Gender differences in the impact of emotional support networks on health-related quality of life among floating elderly: a cross-sectional study in China

Abstract

Background

Emotional support, as a key component of social support, may influence the health of floating elderly individuals. This study aims to examine the characteristics of emotional support networks and assess their impact on health-related quality of life (HRQoL) among the floating elderly, with a focus on gender differences.

Methods

Data were collected through questionnaires from 2,330 floating elderly in Beijing and Nanjing, China. HRQoL was measured using the EuroQol 5-Dimensions 3-Level scale, while emotional support network characteristics were assessed in terms of size, density, composition, heterogeneity, and convergence. Tobit regression models were employed to analyze the impact of emotional support networks on HRQoL.

Results

The mean HRQoL of the participants was 0.884 ± 0.138, with females reporting higher utility values than males (P < 0.05). Emotional support network size was negatively associated with HRQoL (P < 0.01), whereas a larger number of kin members and greater age convergence had positive effects (P < 0.01). Compared to males, higher network density (β=-0.073, P < 0.05) and greater educational heterogeneity (β=-0.116, P < 0.01) were associated with lower HRQoL of female elderly.

Conclusions

The overall HRQoL of floating elderly individuals was relatively good. Their emotional support networks were generally “small in size and high in density”, with greater convergence than heterogeneity. A large emotional support network may not be necessary, and priority attention should be given to those lacking kin-based emotional support. It is also crucial to emphasize the role of peers and consider gender differences when designing emotional support networks for the floating elderly.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-06406-6.

Introduction

The floating elderly population has become a growing concern worldwide. In 2018, the World Health Organization Regional Office for Europe highlighted in its report on the Health Status of Older Refugees and Migrants that both aging and migration are complex multidimensional processes, calling for the full integration of the needs of older refugees and migrants into Europe’s aging policies and practices [1]. In China, the issue is equally significant. According to the 2018 China Floating Population Development Report, the number of floating elderly individuals surged from 5.03 million in 2000 to 13.04 million in 2015 [2]. With the continued implementation of China’s urbanization strategy, the persistence of intergenerational support, and the growing elderly population [3], the number of floating elderly in China is expected to keep rising, highlighting the need for increased attention to this group.

Health-related quality of life (HRQoL), as an important indicator of individual health, has received increasing attention from scholars. HRQoL encompasses self-assessments of physical, emotional, and social well-being [4]. It has been observed that HRQoL in older populations is influenced by several factors, including age, gender, marital status, and socioeconomic status [5–7]. In addition, the positive effects of social support on the health of older adults have been widely acknowledged [8–10], with evidence suggesting that adequate social support can reduce stress and enhance HRQoL among the elderly [11, 12]. Floating elderly individuals, who experience the dual challenges of aging and migration, face significant barriers to social integration [13, 14]. It is necessary to focus on their social support and understand how this influences their health outcomes.

Several studies have highlighted the significant impact of social support on the health of floating elderly. Interactions with residents in the inflow area, for example, have been shown to improve their health [15], and adequate social support can enhance their well-being while increasing health services utilization [16]. These findings highlight social support as a critical factor in understanding the health of the floating elderly. However, two key research gaps remain. On one hand, the specific categorization of social support has not been fully explored. Most studies tend to classify social support as formal vs. informal support or subjective vs. objective [17, 18]. For instance, Xu et al. [19] assessed the impact of using National Basic Public Health Services, a formal social support, on the health outcomes of floating elderly. Other categorizations, such as van der Poel (20) dividing social support into emotional support, instrumental support, and social companionship, or the inclusion of informational and appraisal support [21], have also been proposed. Despite these alternatives, limited research has examined how these specific types of social support affect the health of the floating elderly.

On the other hand, there is limited research on the specific characteristics of social support networks among the floating elderly. Unlike studies that classify social support into emotional, instrumental, and other types, social support networks focus more on the structural characteristics of the social support relationships, emphasizing how individuals access support through their network connections [22]. For example, Wellman and Wortley [23] found that kinship ties vary in their importance and function in providing emotional support, practical assistance, financial aid, and companionship. Granovetter’s theory of “weak ties” also suggests that while weak ties in the social support network, such as casual acquaintances, may offer limited emotional support, they often provide greater instrumental value [24]. In the Chinese context, Danching Ruan was among the first scholars to systematically examine social support networks. Her research on residents in Tianjin [25], later expanded to Beijing and Hong Kong [26], revealed that spouses or partners play a prominent but diffuse role in social support. Other relatives were more involved in providing instrumental support, whereas non-relatives primarily offered emotional support. In addition, Zhou and Kang [27] revealed that informal care networks among functionally impaired older adults exhibit diverse structures and are shaped by a combination of individual, family, and policy factors.

Furthermore, it has been emphasized that the size, density, and types of relationships within social support networks can directly affect the resources available to individuals [28, 29], subsequently impacting health outcomes [30–32]. According to social capital theory, the structure and quality of social networks are closely related to an individual’s level of health [33]. For example, Ali et al. [34] found that, compared to those in large, low-strain networks, older adults in small, restricted, high-contact networks had fewer disabilities but worse mobility, while those in large, high-strain networks also exhibited poorer mobility. In South Korea, studies have shown that diverse or family-based social networks are positively associated with better subjective mental health among older adults [35, 36]. Using data from Alameda County, California, Berkman and Syme [37] found that individuals lacking social ties had a significantly higher risk of mortality during the follow-up period compared to those with close social connections. Additional findings indicated that for individuals under 60, marital status played the most crucial role in health, whereas for those aged 60 and older, close ties with friends or relatives were more protective [38]. Similarly, studies in China have found that neighbor and friend networks had stronger effects on older adults’ health than family-based networks [39, 40]. However, to the best of our knowledge, there is a lack of research on the specific characteristics of social support networks among the floating elderly and their impact on health.

In this study, emotional support networks within social support networks and their impact on the health of floating older adults were focused. As an important form of social support, emotional support refers to the psychological resources that individuals gain from emotional comfort, understanding, and care, which plays a critical role in relieving stress and promoting well-being [22, 41]. Compared to informational, financial, or instrumental support, emotional support is particularly vital for floating older adults, who are often disconnected from familiar social environments and tend to rely more heavily on rebuilding emotional support networks in new settings.

In addition, gender differences were considered in the analytical framework for several reasons. First, men and women differ significantly in both their access to emotional support and the structure of their support networks. Some research suggests that women tend to maintain more emotionally expressive and close-knit social networks, while men are more likely to have less emotionally intimate connections, which may lead to a greater risk of emotional isolation in later life [42, 43]. Second, gender is closely associated with health outcomes. Older women may be more vulnerable to mental health challenges, while older men tend to be more affected by functional decline, especially in the context of reduced social interaction [44–47]. Third, some studies have documented gender differences in the relationship between social support and health [48]. For example, a Japanese study found that social support may promote subjective health more strongly in older men than in women [49]. These gender-based differences are therefore critical for understanding the heterogeneous effects of emotional support on health.

Based on the above, limited research has explored the characteristics of emotional support networks among floating elderly, let alone analyzing their relationship with the health-related quality of life. To address this issue, this study aims to: (1) identify the HRQoL and emotional support network characteristics of floating elderly in China; (2) explore the impact of their emotional support networks on HRQoL; and (3) investigate the gender differences in this impact. The findings of this study can provide valuable insights into the construction of emotional support networks for floating elderly, and ultimately improving their health utility.

Methods

Data source and participants

Beijing and Nanjing, China’s first-tier and new first-tier cities, are major destinations for the inflow of elderly people. Therefore, these two cities were selected for the study, and face-to-face surveys were conducted between June and October 2021 among the floating elderly in both cities. The survey aimed to gather basic information as well as details about the emotional support networks of the respondents. Written informed consent was provided in the introductory section of the questionnaire, and consent was obtained from all participants.

A stratified random sampling method was used for sample selection. Based on the statistical yearbook data and the distribution of the elderly, 8 districts/counties were randomly selected from Beijing and 4 from Nanjing. Subsequently, 5 communities were randomly chosen from each district/county, and 30 non-local residents (60 in Nanjing) aged 60 and above were randomly selected from each community to participate in this survey. The sample size required for the study was calculated using formula (1).

1

where z determines the confidence level, generally 1.96, which corresponds to the 95% confidence level. P is the percentage of a characteristic in the target population and is conservatively set at 0.5. d is the acceptable precision level, generally, we take 0.03. Using this formula, 1,067 samples were required for the study. To account for potential exclusions due to quality issues, the sample size was expanded to 1,200 for each city. In total, 2,448 floating elderly individuals were surveyed, including 1,248 in Beijing and 1,200 in Nanjing. After excluding samples with missing key variables, 2,330 valid samples remained, resulting in an effective response rate of 95.18%.

Measurement of HRQoL

EuroQol 5-Dimensions 3-Level (EQ-5D-3L) questionnaire was used to measure the HRQoL of the floating elderly. The EQ-5D-3L consists of five health dimensions: Mobility, Self-care, Usual activities, Pain/Discomfort, and Anxiety/Depression. Each dimension has three response levels: 1 = no problems, 2 = some/moderate problems, and 3 = extreme problems [50, 51]. HRQoL was calculated using the Chinese version of the EQ-5D-3L utility value system, which was developed by Liu et al. [52] through a time trade-off model (TTO) for the general Chinese population. Considering all possible health states, the corresponding utility scores range from − 0.149 (representing the worst health state: 33333) to 1.000 (representing perfect health: 11111), with higher scores indicating higher health utility. A detailed description of the HRQoL score construction and computational process is in Text S1 and Table S1 of the supplementary material.

Measurement of emotional support network

Similar to previous studies [20, 53], this study identified emotional support network members based on respondents’ answers to the question, “Who is usually with you when you need someone to talk to or chat with?” The emotional support network characteristics of the participants were measured in four dimensions: network size and density, network composition, network heterogeneity, and network convergence.

Network size refers to the total number of individuals who provide emotional support for the floating elderly and is calculated by summing the number of network members. Network density refers to the extent to which members of a network are interconnected, typically measured by the ratio of closely related member pairs to the total number of possible pairs. The density of a network is zero if all its members are connected only to the central individual, and 100% if all members are closely connected. Network density is calculated using the formula (2), where x is the number of pairs of members with close relationships in the network and n is the network size.

2

Network composition refers to the specific relationships between the floating elderly and their emotional support network members, categorized into kinship and non-kinship relationships. Kinship includes family members such as parents, spouses, sons, daughters, daughters-in-law, sons-in-law, grandchildren, siblings, and other relatives. Non-kinship includes individuals such as neighbors, caregivers, fellow townsmen, classmates, co-workers, friends, and others.

Network heterogeneity refers to the variation in demographic characteristics, such as age, gender, and educational level, among all the network members [54]. This indicator represents the probability that two people, randomly selected from one emotional support network, do not belong to the same category in some regard. For example, gender heterogeneity refers to the likelihood that a random selection of two individuals from a network will result in one being male and the other female. If the gender distribution within an emotional support network is homogeneous, it would exhibit 50% gender heterogeneity. In contrast, if the network is composed entirely of the same sex, then gender heterogeneity is zero. The formula for network heterogeneity is (3).

3

where P is the proportion of individuals with a certain characteristic in group i (i = 1, 2, …, k) relative to the overall total, and i is the number of groups the population can be divided into based on the characteristic. Age heterogeneity is measured by the standard deviation (SD) of age among network members. A larger SD indicates stronger age heterogeneity within the network.

Network convergence refers to the similarity between respondents and the members of their emotional support network concerning certain characteristics. It is measured by the percentage of network members who share the same characteristic category as the floating elderly. For example, if a female floating elderly person receives emotional support from two females and two males, the gender convergence of her emotional support network would be 50%. Age convergence refers to the percentage of network members whose age difference from the respondent is no greater than five years.

It is important to note that the convergence and heterogeneity of an emotional support network can only be calculated for respondents with a network size greater than 1. Therefore, floating elderly individuals with a network size of fewer than 2 were excluded from the calculations of heterogeneity and convergence.

Control variables

Based on previous studies [55–57], the control variables in this study included three aspects: demographic characteristics, floating characteristics, and health status of the floating elderly. Demographic characteristics include gender, age, educational level, marital status, Hukou status, economic status, and whether the floating elderly were enrolled in health insurance. Floating characteristics included the scope, duration, and reasons for floating. Health status was assessed based on whether the respondent had hypertension or diabetes, and whether they had experienced illness in the past year. The categorization and assignments of each variable can be found in Table S2 in the supplementary material.

Statistical analysis

Stata/SE 16.0 (Stata Corp, College Station, TX, USA) was used for data analysis. Descriptive statistical analysis was first performed to describe the characteristics of the participants. Continuous and categorical variables were presented as Mean ± SD or frequency (percentage), respectively. The Kolmogorov-Smirnov test indicated that neither HRQoL nor the network characteristics followed a normal distribution (Supplementary Figure S1, Table S3). Thus Manual-Whitney U test and Kruskal-Wallis H test were used to compare HRQoL among individuals with different characteristics.

Regression models were then constructed to examine the relationship between emotional support networks and HRQoL. Given the strong ceiling effect of the EQ-5D scale, regression methods that fail to account for this effect may produce biased coefficient estimates. The Tobit model, which performs more reliably than OLS in such cases [58], was therefore employed in our analysis. Finally, gender differences in the impact of emotional support networks on HRQoL were examined through separate regressions for male and female participants. A two-tailed P-value of less than 0.05 was considered statistically significant.

Results

Characteristics of the participants

Among the 2,330 participants (Table 1), ages ranged from 66 to 89 years, with a mean age of 66.9 years. Of these, 1,220 (52.36%) were male and 1,110 (47.64%) were female. About 38.28% had completed high school or higher education. Additionally, 70.77% lived with their spouses, and 59.40% registered with an agricultural Hukou. Most participants (73.35%) reported being financially stable or wealthy, and 92.83% were enrolled in health insurance.

Table 1.

Characteristics and HRQoL of the participants

Characteristics		N (%)	HRQoL	χ2/Z	P-value
Gender	Male	1220(52.36)	0.879 ± 0.136	−2.465	0.014
	Female	1110(47.64)	0.890 ± 0.140
Age in 2021	60–69 years	1726(74.08)	0.911 ± 0.115	238.150	< 0.001
	70–79 years	540(23.18)	0.814 ± 0.167
	80 years and above	64(2.75)	0.768 ± 0.160
Educational level	Uneducated	166(7.12)	0.850 ± 0.152	8.355	0.015
	Primary or middle school	1272(54.59)	0.885 ± 0.145
	High school or above	892(38.28)	0.889 ± 0.122
Marital status	No spouse	391(16.78)	0.852 ± 0.159	63.460	< 0.001
	Living with spouse	1649(70.77)	0.884 ± 0.134
	Living apart from spouse	290(12.45)	0.928 ± 0.117
Hukou status	Agricultural Hukou	1384(59.40)	0.889 ± 0.135	8.237	0.016
	Non-Agricultural Hukou	941(40.39)	0.877 ± 0.142
	Other Hukou	5(0.21)	0.954 ± 0.103
Economic status	Very poor	211(9.06)	0.867 ± 0.146	7.140	0.129
	Slightly poor	410(17.60)	0.880 ± 0.159
	Break-even	653(28.03)	0.892 ± 0.133
	Slightly rich	730(31.33)	0.882 ± 0.130
	Very rich	326(13.99)	0.890 ± 0.128
Health insurance	No	167(7.17)	0.891 ± 0.143	0.862	0.392
	Yes	2163(92.83)	0.884 ± 0.137
Floating scope	Cross-provincial mobility	1692(72.62)	0.878 ± 0.140	14.599	< 0.001
	Cross-municipal mobility	621(26.65)	0.899 ± 0.131
	Cross-county mobility	17(0.73)	0.946 ± 0.066
Floating duration	6 months and less	221(9.48)	0.891 ± 0.129	0.492	0.782
	6 months but less than 12 months	638(27.38)	0.878 ± 0.147
	12 months and more	1471(63.13)	0.886 ± 0.135
Floating reasons	Work or do business	701(30.09)	0.906 ± 0.119	105.632	< 0.001
	Follow the family	1192(51.16)	0.894 ± 0.131
	Retire	338(14.51)	0.806 ± 0.172
	Other	99(4.25)	0.878 ± 0.121
Hypertension	No	1504(64.55)	0.909 ± 0.126	12.626	< 0.001
	Yes	826(35.45)	0.840 ± 0.147
Diabetes	No	1853(79.53)	0.899 ± 0.128	10.409	< 0.001
	Yes	477(20.47)	0.827 ± 0.158
Whether illness within one year	No	1452(62.32)	0.913 ± 0.118	13.519	< 0.001
	Yes	878(37.68)	0.837 ± 0.154

Regarding floating characteristics, most floating elderly (72.62%) moved across provinces, while only 0.73% floated between counties. 63.13% of the participants had been floating for one year or more. The primary reason for floating was following family members (51.16%), followed by work or business (30.09%). In terms of health, 35.45% of the floating elderly had hypertension, 20.47% had diabetes, and 37.68% had experienced illness in the past year.

EQ-5D-3L distribution of the participants

The mean HRQoL of the participants was 0.884 ± 0.138, with a median of 0.869. Health utility values varied by gender, age, education level, marital status, Hukou status, floating scope, floating reasons, and health status (P < 0.05). As shown in Fig. 1, among the EQ-5D dimensions, pain/discomfort was the most common issue among the floating elderly, with 46.87% reporting moderate or severe problems, followed by anxiety/depression (19.36%). Most of the participants had no difficulties with mobility (87.04%), usual activities (92.02%), or self-care (94.03%).

Fig. 1.

Distribution of EQ-5D dimensions among the floating elderly

Characteristics of emotional support networks among the participants

Network size and density

Table 2 presents the distribution of emotional support network sizes among the participants. Most floating elderly had a network of 2 to 3 members, with an average size of 2.58. 41.24% had a network of three, while 13 participants reported having no emotional support. In terms of network density, the emotional support network for floating elderly exhibited a density greater than 90%. Male floating elderly had a larger emotional support network than females (2.72 vs. 2.43, P < 0.001).

Table 2.

The emotional support network size among the floating elderly

Network size	Total	Male	Female	Z	P-value
0	13(0.56)	6(0.49)	7(0.63)	6.903	< 0.001
1	318(13.65)	143(11.72)	175(15.77)
2	702(30.13)	325(26.64)	377(33.96)
3	961(41.24)	513(42.05)	448(40.36)
4	265(11.37)	178(14.59)	87(7.84)
5	71(3.05)	55(4.51)	16(1.44)
Mean of network size	2.58	2.72	2.43
Mean of network density	91.33	91.66	90.97	−0.786	0.432

Network composition

Table 3 shows the composition of the emotional support network among the floating elderly, with only 3.09% of the participants having no kin members in their network. The emotional support network of males had an average of 2.24 kin members and 0.11 non-kin members, while female elderly had an average of 2.03 kin members and 0.07 non-kin members (P < 0.05). The specific members that make up the emotional support network among floating elderly are shown in Supplementary Table S4.

Table 3.

The composition of the emotional support network among the floating elderly

Type	Numbers	Total	Male	Female	Z	P-value
Kinship	0	72(3.09)	26(2.13)	46(4.14)	4.677	< 0.001
	1	545(23.39)	263(21.56)	282(25.41)
	2	882(37.85)	455(37.30)	427(38.47)
	3	658(28.24)	354(29.02)	304(27.39)
	> 3	173(7.43)	122(10.00)	51(4.59)
	Mean	2.14	2.24	2.03
Non-kinship	0	1960(84.12)	1008(82.62)	952(85.77)	3.256	0.001
	1	345(14.81)	201(16.48)	144(12.97)
	2	23(0.99)	10(0.82)	13(1.17)
	3	2(0.09)	1(0.08)	1(0.09)
	> 3	0(0.00)	0(0.00)	0(0.00)
	Mean	0.17	0.11	0.07

Network heterogeneity and convergence

Table 4 shows the heterogeneity and convergence of emotional support networks. Male floating elderly exhibited higher gender and educational heterogeneity in their networks than females (P < 0.05). The emotional support networks of male floating elderly had a gender convergence of 44.18%, whereas the rate for female elderly was 37.19%, with the difference being statistically significant (P < 0.05).

Table 4.

The heterogeneity and convergence of emotional support networks among the floating elderly (%)

Characteristics		Total (n = 1999)		Male (n = 1071)		Female (n = 928)		Z	P-value
		Mean	SD	Mean	SD	Mean	SD
Network heterogeneity	Gender	34.08	20.38	36.23	18.91	31.60	21.70	3.258	0.001
	Age	14.80	3.57	14.73	3.64	14.89	3.49	−1.283	0.200
	Educational	26.00	21.98	27.88	21.85	23.81	21.84	3.938	0.000
Network convergence	Gender	40.94	26.72	44.18	25.59	37.19	27.51	5.615	0.000
	Age	28.46	23.10	28.44	22.95	28.47	23.28	−0.545	0.586
	Educational	49.46	36.67	48.84	35.40	50.22	38.10	−0.625	0.532

Association between emotional support networks and HRQoL

Table 5 demonstrates the relationship between the emotional support networks of male and female participants and their health-related quality of life. After adjusting for control variables, the size of the emotional support network was negatively associated with HRQoL in both male and female floating elderly (P < 0.01). The number of kinship members and age convergence within the network were positively associated with HRQoL (P < 0.01). For female floating elderly, higher network density was associated with poorer HRQoL (β=−0.073, P < 0.05), while greater educational heterogeneity among network members negatively affected their HRQoL (β=−0.116, P < 0.01).

Table 5.

Association between emotional support networks and HRQoL among male and female floating elderly

Characteristics	Total		Male		Female
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Network size	−0.064*	−0.050*	−0.070*	−0.054*	−0.050*	−0.035*
	(−0.083,−0.045)	(−0.068,−0.032)	(−0.095,−0.046)	(−0.078,−0.031)	(−0.082,−0.019)	(−0.064,−0.005)
Number of kinship members	0.061*	0.045*	0.049*	0.037*	0.077*	0.057*
	(0.041,0.0810)	(0.026,0.064)	(0.024,0.075)	(0.013,0.061)	(0.044,0.109)	(0.027,0.087)
Network density	−0.046*	−0.047*	0.003	−0.024	−0.100*	−0.073*
	(−0.090,−0.003)	(−0.089,−0.006)	(−0.054,0.060)	(−0.078,0.031)	(−0.168,−0.032)	(−0.137,−0.010)
Gender heterogeneity	−0.045	−0.051	−0.059	−0.043	−0.065	−0.083
	(−0.109,0.019)	(−0.111,0.008)	(−0.145,0.026)	(−0.123,0.038)	(−0.165,0.034)	(−0.174,0.008)
Age heterogeneity	0.002	0.001	0.002	0.002	0.001	0.001
	(−0.002,0.005)	(−0.002,0.004)	(−0.002,0.006)	(−0.002,0.006)	(−0.004,0.007)	(−0.004,0.006)
Educational heterogeneity	−0.081*	−0.068*	−0.025	−0.025	−0.145*	−0.116*
	(−0.134,−0.028)	(−0.119,−0.016)	(−0.094,0.044)	(−0.093,0.043)	(−0.228,−0.063)	(−0.196,−0.037)
Gender convergence	0.022	0.035	−0.015	0.028	0.061	0.058
	(−0.020,0.065)	(−0.004,0.074)	(−0.070,0.040)	(−0.024,0.080)	(−0.007,0.129)	(−0.007,0.122)
Age convergence	0.174*	0.111*	0.186*	0.112*	0.165*	0.126*
	(0.123,0.225)	(0.060,0.161)	(0.120,0.253)	(0.045,0.179)	(0.086,0.245)	(0.048,0.204)
Educational convergence	−0.026	0.007	−0.012	0.009	−0.043	−0.006
	(−0.055,0.002)	(−0.039,0.053)	(−0.050,0.026)	(−0.051,0.069)	(−0.086,0.000)	(−0.077,0.066)
Controls	No	Yes	No	Yes	No	Yes

(1) Model 1 was unadjusted, and Model 2 was controlled for the demographics, floating characteristics, and health status of the floating elderly. (2) Clustered standard errors in parentheses. (3) *P < 0.05

Discussion

Using a quantitative survey, this study examined the characteristics of emotional support networks among the floating elderly and the impact on their health-related quality of life. To our knowledge, this is the first to apply a nomothetic approach to measure emotional support networks in this population, identifying key structural characteristics such as network size, density, composition, heterogeneity, and convergence. More importantly, this study highlights that the structural configuration of emotional support networks is a critical factor influencing HRQoL. Although the importance of emotional support for the health of the floating elderly has been recognized, existing research often overlooks the structural characteristics of support networks. By shifting the focus from the general availability of support to its network structure, and by further incorporating gender-specific analysis, this study offers a more nuanced understanding of how emotional support affects HRQoL among this population. Situated in the context of rapid urbanization and population mobility in China and beyond, the findings can provide empirical evidence to inform the design of effective emotional support networks for the floating elderly.

Specifically, the HRQoL of the floating elderly was generally good in this study, with a mean score of 0.884. Using the same utility value system, the health utility was slightly higher than those reported by Tan, Chen (59) for elderly people in Shandong province (0.870) and by Wang et al. [60] in Beijing, China (0.853). This discrepancy may be attributed to the higher proportion of individuals under 70 years of age in this study, as only those in relatively better health status are likely to float, supporting the concept of “healthy immigration” [61]. However, while pain and discomfort were the most significant challenges, consistent with most studies [62, 63], anxiety and depression were more prevalent among the floating elderly, contrasting with findings where these were least reported [59, 64]. This difference may be due to the challenges of relocation, as the floating elderly must adapt to new environments and face stressors such as lifestyle changes and the loss of former social networks. These factors can lead to loneliness and anxiety, which negatively impact their mental health.

Regarding the emotional support network characteristics, this study found that the emotional support network of the floating elderly was small, with an average size of no more than three people. The network members were primarily relatives and closely connected. Due to factors such as the limited scope of activities and the social environment of their new residence, the social support networks of the floating elderly tend to be small after relocating to the inflow area. Additionally, some former network members may become estranged due to increasing geographic separation [65]. For floating populations, kinship ties are more important than geospatial proximity [66], and thus kin members make up the vast majority of the network. Moreover, the emotional support network of the floating elderly exhibited greater convergence than heterogeneity, suggesting they tend to receive emotional support from individuals with similar gender, age, and education backgrounds. From the perspective of social capital theory, this implies that bonding social capital outweighs bridging social capital in their emotional support networks [67]. In other words, the floating elderly had limited interaction with broader social circles, and the social resources within their emotional support networks were relatively inadequate.

After controlling for potential confounders, the emotional support networks of the floating elderly had a significant impact on their health-related quality of life, with notable gender differences in this relationship. These findings can be understood within the frameworks of the convoy model of social relations and social capital theory, and further interpreted through the lens of gender role theory. Social support networks, as a form of social capital, have been recognized as an important social determinant of health [33]. In this study, a larger emotional support network was found to be detrimental to the health of both male and female floating elderly, whereas more kin members in the network were associated with higher health utility. This finding aligns with social capital theory and suggests that for the floating elderly, a larger support network is not necessarily better, and the presence of close family members may play a more crucial role in promoting their health. According to the convoy model of social relations, individuals in later life tend to intentionally narrow their social networks and prioritize emotionally close companions, as these relationships are more beneficial for their health and subjective well-being [43, 68]. The findings of this study provide empirical support for this hypothesis. Larger networks are likely to include more weak ties, the maintenance of which often requires greater emotional investment and social skills [24, 69], potentially leading to psychological stress or an increased sense of burden for the elderly. Furthermore, while Stevens and Westerhof [70] found that social support from friends reduced loneliness in older adults, with family support not directly linked to it, the context of traditional Chinese values may provide a different perspective. In Chinese culture, family support is considered the primary source of care for the elderly [71]. Support from friends, neighbors, or social institutions is often more materialistic and may not provide adequate emotional support. In contrast, support from family members tends to fulfill deeper emotional needs, reducing feelings of isolation and psychological burden, and ultimately improving the health of the elderly.

In addition, our findings suggest that higher densities of emotional support networks were linked to poorer HRQoL for female participants, but not for males. This significant gender difference can be understood within the frameworks of gender role theory and the concept of relational maintenance responsibilities [72]. On one hand, compared to men, women are more frequently engaged in caregiving-related occupations and family roles, such as being primary caregivers or homemakers [73]. These roles not only make them recipients of emotional support but also place them in the position of maintaining and coordinating relationships [74]. A high-density network may require substantial time and emotional investment to sustain, placing an excessive burden of “emotional labor” on women and creating role pressure that contributes to increased emotional strain [75]. On the other hand, women are more likely to express their emotions and seek support when facing stress. However, in a densely connected network, excessive emotional expression may be perceived as vulnerability or dependence, which negatively affects their self-esteem and identity [76]. Thus, while higher network densities may offer more emotional support to female floating seniors, if not balanced with clear boundaries and healthy interactions, this support may have negative effects on their health utility.

Finally, the study found that higher age convergence in the emotional support networks of both male and female floating elderly was associated with better health utility. This suggests that receiving emotional support from peers may be particularly beneficial for the health of the floating elderly. According to the convoy model, older adults tend to prioritize familiar social partners who offer emotional satisfaction. For the floating elderly, who often face the stress of adapting to new environments after migration, interacting with peers of similar age and life experience can foster stronger empathy, shared topics, and a sense of identity. This form of bonding social capital can effectively fulfill their emotional needs. While emotional support from children is often seen as crucial for promoting better health in older adults [77, 78], we do not deny this view. However, compared to child relationships, peer support offers a more reciprocal exchange of social resources between equals, which is also effective in enhancing an individual’s sense of self-worth, dignity, and accessibility [79]. This, in turn, positively impacts health outcomes. Therefore, the findings highlight the important role of peers in improving the health utility of the floating elderly, emphasizing the need for emotional support from peers.

At the same time, for female floating elderly, a greater number of individuals with varying education levels in their emotional support network was associated with poorer HRQoL, whereas no such effect was observed in males. High educational heterogeneity in a network may reflect the presence of more potential bridging social capital, which can facilitate better connections between the floating elderly and the broader society. However, differences in communication styles, worldviews, and problem-solving approaches among individuals with diverse educational backgrounds may also arise [80]. Women, in particular, tend to value interpersonal harmony, emotional depth, and empathetic communication. When their support networks include more members with differing educational levels, they may encounter greater communication challenges and interpersonal tension, which could negatively affect their well-being. In contrast, men tend to have more straightforward, problem-focused social interactions, emphasizing practical support rather than emotional communication. They may experience less emotional distress and psychological stress when interacting with individuals of different educational backgrounds, making them less likely to suffer the negative health effects that arise from educational differences within their support networks.

This study sheds some light on enhancing the HRQoL of floating elderly. First, although the family’s informal social support role has been weakened in China due to the one-child policy and increased geographical mobility [81], family emotional support remains irreplaceable. A large emotional support network may not be necessary, and attention should be given to floating elderly individuals who lack kin members to provide emotional support. Second, the study highlights the significant role of peers in offering emotional support and maintaining the health of the floating elderly. Finally, for the floating elderly, an ideal emotional support network should be both heterogeneous and convergent, both building close relationships and facilitating access to a variety of resources. For women, efforts should be made to diversify their emotional support networks while minimizing conflicts arising from educational differences among members.

However, several limitations should be acknowledged. First, the study relies on self-reported data and pre-defined classification schemes, which may introduce potential recall bias and social desirability bias, particularly among older respondents who may underreport or misreport their emotional support experiences. Second, EQ-5D-5L has been reported to show superior performance over EQ-5D-3L in terms of lower ceiling effect and better discriminatory power [82, 83], thus EQ-5D-5L could be considered for future studies. Third, a lack of comparable studies makes it difficult to determine whether the emotional support network characteristics observed among the floating elderly are relatively high or low. Future studies could expand the sample size and include different populations to make comparisons across groups and regions. Finally, this is a cross-sectional study and cannot establish causal relationships between emotional support networks and HRQoL. Although we attempted to interpret the findings from the theoretical perspective, these interpretations remain tentative and should be treated with caution. Future studies could incorporate qualitative interviews or mixed-methods approaches to validate these assumptions and better uncover the underlying mechanisms.

Conclusions

In this study, HRQoL among the floating elderly was generally good. Their emotional support networks were characterized by being “small in size and high in density”, with higher converge than heterogeneity. A larger size of emotional support networks was negatively associated with health utility for both genders, whereas a greater number of kin members and higher age convergence within the network were linked to better health outcomes. For female floating seniors, higher network density and greater educational heterogeneity had adverse health effects. A large emotional support network may not be necessary for the floating elderly, and attention should be given to those lacking emotional support from kin members. Moreover, efforts to optimize the structure of their emotional support networks, particularly by acknowledging the positive role of peers, are crucial. It is also important to consider gender differences in the design and implementation of such support systems.

Abbreviations

EQ-5D-3L

EuroQol 5-Dimensions 3-Level

HRQoL

Health-related quality of life

Author contributions

J.X. undertook the data analysis and wrote the manuscript. Z.K. performed data collection and reviewed the methodology. X.H. designed the study and was responsible for funding acquisition. B.C. and L.Z. provided additional support during data entry and analysis. All authors reviewed and approved the final manuscript.

Funding

The research was supported by: National Natural Science Foundation of China (71774034, 72074064).

Data availability

The datasets used in the current study are not publicly available due to the confidential policy but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This quantitative study was conducted in strict adherence to the Declaration of Helsinki. Ethical approval for the study protocol was obtained from the Ethics Committee of Harbin Medical University. Informed consent was obtained from all participants before the survey began.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.