependent variable: quality of life (n=3690)
Total effect	1.506*	0.144	1.225	1.787
Direct effect	0.759*	0.094	0.576	0.943
Total indirect effects	0.747*	0.084	0.582	0.911
Indirect effect 1: functional limitations - social support - QoL	0.190*	0.034	0.123	0.256
Indirect effect 2: functional limitations - depressive symptoms - QoL	0.445*	0.053	0.342	0.548
Indirect effect 3: functional limitations - social support - depressive symptoms - QoL	0.112*	0.019	0.075	0.149
Proportion mediated (%)	49.6

CIs for indirect effects were bootstrapped with 5000 replications. The standardised indirect effects were found to be similar to bootstrapped values.

Mediation analysis includes covariates sex, ethnicity, income level, education level, marital status and employment status.

*P value<0.001.

IADL, instrumental activities of daily living ; LL, lower limit; PADL, personal activities of daily living ; QoL, quality of life; SE, robust SE; UL, upper limit; β, coefficient.

Replicating the mediation models without adjusting for covariates yielded a consistently significant result. The correlation coefficients necessary to nullify the indirect effect across all paths range between −0.2 and 0.4 (online supplemental appendix H).

Discussion

Examining the complex relationship between various health factors through a mediation analysis revealed social support’s crucial role in influencing the well-being of community-dwelling older people. The presence of social support mediated the relationship between functional limitations and depressive symptoms, with social support indirectly accounting for reduced depressive symptoms. Similarly, social support also mediated the relationship between functional limitations and quality of life, indirectly improving the quality of life in the relationship. The serial mediation analysis found that social support affects depressive symptoms, influencing the quality of life. This finding suggests a serial mediation relationship between functional limitations and quality of life.

Higher functional limitations were significantly associated with lower depressive symptoms and better quality of life, confirming the findings in various studies.^18–21 The presence of social support explained the relationship. This finding reciprocated evidence from other studies.34 35 45 63 64 Earlier work in 1996 found that lower reported social support was an essential reason for decreases in life satisfaction and increases in depressive symptoms among people in the USA.³⁴ In a Norwegian study, perceived social support was a significant mediator between functional limitations and depression among those aged 45 and above.⁴⁵ Similarly, a study among older people in China found that subjective support and utilisation of social support partially mediated the relationship between physical disability and depressive symptoms. In contrast, objective social support played no significant role.³⁵ Another study among older women in the USA found no significant role played by perceived social support in explaining depression.⁶⁴ It is important to note that all studies listed used different scales to measure all variables of interest and had diverse target populations and analytical methods in testing the mediation role, which may affect the similarities or differences observed from the current study.

Despite contrasting findings, the evidence thus far points towards social support explaining or being involved in the path of the relationship between depressive symptoms and quality of life among older people rather than merely a risk factor. As a person ages, many stressful life changes emerge, such as losing a spouse, retirement, relocations and changes in daily routines.³² Stressors of this type are hypothesised to directly affect the risk and severity of illnesses, especially among those with functional limitations.⁶⁵ It has been proposed that, from a social aspect, having emotionally significant social relationships may be required to elicit the resilience necessary for successful adjustment to some of the stressors associated with ageing.⁶⁶ Biopsychosocial models of depressive symptoms and quality of life explain that disintegration in social support leads to social loneliness and unfulfillment of various needs of older people, making them vulnerable to declined mental health and life satisfaction.^67–69 Furthermore, in separate path analyses, this study explored two distinct categories of functional limitations in older adults: PADL and IADL. PADL encompasses essential self-care tasks, while IADL involves more complex activities supporting independent living. The research underscores the pivotal role of social support as a mediator for depressive symptoms and quality of life across both PADL and IADL domains. It emphasises the significance of integrating social support interventions into caregiving strategies, recognising their potential to mitigate functional limitations’ impact on older people’s well-being.

The implications of social support as a mediator are far-reaching. By identifying social support as a third variable in the relationship between functional limitations, depressive symptoms and quality of life, it becomes apparent that prevention and management activities should be extended beyond health services alone. This understanding allows for developing comprehensive strategies encompassing broader social interventions, ultimately leading to more optimal outcomes in later life.³² In the current study, social support partially mediates, indicating that functional limitations were associated with depressive symptoms and quality of life independently and must be addressed vigorously. Nevertheless, the mediation analysis findings show that improving social support is crucial. In particular, the social aspect being tested in the current study revolved around perceived social support, requiring an understanding of its components to address the issue accurately. Perceived social support refers to the satisfaction of feeling understood, respected and supported by a person’s social network.³⁷ It does not involve the provision of aid, financial assistance or any other forms of tangible support. In fact, the economic factor was controlled as one of the potential confounders, which may be interpreted as regardless of financial background and other socio-demographic factors, variations in perceived social support played a role in influencing depressive symptoms and quality of life.

The prospects of improving the population’s health have shifted from improving healthcare alone to addressing the major social determinants of health, including accommodating better social support.⁷⁰ The more important questions to be addressed are: who is responsible for providing adequate social support, and what measures can be taken to improve the delivery of social support? In the context of the current study, whereby Malaysia is part of Asian countries, the responsibility of caring for and supporting the older people often rests on their children, although becoming more challenging due to urbanisation, the growing demands of the workforce and the relocation of children away from parents.⁷¹ Nevertheless, over 70% of older Malaysians were found to be living with family and other members of their home, 21% were living with their spouse and only 9% were living alone.⁴⁰ Some strategies to strengthen family support towards older people include financial incentives and tax reliefs to families who continue caring for their older relatives.⁷² Strengthening social activities for older people living in the community, such as social networking and educational classes held in community centres, has been associated with good outcomes among older people.^{73 74}

In the Malaysian context, various efforts have been put in place to improve social support for older people. For example, the Ministry of Women, Family and Community Development and Ministry of Health introduced various community-based programmes for older people, such as group physical activities, social gatherings and the establishment of committees overseeing and addressing older people’s needs in the community.⁷⁵ However, community-based services must be made widely accessible and affordable for older people for a successful outcome. In Malaysia, community-based initiatives, such as senior citizens’ activity centres, were gradually introduced in phases beginning in 2002.⁷⁶ The coverage has expanded over the years, but the implementation has faced challenges in coordinating activities across the social and health sectors.^{76 77}

A meta-analysis analysing social care services, which are defined as care that supports persons’ day-to-day needs delivered in their homes, found that the service led to lower hospital readmission rates and length of stay.⁷⁸ This profound evidence encapsulates the ability of persons, families, communities and governing bodies to enhance the delivery of social support for older people, ideally through an integrated approach that offers comprehensive support. In Malaysia, incentives and aid were available, including meals at home and monetary assistance to support the well-being of older people in the community.^{75 79} However, the perceived social support remains low. In 2018, 30.8% of Malaysians above 60 perceived having poor social support,¹⁷ highlighting the need for a more integrated approach to address the issue. A strategic consideration involves adopting a social prescribing service. In this paradigm, the health sector seamlessly integrates referrals for community-based social activities into the treatment plans for older people.^{80 81} To effectively implement this initiative in Malaysia, a comprehensive approach is essential, beginning with the mapping and garnering support from various facilities that offer relevant services. This entails identifying and engaging stakeholders across different sectors to ensure alignment and collaboration. Another strategic move involves integrating assessments for the level of social support into the healthcare system’s protocols for addressing the health concerns of older people. Specifically, healthcare providers should be equipped to identify older people experiencing depression and low quality of life and assess their level of social support. This may involve implementing standardised screening tools and training healthcare professionals to recognise indicators of inadequate social support. To ensure effective intervention, seamless referral pathways should be established, directing older people with insufficient social support to appropriate resources such as social welfare departments or community support services. This necessitates raising awareness and training healthcare and social service providers about the importance of social support in improving the well-being of older people.

Depressive symptoms were also significantly associated with older people’s quality of life. The serial mediation analysis found that the multiple mediations by social support-depressive symptoms explain part of the quality of life among those with functional limitations. The regression analysis showed that social support was significantly associated with depressive symptoms, explaining the relationship. This relationship may be attributed to the positive influence of social support, as it enables older persons to engage more actively in health-promoting activities and access various healthcare or health-related resources.⁸² Social support also provides coping mechanisms among older people in dealing with health or non-health-related issues, including depression.⁸³ Depression, conversely, excludes the experience of positive well-being and negatively influences a person’s living conditions, thus resulting in poorer quality of life.⁸⁴ This relationship highlights the importance of adopting a holistic and integrated approach to improve the quality of life of older people. Vigorous management of functional limitations should be accompanied by efforts to strengthen social support and manage depressive symptoms to enhance older people’s overall quality of life.

NHMS 2018 was conducted before the onset of the COVID-19 pandemic. Subsequent Malaysian studies have revealed the profound impact of the pandemic on the mental health and quality of life of Malaysians, mainly attributed to factors such as social isolation and economic instability.^85–87 Although the role of social support as a mediator during this period in Malaysia remains unexplored, existing research underscores its crucial role in preserving mental well-being and overall quality of life, particularly among older people.^{88 89} A Malaysian study showed that older people with good social networks during the pandemic had better psychosocial outcomes.⁹⁰ Thus, it is postulated that the pandemic would have amplified the significance of social support as a mediator in mitigating these adverse effects.

While the NHMS 2018 is a large and nationally representative survey, it only captures older people living in the community, thus excluding older people residing in care facilities due to more severe functional limitations or depressive symptoms. Second, due to the study’s cross-sectional nature, causal establishment between relevant variables cannot be determined. The functional limitations, social support, depression and quality of life measurement tools were rated based on respondents’ perceptions and interpretations, thus subject to personal bias. Clinical assessment of depression and functional limitations and an observational or qualitative study to explore social support and quality of life may offer more accurate measurements. This study only analysed the perceived social support among the participants while acknowledging that various other forms of social support exist. Further research incorporating different forms of social support may offer a more comprehensive overview. Finally, our study was based on Malaysia’s unique cultural practice and healthcare systems, where countries with similar backgrounds may relate to the outcomes. Regardless, our analysis is one of the few examining the mediation analysis between functional limitations with two outcomes, depressive symptoms and quality of life, with additional serial mediation analysis.

Conclusion

This study reveals that perceived social support alleviates the relationship between functional limitations and depressive symptoms, as well as quality of life among older people. These findings provide valuable insights for designing and developing preventive measures and healthcare interventions to address depressive symptoms and improve quality of life. Consequently, promoting social support in the community-dwelling older population becomes vital for enhancing their well-being. Furthermore, by exploring the serial mediation roles of social support-depressive symptoms towards the quality of life, the study contributes to a better understanding of the complex interactions between various factors in influencing older people’s well-being. Enhancing social support for older people is vital for safeguarding their mental health and overall well-being. This involves expanding and formally integrating social activities into health and social sectors and increasing awareness for informal support within families and caregivers. Assessment of social support levels during regular medical visits and prescribing social activities may help older people feel more supported. Health and social care providers can improve their well-being and sense of belonging by understanding their social connections and suggesting activities they enjoy.

Data availability statement

Data are available upon reasonable request. The data set analysed for this article is part of the National Health and Morbidity Survey 2018: Elderly Health study and belongs to the Ministry of Health Malaysia. Requests for the data can be obtained from the Sector for Biostatistics & Data Repository, National Institute of Health, Ministry of Health Malaysia, accessible at https://nihdars.nih.gov.my/, with permission from the Director-General of Health, Malaysia.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

All participants provided written informed consent before interviews during the NHMS 2018’s data collection. The NHMS survey protocol was approved by the Medical Research and Ethics Committee (MREC), Ministry of Health Malaysia (NMRR-17-2655-39047). The study was conducted by Good Clinical Practice guidelines and the Declaration of Helsinki.