Abstract
Background
Community-dwelling older adults differ in their social, psychological, and cognitive resources, which may influence health-related quality of life and healthy aging. Variable-centered approaches may not fully capture this heterogeneity. This study aimed to identify psychosocial and cognitive profiles among community-dwelling older adults and examine differences in health-related quality of life and healthy aging across profiles.
Methods
This study was a secondary analysis of de-identified data from a nationwide cross-sectional survey of 510 community-dwelling older adults in Korea. Latent profile analysis was conducted using social network, perceived social support, depressive symptoms, and subjective cognitive function as profile indicators. Health-related quality of life, measured by the Physical Component Summary and Mental Component Summary of the 12-item Short Form Health Survey, and healthy aging were examined as distal outcomes. Covariate effects of age, sex, education years, living status, and economic status were removed using linear regression residuals. Welch’s one-way analysis of variance and Games–Howell post hoc tests were used to compare residualized outcomes across profiles.
Results
A five-profile solution was selected based on model fit, entropy, minimum profile size, parsimony, and interpretability. The profiles were labeled as the High-Depression Vulnerable Group (n = 29), Moderate-Depression Socially Connected High-Resource Group (n = 53), Low-Depression Socially Constrained Group (n = 129), Elevated-Depression Intermediate-Resource Group (n = 41), and Low-Depression, High-Support, Limited-Network Group (n = 258). Profile membership was significantly associated with the Physical Component Summary, F(4, 107) = 81.66, p < .001; Mental Component Summary, F(4, 107) = 10.27, p < .001; and healthy aging, F(4, 110) = 53.51, p < .001. The Low-Depression, High-Support, Limited-Network Group had the highest healthy aging score, significantly exceeding all other profiles. Physical health-related quality of life was similarly high in the Moderate-Depression Socially Connected High-Resource Group and the Low-Depression, High-Support, Limited-Network Group. The High-Depression Vulnerable Group had the lowest Physical Component Summary and healthy aging scores.
Conclusions
Community-dwelling older adults showed distinct psychosocial and cognitive profiles that were differentially associated with health-related quality of life and healthy aging. Healthy aging was most favorable in the group characterized by limited social network resources but high perceived social support, low depressive symptoms, and favorable subjective cognitive function. These findings suggest that healthy aging may depend not only on social network resources, but also on perceived support, depressive symptom burden, and subjective cognitive function. Person-centered assessment may help identify older adults who require tailored strategies to promote quality of life and healthy aging in community settings.
Keywords: Community-dwelling older adults, Healthy aging, Health-related quality of life, Social support, Social network, Depression, Subjective cognitive function, Latent profile analysis
Introduction
Population aging has increased the need to understand not only how long people live but also how older adults maintain health, functioning, autonomy, social connectedness, and meaning in later life.
Contemporary gerontological research conceptualizes healthy aging as a multidimensional process shaped by intrinsic capacity, environmental resources, and social participation. Beard et al. described the WHO healthy aging framework as centered on functional ability, reflecting the interaction between individual capacity and environment [1]. From this perspective, healthy aging is not simply the absence of disease, but the ability to engage in valued activities, maintain relationships, participate in society, adapt to age-related changes, and preserve meaning in later life.
Health-related quality of life is also a key outcome in aging research because it reflects older adults’ perceived physical, psychological, and social functioning. The WHOQOL Group defined quality of life as individuals’ perception of their position in life in relation to their culture, value systems, goals, expectations, standards, and concerns [2].
Although health-related quality of life and healthy aging are closely related, they are conceptually distinct. Health-related quality of life primarily reflects perceived health and functioning, whereas healthy aging captures a broader and more personally meaningful sense of living well in later life. Examining both outcomes may therefore provide a more comprehensive understanding of well-being among community-dwelling older adults.
Social isolation and limited social network resources are major concerns among older adults living in the community. Social isolation generally refers to objective deficits in social contact, social participation, or social network resources, whereas loneliness refers to the subjective experience of perceived social disconnection. Meta-analytic evidence has shown that both actual and perceived social isolation are associated with increased mortality risk [3], and that stronger social relationships are associated with a greater likelihood of survival [4].
Poor social relationships, loneliness, and social isolation have also been linked to increased risks of coronary heart disease, stroke, cognitive decline, and dementia [5, 6]. These findings suggest that social connectedness is not only an emotional resource, but also an important determinant of health, functioning, and well-being in later life.
Both social network resources and perceived social support may influence aging-related outcomes. Social network reflects the structural and functional resources available through social relationships, whereas perceived social support reflects the subjective availability of emotional, informational, and instrumental assistance. Older adults with similar social network resources may differ substantially in perceived support, coping capacity, and psychological well-being. Previous studies have shown that social participation, social support, and perceived connectedness are associated with better quality of life among older adults [7, 8]. Therefore, both social network resources and subjective social support should be considered when examining quality of life and healthy aging.
Psychological and cognitive factors are also important. Depression is common in later life and is consistently associated with poorer quality of life [9]. Depressive symptoms are also associated with functional disability, reduced activity engagement, poorer self-care, and lower social functioning [10, 11]. In contrast, perceived social support may serve as a protective resource by supporting psychological adaptation, coping, well-being, and functional maintenance in the face of illness, loss, and functional decline [12]. Subjective cognitive function is also relevant because perceived cognitive difficulties have been associated with affective symptoms, reduced activity engagement, functional difficulties, and poorer quality of life, even among older adults without clinically established cognitive impairment [13, 14].
Although previous research has identified social, psychological, and cognitive factors associated with quality of life and healthy aging, many studies have used variable-centered approaches. These approaches are useful for estimating independent or conditional associations among variables, but they may not fully capture heterogeneity among older adults who experience different combinations of resources and vulnerabilities. Person-centered methods, including latent profile analysis, are specifically designed to model such heterogeneity by identifying unobserved subgroups of individuals with similar patterns across multiple indicators [15, 16]. Previous studies have shown that older adults can be classified into distinct psychosocial, cognitive, and social subgroups, and that these subgroups differ in quality of life and related health outcomes [17, 18]. Prior research supports the need to distinguish social network resources from perceived social support, as they are related but conceptually distinct dimensions of social connectedness [19, 20].
Latent profile analysis may therefore provide useful insight into how combinations of social, psychological, and cognitive characteristics are associated with later-life outcomes. For example, some older adults may have relatively strong social network resources but still report depressive symptoms or low perceived support, whereas others may have limited social networks but favorable subjective psychosocial resources. Subjective cognitive function is also important because perceived cognitive difficulties may affect confidence, autonomy, social participation, and perceived ability to manage daily life [13, 14].
In this study, social network, perceived social support, depressive symptoms, and subjective cognitive function were used as profile indicators because they represent objective and subjective dimensions of social, psychological, and cognitive functioning. Identifying distinct profiles may clarify how different combinations of resources and vulnerabilities are associated with health-related quality of life and healthy aging.
Therefore, this study aimed to identify latent profiles of community-dwelling older adults based on social network, perceived social support, depressive symptoms, and subjective cognitive function, and to examine differences in health-related quality of life and healthy aging across profiles. This secondary analysis used data from a nationwide survey of community-dwelling older adults in Korea and adjusted for age, sex, education years, living status, and economic status. Findings may extend previous variable-centered evidence by identifying heterogeneous subgroups of older adults and suggesting profile-specific strategies to promote quality of life and healthy aging in community settings.
Methods
Study design and data source
This study was a secondary analysis of de-identified data from an unpublished nationwide cross-sectional survey of community-dwelling older adults in Korea. The parent study used a correlational survey design to examine psychosocial, cognitive, and health-related characteristics of older adults living in urban and rural communities.
The present study used variables collected as part of the approved parent study protocol. It did not involve new participant recruitment, additional data collection, or analysis of personally identifiable information. This secondary analysis addressed a distinct research question from the parent study by applying a person-centered analytic approach to the existing survey dataset.
Specifically, latent profile analysis was conducted to identify subgroups of older adults based on social network resources, perceived social support, depressive symptoms, and subjective cognitive function. Health-related quality of life and healthy aging were then compared across the identified profiles.
Participants
The parent dataset included 510 community-dwelling adults aged 65 years or older.
Participants were eligible if they lived in the community, were able to communicate with trained interviewers, and provided written informed consent to participate in face-to-face interviews and complete standardized questionnaires. Older adults were excluded if they were hospitalized, institutionalized in nursing homes or long-term care facilities, or unable to communicate or comprehend the questionnaire.
Sample size calculation
The available sample of 510 older adults was considered adequate for latent profile analysis. Methodological and simulation studies indicate that sample adequacy in mixture modeling depends not only on total sample size, but also on the number of profiles, number and quality of indicators, profile separation, classification accuracy, and minimum profile size [16, 21].
In the present study, four profile indicators were used, the selected five-profile model showed high entropy (0.907), and the final solution was retained after considering model fit, classification quality, minimum profile size, parsimony, and substantive interpretability.
Data collection procedure
Data were originally collected from September 15 to December 30, 2024, by a contracted professional survey agency with an established nationwide survey network. The agency was engaged to support participant recruitment and standardized face-to-face data collection across Korea under a data confidentiality agreement that prohibited the unauthorized use or disclosure of participant data.
The survey was conducted in diverse community settings across four metropolitan cities, ten small- and medium-sized cities, and nine county-level areas. Eligible community-dwelling older adults were recruited from cities and provinces to reflect variation in age group, sex, and geographic region. Recruitment notices were posted in community settings, including apartment complexes, senior centers, shopping areas, and bus terminals.
Individuals who expressed interest in the study received detailed explanations of the study purpose, procedures, voluntary nature of participation, confidentiality protections, and their right to withdraw at any time. Written informed consent was obtained before participation. Trained interviewers then conducted face-to-face interviews using structured questionnaires. Each interview lasted approximately 30–40 min.
The professional survey agency was involved only in recruitment support and data collection.
It had no role in study conceptualization, data analysis, interpretation of findings, manuscript preparation, or the decision to submit the manuscript.
Ethical considerations
This study was granted an exemption from additional ethical review by the Institutional Review Board of Chungnam National University (No. 202407-SB-099-01) and was conducted in accordance with the principles of the Declaration of Helsinki. The parent study had obtained written informed consent from all participants before data collection. The present study involved a secondary analysis of de-identified data from the approved parent study protocol. No additional participant recruitment, data collection, or use of personally identifiable information was conducted.
Measures
General characteristics
Sociodemographic characteristics included age, sex, education, cohabitation status, and economic status. Age was measured in years. Sex was coded as male, with female as the reference category. Education was assessed as total years of formal education. Cohabitation status was coded as living alone, with living with others as the reference category. Economic status was coded as low economic status, with middle or high economic status as the reference category.
Social support
Social support was assessed using a 10-item social support scale representing emotional, informational, and instrumental support [22]. This scale was developed based on the Medical Outcomes Study Social Support Survey [23] and revised to reflect the cultural context of community-dwelling older adults in Korea [22]. Items were rated on a 5-point Likert scale ranging from 1 (“not at all”) to 5 (“very much”), with higher scores indicating greater perceived social support. Cronbach’s α was 0.88 in the present study.
Social network
Social network was assessed using the Lubben Social Network Scale-Revised (LSNS-R), which was developed and validated for older adults [24]. The LSNS-R assesses family and friend networks, including the number of relatives and friends seen or heard from regularly, the number with whom participants feel at ease discussing private matters, and the number they can call on for help or important decisions. The scale consists of two subscales: family network and friend network, with six items in each subscale. Each item is rated on a 6-point scale ranging from 0 (“none”) to 5 (“very often”). Total scores were calculated by summing all items, with higher scores indicating stronger social network resources. The scale was used with the author’s permission. Cronbach’s α was 0.89 in the present study.
Depression
Depressive symptoms were assessed using the Geriatric Depression Scale Short Form developed by Sheikh and Yesavage [25]. It consists of 15 yes/no items, with depressive responses scored as 1. The Korean short-form version was published by Kee [26] and demonstrated good reliability, with Cronbach’s α = 0.88, split-half reliability = 0.85, and test–retest reliability = 0.90. Items with positive wording were reverse-coded so that higher total scores indicated more severe depressive symptoms. Cronbach’s α was 0.89 in the present study.
Subjective cognitive function
Subjective cognitive function was assessed using the Korean version of the Mild Cognitive Impairment Questionnaire (MCQ), originally developed to measure cognition-related quality of life among individuals with cognitive impairment [27]. In this study, the MCQ assessed emotional and practical concerns related to memory problems. Items are rated on a 4-point Likert scale ranging from 1 (“almost never”) to 4 (“always”). Scores were reverse-coded so that higher scores indicated more favorable subjective cognitive function. The Korean version of the MCQ has demonstrated excellent reliability in a previous study, with Cronbach’s α of 0.93 [28]. Cronbach’s α was 0.92 in the present study.
Health-related quality of life
Health-related quality of life was assessed using the 12-item Short Form Health Survey (SF-12) [29]. The instrument was used with permission from QualityMetric. Physical Component Summary (PCS) and Mental Component Summary (MCS) scores were calculated using the standardized scoring algorithm. Scores range from 0 to 100, with higher scores indicating better health-related quality of life. In the present secondary analysis, PCS and MCS were examined as distal outcomes across latent profiles.
Healthy aging
Healthy aging was measured using the Healthy Aging Scale developed and validated among frail older adults in Korea [22]. This 12-item instrument uses a 4-point Likert scale ranging from 1 (“not at all”) to 4 (“very much so”), yielding a total score from 12 to 48. Higher scores indicate better healthy aging. Cronbach’s α was 0.81 in the development study and 0.88 in the present study.
Statistical analysis
Data were analyzed using IBM SPSS Statistics version 29.0 and R version 4.6.0. Latent profile analysis was conducted using the tidyLPA package in R. Descriptive statistics were used to summarize participant characteristics and study variables. Continuous variables were reported as means and standard deviations, and categorical variables were reported as frequencies and percentages. No missing values were identified across all study variables; therefore, complete data from all 510 participants were included in the analysis.
Latent profile analysis was conducted using depressive symptoms, social network, perceived social support, and subjective cognitive function as profile indicators. Because the indicators had different measurement ranges, all profile indicators were standardized before analysis. Models with three to eight profiles were estimated and compared. Model selection was based on the Bayesian information criterion, sample-size-adjusted Bayesian information criterion, entropy, minimum profile size, parsimony, and substantive interpretability. Lower BIC and SABIC values indicated better relative model fit, whereas higher entropy indicated clearer classification. Solutions with very small profiles were considered potentially unstable and were not preferred unless they provided clear theoretical and practical meaning.
After the optimal profile solution was selected, participants were assigned to their most likely profile based on posterior probabilities. Differences in covariates across latent profiles were examined to assess potential confounding. Continuous covariates, including age and education years, were compared across profiles using Welch’s one-way ANOVA because this approach is robust to unequal group sizes and heterogeneity of variance. Binary covariates, including sex, living alone, and low economic status, were compared using Chi-square tests.
Profile differences in PCS, MCS, and healthy aging were then examined. Before comparing outcomes across profiles, associations between covariates and each outcome were assessed to determine whether profile differences could be partly attributable to demographic or socioeconomic factors. Age, sex, education years, living status, and economic status were included as covariates. Because each covariate was significantly associated with at least one outcome, covariate effects were adjusted by fitting a separate linear regression model for each outcome and extracting the residuals. The residualized PCS, MCS, and healthy aging scores were then used in profile comparisons.
Before conducting group comparisons, the assumptions of normality and homogeneity of variance were examined. Q-Q plots showed that the residualized outcomes were approximately normally distributed, although slight deviations were observed in the tails. Given the large sample size, these deviations were considered acceptable. Homogeneity of variance was assessed using Bartlett’s test, which indicated unequal variances across profiles for the outcome variables. Therefore, Welch’s one-way ANOVA was used to compare residualized PCS, MCS, and healthy aging scores across profiles. Games–Howell post hoc tests were conducted for pairwise comparisons because this method does not assume equal variances or equal group sizes. For interpretability, covariate-adjusted outcome means were presented on the original scale of each outcome. Significant pairwise comparisons were reported based on Games–Howell post hoc tests.
Results
Participant characteristics
The analytic sample included 510 community-dwelling older adults (Table 1). The mean age was 71.60 years (SD = 4.66), and the mean education level was 9.99 years (SD = 3.64). Half of the participants were male (n = 255, 50.0%). A total of 168 participants (32.9%) lived alone, and 212 participants (41.6%) reported low economic status.
Table 1.
Characteristics of the analytic sample
| Characteristics | Total sample (N = 510) | ||
|---|---|---|---|
| Sociodemographic characteristics | Mean ± SD or n (%) | ||
| Age, years | 71.60 ± 4.66 | ||
| Education, years | 9.99 ± 3.64 | ||
| Male | 255 (50%) | ||
| Living alone | 168 (32.9%) | ||
| Low economic status | 212 (41.6%) | ||
| Profile indicators | Possible range | Mean ± SD | |
| Social network | 0–60 | 35.50 ± 7.91 | |
| Social support | 10–40 | 30.40 ± 4.94 | |
| Depressive symptoms | 0–15 | 2.90 ± 3.57 | |
| Subjective cognitive function | 13–65 | 50.20 ± 6.40 | |
| Distal outcomes | Possible range | Mean ± SD | |
| PCS | 0-100 | 42.41 ± 12.45 | |
| MCS | 0-100 | 43.10 ± 7.46 | |
| Healthy aging | 12–48 | 36.94 ± 4.16 | |
PCS Physical Component Summary, MCS Mental Component Summary
For the profile indicators, the mean social network score was 35.50 (SD = 7.91), the mean perceived social support score was 30.40 (SD = 4.94), the mean depressive symptoms score was 2.90 (SD = 3.57), and the mean subjective cognitive function score was 50.20 (SD = 6.40). For the distal outcomes, the mean Physical Component Summary score was 42.41 (SD = 12.45), the mean Mental Component Summary score was 43.10 (SD = 7.46), and the mean healthy aging score was 36.94 (SD = 4.16).
Latent profile model selection
As shown in Table 2, the five-profile solution was selected as the final model. This model showed the lowest BIC value (BIC = 5123) and the highest entropy value (entropy = 0.907), indicating favorable model fit and clear classification. Although the eight-profile model had the lowest SABIC value, it included a very small profile and showed lower entropy, suggesting potential over-extraction. The six- and seven-profile models did not improve the BIC and showed poorer classification quality. Therefore, the five-profile solution was retained as the most parsimonious and substantively meaningful model.
Table 2.
Fit indices for latent profile models
| Number of profiles | BIC | SABIC | Entropy | Smallest Profile (proportion) | Interpretation |
|---|---|---|---|---|---|
| 3 | 5338 | 5281 | 0.862 | 0.09 | Too simple and too high BIC |
| 4 | 5218 | 5145 | 0.866 | 0.09 | Good, but too high BIC |
| 5 | 5123 | 5035 | 0.907 | 0.06 | Selected final model |
| 6 | 5149 | 5045 | 0.784 | 0.06 | BIC increased; entropy decreased |
| 7 | 5149 | 5028 | 0.759 | 0.05 | Low entropy |
| 8 | 5160 | 5024 | 0.775 | 0.03 | The minimum number is too small |
BIC Bayesian information criterion, SABIC Sample-size adjusted Bayesian information criterion
Profile indicator patterns and profile labels
Table 3 presents the raw means of the profile indicators across the five latent profiles. Figure 1 shows the standardized profile patterns using both a line plot and a heatmap of the standardized means (z-scores) of the four profile indicators. These visualizations were used to support profile labeling. Because these indicators were used to derive the latent profiles, between-profile significance tests for the profile indicators were not conducted.
Table 3.
Profile indicator means by latent profile
| Indicator | Profile 1 n = 29 |
Profile 2 n = 53 |
Profile 3 n = 129 |
Profile 4 n = 41 |
Profile 5 n = 258 |
|---|---|---|---|---|---|
| Mean(SD) | Mean(SD) | Mean(SD) | Mean(SD) | Mean(SD) | |
| Social network | 34.21(8.08) | 50.43(4.82) | 33.75(5.54) | 36.80(6.42) | 33.26(6.12) |
| Social support | 20.90(3.21) | 33.89(2.97) | 25.17(2.13) | 28.56(2.90) | 33.63(2.08) |
| Depression | 12.90(1.80) | 4.32(1.81) | 2.27(1.81) | 8.44(1.78) | 0.92(1.20) |
| Cognitive function | 40.41(6.51) | 52.43(5.13) | 48.31(4.87) | 44.95(5.44) | 52.63(5.55) |
Profile indicators were used to derive the latent profiles; therefore, between-profile significance tests are not reported
Profile 1 = High-Depression Vulnerable Group; Profile 2 = Moderate-Depression Socially Connected High-Resource Group; Profile 3 = Low-Depression Socially Constrained Group; Profile 4 = Elevated-Depression Intermediate-Resource Group; Profile 5 = Low-Depression, High-Support, Limited-Network Group
Fig. 1.
Latent profile patterns based on standardized indicator scores. a Line plot and b heatmap of standardized mean scores for the four profile indicators across the five latent profiles. Values are z scores; higher values indicate higher levels of each indicator. For depressive symptoms, higher scores indicate greater depressive symptoms. In the heatmap, blue indicates values below the sample mean, and red indicates values above the sample mean. Both panels present the same standardized profile patterns to support profile labeling and interpretation. Note. Profile 1 = High-Depression Vulnerable Group; Profile 2 = Moderate-Depression Socially Connected High-Resource Group; Profile 3 = Low-Depression Socially Constrained Group; Profile 4 = Elevated-Depression Intermediate-Resource Group; Profile 5 = Low-Depression, High-Support, Limited-Network Group
Profile 1: high-depression vulnerable group
Included 29 participants. This profile was characterized by the highest depressive symptoms, lowest perceived social support, lowest subjective cognitive function, and relatively limited social network resources. This pattern indicated combined psychological, social, and cognitive vulnerability.
Profile 2: moderate-depression socially connected high-resource group
Included 53 participants. This profile showed the strongest social network resources and high perceived social support, relatively high subjective cognitive function, and moderately elevated depressive symptoms. This pattern suggested strong social and cognitive resources despite some depressive symptoms.
Profile 3: low-depression socially constrained group
Included 129 participants. This profile showed relatively low depressive symptoms, but limited social network resources, lower perceived social support, and moderate subjective cognitive function. This pattern indicated relatively favorable emotional status but constrained social resources.
Profile 4: elevated-depression intermediate-resource group
Included 41 participants. This profile was characterized by elevated depressive symptoms, moderate social network resources, moderate perceived social support, and relatively low subjective cognitive function.
This pattern suggested intermediate social resources but continued psychological and cognitive vulnerability.
Profile 5: low-depression, high-support, limited-network group
Was the largest profile in the sample, including 258 participants. This profile showed the most limited social network resources, but high perceived social support, the lowest depressive symptoms, and the highest subjective cognitive function. This pattern suggested that limited social network resources did not necessarily correspond to low perceived support or poor psychological functioning.
Demographic differences across latent profiles
Demographic characteristics differed significantly across the five latent profiles, except for sex (Table 4). Age differed significantly across profiles (p < .001). The High-Depression Vulnerable Group had the highest mean age (74.48 ± 5.02 years), whereas the Moderate-Depression Socially Connected High-Resource Group had the lowest mean age (68.85 ± 3.31 years). Education years also differed significantly across profiles (p < .001), with the Moderate-Depression Socially Connected High-Resource Group showing the highest mean education level (12.79 ± 1.83 years).
Table 4.
Demographic differences across latent profiles
| Variable | Profile 1 n = 29 |
Profile 2 n = 53 |
Profile 3 n = 129 |
Profile 4 n = 41 |
Profile 5 n = 258 |
p |
|---|---|---|---|---|---|---|
| Mean(SD) | Mean(SD) | Mean(SD) | Mean(SD) | Mean(SD) | ||
| Age | 74.48(5.02) | 68.85(3.31) | 70.96(4.03) | 71.93(4.22) | 72.11(4.92) | < 0.001 |
| Education years | 10.48(2.08) | 12.79(1.83) | 9.25(4.06) | 10.46(3.55) | 9.66(3.58) | < 0.001 |
| n (%) | n (%) | n (%) | n (%) | n (%) | ||
| Sex | ||||||
| Female | 12(41.38) | 22(41.51) | 60(46.51) | 26(63.41) | 135(52.33) | 0.163 |
| Male | 17(58.62) | 31(58.49) | 69(53.49) | 15(36.59) | 123(47.67) | |
| Living status | ||||||
| Living together | 9(31.03) | 50(94.33) | 57(44.19) | 28(68.29) | 198(76.74) | < 0.001 |
| Living alone | 20(68.97) | 3(5.66) | 72(55.81) | 13(31.71) | 60(23.26) | |
| Economic status | ||||||
| Middle/high | 11(37.93) | 36(67.92) | 47(36.43) | 18(43.90) | 186(72.09) | < 0.001 |
| Low | 18(62.07) | 17(32.08) | 82(63.57) | 23(56.10) | 72(27.91) | |
Values are presented as mean ± SD for continuous variables and n (%) for categorical variables. Differences across profiles were examined using one-way ANOVA or Welch’s ANOVA for continuous variables, and chi-square tests for categorical variables
Profile 1 = High-Depression Vulnerable Group; Profile 2 Moderate-Depression Socially Connected High-Resource Group; Profile 3 = Low-Depression Socially Constrained Group; Profile 4 = Elevated-Depression Intermediate-Resource Group; Profile 5 = Low-Depression, High-Support, Limited-Network Group.
Living status differed significantly across profiles (p < .001). The proportion of participants living alone was highest in the High-Depression Vulnerable Group (68.97%) and lowest in the Moderate-Depression Socially Connected High-Resource Group (5.66%).
Economic status also differed significantly across profiles (p < .001). Low economic status was most common in the Low-Depression Socially Constrained Group (63.57%) and the High-Depression Vulnerable Group (62.07%), and least common in the Low-Depression, High-Support, Limited-Network Group (27.91%). Sex distribution did not differ significantly across profiles (p = .163).
Differences in health-related quality of life and healthy aging across profiles
Before conducting profile comparisons, covariate associations with the outcome variables were examined.
Older age, female sex, and living alone were associated with lower PCS scores. Middle or high economic status was associated with higher PCS, MCS, and healthy aging scores compared with low economic status. Education years were positively associated with PCS but negatively associated with healthy aging. Because the covariates were associated with at least one outcome, subsequent profile comparisons were conducted using covariate-adjusted outcome scores.
Figure 2 presents the covariate-adjusted mean scores and 95% confidence intervals for PCS, MCS, and healthy aging across the five profiles, expressed on the original scale of each outcome for interpretability. Adjusted means, 95% confidence intervals, and significant pairwise comparisons are presented in Table 5.
Fig. 2.
Adjusted Mean Scores with 95% Confidence of PCS, MCS, and Healthy Aging Across Latent Profiles. Note. PCS = Physical Component Summary; MCS = Mental Component Summary
Table 5.
Differences in PCS, MCS, and healthy aging across latent profiles
| Outcome | Profile | Adjusted Mean |
SE | 95% CI | F | Partial η² |
Significant pairs |
|---|---|---|---|---|---|---|---|
| PCS | 1 | 22.32 | 1.32 | [19.61, 25.03] | 81.66* | 0.754 | 2,3,4,5 > 1 |
| 2 | 46.48 | 1.62 | [43.23, 49.72] | 2,5 > 3,4 | |||
| 3 | 40.72 | 0.80 | [39.14, 42.31] | 3 > 4 | |||
| 4 | 31.74 | 1.88 | [27.95, 35.53] | ||||
| 5 | 46.38 | 0.52 | [45.36, 47.40] | ||||
| MCS | 1 | 43.01 | 1.74 | [39.44, 46.57] | 10.27* | 0.277 | 5 > 3 |
| 2 | 42.99 | 0.95 | [41.09, 44.89] | ||||
| 3 | 40.18 | 0.55 | [39.09, 41.27] | ||||
| 4 | 41.98 | 1.09 | [39.78, 44.18] | ||||
| 5 | 44.78 | 0.46 | [43.88, 45.68] | ||||
| Healthy aging | 1 | 32.30 | 0.82 | [30.61, 33.98] | 53.51* | 0.660 | 5 > 1,2,3,4 |
| 2 | 35.28 | 0.32 | [34.63, 35.92] | 2,3 > 1 | |||
| 3 | 35.63 | 0.30 | [35.03, 36.23] | 3 > 4 | |||
| 4 | 33.96 | 0.44 | [33.07, 34.86] | ||||
| 5 | 38.93 | 0.20 | [38.53, 39.34] |
Adjusted means are presented on the original scale of each outcome. Effect sizes were reported as partial η² for the overall profile effect for each outcome
Profile 1 = High-Depression Vulnerable Group; Profile 2 = Moderate-Depression Socially Connected High-Resource Group; Profile 3 = Low-Depression Socially Constrained Group; Profile 4 = Elevated-Depression Intermediate-Resource Group; Profile 5 = Low-Depression, High-Support, Limited-Network Group
PCS Physical Component Summary, MCS Mental Component Summary, CI Confidence interval, Adjusted mean residualized outcome means with covariate effects removed, expressed in the original scale of each outcome, SE Standard error
*p < .001. Significant pairs indicate profile pairs with statistically significant differences based on Games-Howell post-hoc tests (adjusted p < .05)
Physical component of health-related quality of life
PCS differed significantly across profiles, F(4, 107) = 81.66, p < .001, partial η² = 0.754. The Moderate-Depression Socially Connected High-Resource Group (Profile 2) and the Low-Depression, High-Support, Limited-Network Group (Profile 5) showed the highest PCS scores (M = 46.48, SE = 1.62, 95% CI [43.23, 49.72] and M = 46.38, SE = 0.52, 95% CI [45.36, 47.40], respectively), with no significant difference between them. Both profiles had significantly higher PCS scores than Profiles 3 and 4.
The High-Depression Vulnerable Group (Profile 1) showed the lowest PCS score (M = 22.32, SE = 1.32, 95% CI [19.61, 25.03]) and scored significantly lower than all other profiles. The Low-Depression Socially Constrained Group (Profile 3; M = 40.72, SE = 0.80, 95% CI [39.14, 42.31]) had significantly higher PCS scores than Profiles 1 and 4, but significantly lower PCS scores than Profiles 2 and 5. The Elevated-Depression Intermediate-Resource Group (Profile 4; M = 31.74, SE = 1.88, 95% CI [27.95, 35.53]) scored higher than Profile 1 but lower than Profiles 2, 3, and 5.
Mental component of health-related quality of life
MCS differed significantly across profiles, F(4, 107) = 10.27, p < .001, partial η² = 0.277. The Low-Depression, High-Support, Limited-Network Group (Profile 5) showed the highest MCS score (M = 44.78, SE = 0.46, 95% CI [43.88, 45.68]), while the Low-Depression Socially Constrained Group (Profile 3) showed the lowest score (M = 40.18, SE = 0.55, 95% CI [39.09, 41.27]). Pairwise comparisons showed that Profile 5 had a significantly higher MCS score than Profile 3. Other pairwise differences were not statistically significant (Table 5).
Healthy aging
Healthy aging differed significantly across profiles, F(4, 110) = 53.51, p < .001, partial η² = 0.660. The Low-Depression, High-Support, Limited-Network Group (Profile 5) had the highest healthy aging score (M = 38.93, SE = 0.20, 95% CI [38.53, 39.34]) and scored significantly higher than all other profiles.
The High-Depression Vulnerable Group (Profile 1) had the lowest healthy aging score (M = 32.30, SE = 0.82, 95% CI [30.61, 33.98]) and scored significantly lower than the Moderate-Depression Socially Connected High-Resource Group (Profile 2; M = 35.28, SE = 0.32, 95% CI [34.63, 35.92]), the Low-Depression Socially Constrained Group (Profile 3; M = 35.63, SE = 0.30, 95% CI [35.03, 36.23]), and the Low-Depression, High-Support, Limited-Network Group. The Low-Depression Socially Constrained Group also had a significantly higher healthy aging score than the Elevated-Depression Intermediate-Resource Group (Profile 4; M = 33.96, SE = 0.44, 95% CI [33.07, 34.86]). Other pairwise comparisons were not statistically significant.
Discussion
This study identified five distinct psychosocial and cognitive profiles among community-dwelling older adults based on social network resources, perceived social support, depressive symptoms, and subjective cognitive function. These profiles were significantly associated with health-related quality of life and healthy aging. These findings demonstrate that older adults living in the community are not a homogeneous group, but differ in the ways social, psychological, and cognitive resources coexist. Thus, a person-centered approach may be useful for identifying subgroups of older adults who require different health promotion strategies.
A key finding was that the Low-Depression, High-Support, Limited-Network Group showed the most favorable healthy aging outcome, despite having the lowest social network score. This group also had high perceived social support, the lowest depressive symptoms, and the highest subjective cognitive function. This pattern suggests that social network resources alone may not fully explain older adults’ ability to age well. Prior research supports the need to distinguish social network resources from perceived social support, as these are related but conceptually distinct dimensions of social connectedness with potentially different implications for later-life health and well-being [19, 20]. In this study, the quality of perceived support appeared to be more closely aligned with healthy aging than the size or extent of social network resources. Thus, older adults with limited social network resources may still experience favorable healthy aging when they perceive their available relationships as supportive and have favorable psychological and cognitive indicators.
This finding also highlights the distinction between health-related quality of life and healthy aging. Health-related quality of life primarily reflects perceived physical and mental health status, whereas healthy aging is broader and includes the capacity to live well, adapt to age-related changes, maintain valued activities, preserve relationships, and experience meaning in later life. This interpretation is consistent with the WHO healthy aging framework, which conceptualizes healthy aging in terms of functional ability shaped by intrinsic capacity and environmental resources. It is also consistent with older adults’ own perspectives on aging well, which emphasize independence, positive attitude, social engagement, purpose, and meaning in addition to physical and cognitive functioning [2, 30, 31]. Therefore, the favorable healthy aging score in the Low-Depression, High-Support, Limited-Network Group may reflect not only better perceived health status, but also psychological adaptation, cognitive confidence, and perceived relational support. Strategies to promote healthy aging may therefore need to go beyond increasing the number of social contacts and should also address perceived support, depressive symptoms, subjective cognitive function, and opportunities to maintain meaning in everyday life.
The distinction between social network resources and perceived social support was also evident in the Moderate-Depression Socially Connected High-Resource Group. This group had the strongest social network resources, high perceived social support, and high subjective cognitive function, and its PCS was comparable to that of the Low-Depression, High-Support, Limited-Network Group. These findings suggest that strong social and cognitive resources may support the physical aspects of quality of life, consistent with previous studies linking social network characteristics and social support to health-related quality of life among older adults [32, 33]. However, the presence of moderately elevated depressive symptoms indicates that social connectedness alone may not be sufficient to ensure emotional well-being or healthy aging. Community-based health promotion strategies should therefore assess social network resources, perceived support, and emotional status together [12, 19, 34]. Although this group showed a relatively favorable MCS score, this should be interpreted cautiously because the SF-12 MCS reflects broad mental health-related quality of life rather than depressive symptoms alone.
In contrast, the High-Depression Vulnerable Group showed the least favorable outcome pattern. This group was characterized by the highest depressive symptoms, lowest perceived social support, lowest subjective cognitive function, relatively limited social network resources, the highest mean age, and high proportions of participants living alone and reporting low economic status. It also had the lowest PCS and healthy aging scores. Although its MCS score appeared relatively favorable, pairwise comparisons showed that this group did not differ significantly from most other profiles. Therefore, the MCS result should not be interpreted as evidence of favorable emotional status. These findings suggest that combined psychological, social, cognitive, and socioeconomic vulnerability may place older adults at particular risk for poor aging-related outcomes. This is consistent with evidence linking depression, subjective cognitive impairment, and combined depression-cognitive vulnerability to poorer quality of life, functional disability, and mortality risk [9, 13, 35]. For this group, strategies focused only on increasing social participation may be insufficient.
Integrated approaches that include depression screening, referral pathways, perceived support enhancement, cognitive confidence-building, and accessible low-intensity activity programs may be more appropriate [36, 37].
The two remaining profiles further demonstrate that older adults may have mixed patterns of resources and vulnerabilities. The Low-Depression Socially Constrained Group had relatively low depressive symptoms but constrained social resources and moderate subjective cognitive function. This pattern suggests that some older adults may maintain emotional stability despite limited social resources; however, their lower PCS and MCS compared with more favorable profiles indicate that social constraints may still limit overall well-being. For this group, strategies may focus on expanding meaningful social opportunities, strengthening community engagement, and improving access to health-promoting activities [38, 39]. In contrast, the Elevated-Depression Intermediate-Resource Group had moderate social resources but elevated depressive symptoms and relatively low subjective cognitive function. This group may benefit from interventions that prioritize emotional support, depression management, cognitive confidence, and functional recovery, rather than assuming that moderate social resources are sufficient [13, 40].
The outcome patterns also suggest that PCS, MCS, and healthy aging captured different aspects of later-life well-being. The Low-Depression, High-Support, Limited-Network Group showed the most favorable outcome for healthy aging, with a score significantly higher than all other profiles. However, this pattern was not identical across the health-related quality of life outcomes. For PCS, both the Moderate-Depression Socially Connected High-Resource Group and the Low-Depression, High-Support, Limited-Network Group showed similarly high scores, with no significant difference between these two profiles. This suggests that favorable physical health-related quality of life may be associated with more than one resource pattern: strong social network resources and high perceived support in the Moderate-Depression Socially Connected High-Resource Group, and high perceived support, low depressive symptoms, and favorable subjective cognitive function in the Low-Depression, High-Support, Limited-Network Group.
In contrast, MCS showed a less distinct pattern. Although the Low-Depression, High-Support, Limited-Network Group had the highest MCS score numerically, the differences from the High-Depression Vulnerable Group and the Moderate-Depression Socially Connected High-Resource Group were small and should not be overinterpreted. Importantly, the MCS pattern did not closely correspond to the depressive symptom profiles. In the present study, the correlation between MCS and depressive symptoms measured by the GDS-SF was weak (r = − .108), suggesting that these measures captured related but distinct aspects of mental health and well-being. The GDS-SF directly assesses depressive symptoms, whereas the SF-12 MCS is a broader summary measure of mental health-related quality of life. Therefore, profiles with elevated depressive symptoms did not necessarily show proportionally low MCS scores. This distinction helps explain why the High-Depression Vulnerable Group and the Moderate-Depression, Socially Connected High-Resource Group showed MCS scores that appeared relatively favorable despite elevated depressive symptoms.
One possible interpretation is that the SF-12 MCS captures mental health-related functioning, whereas healthy aging captures a broader, meaning-oriented evaluation of living well in later life [1, 29]. Accordingly, the MCS findings should be interpreted cautiously and together with the depressive symptom profile, rather than as a direct indicator of depression severity. Overall, these findings indicate that different outcome measures yielded different profile patterns: the Low-Depression, High-Support, Limited-Network Group was most favorable for healthy aging, the Moderate-Depression Socially Connected High-Resource Group and the Low-Depression, High-Support, Limited-Network Group were similarly favorable for PCS, and MCS showed only modest differences across profiles.
Taken together, these findings show that healthy aging among community-dwelling older adults is shaped by more than physical and mental health status alone. The most favorable healthy aging outcome was not observed in the group with the strongest social network resources, but in the group with high perceived support, low depressive symptoms, and favorable subjective cognitive function despite limited social network resources. This finding has important implications for community-based gerontological practice because it suggests that the subjective quality of support, psychological well-being, and cognitive confidence may be as important as the size of social networks. Strategies to promote healthy aging may therefore need to support meaningful relationships, perceived support, depressive symptom management, cognitive confidence, and engagement in personally valued activities.
Limitations
This study has several limitations. First, the cross-sectional design limits causal interpretation.
Although latent profiles were associated with health-related quality of life and healthy aging, the directionality of these associations cannot be determined. Second, this study was a secondary analysis of an existing dataset; therefore, the selection of profile indicators, covariates, and outcomes was limited to variables available in the parent survey. Third, subjective cognitive function was assessed using perceived cognitive concerns rather than objective neuropsychological testing. However, subjective cognitive function remains meaningful because perceived cognitive difficulties may influence confidence, autonomy, social participation, and perceived ability to manage daily life. Fourth, the MCS findings should be interpreted with caution because the SF-12 MCS is a broad summary measure of mental health-related quality of life and may not correspond directly to depressive symptom severity. This limitation is particularly relevant when interpreting profiles with elevated depressive symptoms but relatively preserved MCS scores.
Finally, the findings may not be generalizable to institutionalized older adults, older adults in other cultural contexts, or the broader population of older adults in Korea. Although the sample included participants from both urban and rural communities, recruitment was conducted through community-based settings, and participants were required to communicate with trained interviewers and complete standardized questionnaires. Therefore, the sample may have overrepresented older adults who were relatively active, independent, and willing to participate in community-based research, while underrepresenting homebound, highly frail, socially isolated, or cognitively and physically impaired older adults. Accordingly, the relative size of each latent profile should be interpreted as sample-specific. Although the Low-Depression, High-Support, Limited-Network Group was the largest profile in this study, this should not be interpreted as evidence that it is the most common profile among older adults in Korea. Future studies using population-based or home-visit recruitment strategies are needed to examine the reproducibility and population prevalence of these profiles.
Implications
Despite these limitations, this study contributes to gerontological research by identifying distinct psychosocial and cognitive profiles among community-dwelling older adults and showing that these profiles are differentially associated with health-related quality of life and healthy aging. The findings extend previous variable-centered evidence by demonstrating that combinations of social network, perceived social support, depressive symptoms, and subjective cognitive function provide meaningful information for understanding later-life outcomes.
Importantly, the findings suggest that healthy aging is not merely a reflection of physical or mental health-related quality of life. Rather, healthy aging appears to be a broader outcome shaped by internal and external resources that help older adults live meaningfully in the community. Older adults with limited social network resources may still report favorable healthy aging when they have strong perceived support, low depressive symptoms, and favorable subjective cognitive function. Therefore, community-based health promotion strategies should not focus only on increasing social contact, but should also strengthen perceived support, emotional well-being, cognitive confidence, and meaningful engagement.
Person-centered assessment may help identify older adults who require different types and levels of support. Older adults in high-depression and low-resource profiles may benefit from integrated approaches that include depression screening, referral pathways, perceived support enhancement, cognitive confidence-building, and accessible activity programs. In contrast, older adults with limited social network resources but strong perceived support and favorable psychological and cognitive indicators may benefit from strategies aimed at maintaining existing strengths and preventing future social isolation. Future longitudinal and intervention studies are needed to examine whether profile-specific strategies can improve health-related quality of life and healthy aging among community-dwelling older adults.
Conclusion
This study identified five distinct psychosocial and cognitive profiles among community-dwelling older adults in Korea and showed that these profiles were differentially associated with health-related quality of life and healthy aging. The findings indicate that community-dwelling older adults are not a homogeneous group, but differ in their combinations of social network resources, perceived social support, depressive symptoms, and subjective cognitive function.
The most favorable healthy aging outcome was observed in the Low-Depression, High-Support, Limited-Network Group. This finding suggests that healthy aging may depend not only on social network resources, but also on perceived support, depressive symptom burden, and subjective cognitive function. In contrast, the High-Depression Vulnerable Group had the poorest outcomes, underscoring the need for early identification and integrated support for older adults with combined psychological, social, cognitive, and socioeconomic vulnerabilities.
These findings support the use of person-centered approaches to identify subgroups of older adults who may require different types of health promotion strategies. Community-based interventions to promote healthy aging should move beyond uniform approaches and address profile-specific needs, including strengthening perceived support, reducing depressive symptoms, supporting subjective cognitive function, reducing barriers to participation, and promoting meaningful engagement in later life. Future longitudinal and intervention studies are needed to examine whether profile-specific strategies can improve quality of life and healthy aging among community-dwelling older adults.
Abbreviations
- WHO
World Health Organization
- WHOQOL
World Health Organization Quality of Life
- PCS
Physical Component Summary
- MCS
Mental Component Summary
- HRQOL
Health-related quality of life
- LPA
Latent profile analysis
- BIC
Bayesian information criterion
- SABIC
Sample-size adjusted Bayesian information criterion
- CI
Confidence interval
- SD
Standard deviation
- SE
Standard error
Authors' contributions
Y.C. and R.S. conceived of the study, and R.S. received the funding to conduct the study. S.L. carried out statistical analysis and visual presentation of the tables and figures. R.S., S.L., and Y.C. participated in its design and coordination, helped to draft the manuscript. All authors meet authorship criteria and approved the final manuscript.
Funding
This research was supported by National Research Foundation of Korea (Grant number 2022R1A2C2011502).
Data availability
The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
Approval for the study protocol was obtained from the Institutional Review Board of Chungnam National University (No. 202407-SB-099-01). All participants were fully informed about the purpose and procedures of the study, and provided written consent.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Beard JR, Officer A, de Carvalho IA, Sadana R, Pot AM, Michel JP, Lloyd-Sherlock P, Epping-Jordan JE, Peeters G, Mahanani WR, et al. The world report on ageing and health: a policy framework for healthy ageing. Lancet. 2016;387(10033):2145–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.WHOQOL_Group. The World Health Organization Quality of Life assessment (WHOQOL): position paper from the World Health Organization. Soc Sci Med. 1995;41(10):1403–9. [DOI] [PubMed]
- 3.Holt-Lunstad J, Smith TB, Baker M, Harris T, Stephenson D. Loneliness and social isolation as risk factors for mortality: a meta-analytic review. Perspect Psychol Sci. 2015;10(2):227–37. [DOI] [PubMed] [Google Scholar]
- 4.Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS Med. 2010;7(7):e1000316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kuiper JS, Zuidersma M, Zuidema SU, Burgerhof JG, Stolk RP, Oude Voshaar RC, Smidt N. Social relationships and cognitive decline: a systematic review and meta-analysis of longitudinal cohort studies. Int J Epidemiol. 2016;45(4):1169–206. [DOI] [PubMed] [Google Scholar]
- 6.Valtorta NK, Kanaan M, Gilbody S, Ronzi S, Hanratty B. Loneliness and social isolation as risk factors for coronary heart disease and stroke: systematic review and meta-analysis of longitudinal observational studies. Heart. 2016;102(13):1009–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tan SS, Fierloos IN, Zhang X, Koppelaar E, Alhambra-Borras T, Rentoumis T, Williams G, Rukavina T, van Staveren R, Garces-Ferrer J, et al. The association between loneliness and health related quality of life (HR-QoL) among community-dwelling older citizens. Int J Environ Res Public Health. 2020;17(2):600. 10.3390/ijerph17020600. [DOI] [PMC free article] [PubMed]
- 8.Das P, Saha S, Das T, Das P, Roy R, Roy TB. A systematic review and meta-analysis on association between social non-participation and falling in depressive state among the older adult people. Discov Ment Health. 2025;5(1):108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sivertsen H, Bjørkløf GH, Engedal K, Selbæk G, Helvik AS. Depression and quality of life in older persons: a review. Dement Geriatr Cogn Disord. 2015;40(5–6):311–39. [DOI] [PubMed] [Google Scholar]
- 10.Alexopoulos GS, Vrontou C, Kakuma T, Meyers BS, Young RC, Klausner E, Clarkin J. Disability in geriatric depression. Am J Psychiatry. 1996;153(7):877–85. [DOI] [PubMed] [Google Scholar]
- 11.Yuen GS, Bhutani S, Lucas BJ, Gunning FM, AbdelMalak B, Seirup JK, Klimstra SA, Alexopoulos GS. Apathy in late-life depression: common, persistent, and disabling. Am J Geriatr Psychiatry. 2015;23(5):488–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tengku Mohd TAM, Yunus RM, Hairi F, Hairi NN, Choo WY. Social support and depression among community dwelling older adults in Asia: a systematic review. BMJ Open. 2019;9(7):e026667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hill NL, McDermott C, Mogle J, Munoz E, DePasquale N, Wion R, Whitaker E. Subjective cognitive impairment and quality of life: a systematic review. Int Psychogeriatr. 2017;29(12):1965–77. [DOI] [PubMed] [Google Scholar]
- 14.Rotenberg S, Maeir A, Dawson DR. Changes in activity participation among older adults with subjective cognitive decline or objective cognitive deficits. Front Neurol. 2019;10:1393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Berlin KS, Williams NA, Parra GR. An introduction to latent variable mixture modeling (Part 1): overview and cross-sectional latent class and latent profile analyses. J Pediatr Psychol. 2014;39(2):174–87. [DOI] [PubMed] [Google Scholar]
- 16.Spurk D, Hirschi A, Wang M, Valero D, Kauffeld S. Latent profile analysis: a review and how to guide of its application within vocational behavior research. J Vocat Behav. 2020;120:103445. [Google Scholar]
- 17.Lee C, Cho B, Yang Q, Chang SJ, Ko H, Yi YM, Cho H-R, Park Y-H. Psychosocial risk profiles among older adults living alone in South Korea: a latent profile analysis. Arch Gerontol Geriatr. 2021;95:104429. [DOI] [PubMed] [Google Scholar]
- 18.Liu C, Luo Q, Luo D, Zhou Y, Feng X, Wang Z, Xiao J, Bi Q, Smith GD. Quality of life profiles and its association with predictors amongst Chinese older adults in nursing homes: a latent profile analysis. BMC Geriatr. 2023;23(1):740. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cho JH, Olmstead R, Choi H, Carrillo C, Seeman TE, Irwin MR. Associations of objective versus subjective social isolation with sleep disturbance, depression, and fatigue in community-dwelling older adults. Aging Ment Health. 2019;23(9):1130–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rhee TG, Marottoli RA, Monin JK. Diversity of social networks versus quality of social support: Which is more protective for health-related quality of life among older adults? Prev Med. 2021;145:106440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gudicha DW, Tekle FB, Vermunt JK. Power and sample size computation for wald tests in latent class models. J Classif. 2016;33(1):30–51. [Google Scholar]
- 22.Seo J, Joo K, Li Y, Kim N, Oh E, Gansukh L, Song R. Healthy aging in frail older adults: active aging project of a national survey. Int J Nurs Stud Adv. 2025;8:100302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sherbourne CD, Stewart AL. The MOS social support survey. Soc Sci Med. 1991;32(6):705–14. [DOI] [PubMed] [Google Scholar]
- 24.Lubben J, Blozik E, Gillmann G, Iliffe S, von Renteln Kruse W, Beck JC, Stuck AE. Performance of an abbreviated version of the Lubben social network scale among three European community-dwelling older adult populations. Gerontologist. 2006;46(4):503–13. [DOI] [PubMed] [Google Scholar]
- 25.Sheikh JI, Vysavage JA. Geriatric Depression Scale (GDS): recent evidence and development of a shorter version. Clin Gerontol. 1986;5(1–2):165–73. [Google Scholar]
- 26.Kee BS. A preliminary study for the standardization of geriatric depression scale short form-Korea version. J Korean Neuropsychiatr Assoc. 1996;35(2):298–307. [Google Scholar]
- 27.Dean K, Jenkinson C, Wilcock G, Walker Z. The development and validation of a patient-reported quality of life measure for people with mild cognitive impairment. Int Psychogeriatr. 2014;26(3):487–97. [DOI] [PubMed] [Google Scholar]
- 28.Song R, Gang M, Park M, Park M, Jang M, Hwang IO, Kim JL. Validity and reliability of K-MCQ to assess quality of life of older adults with mild cognitive impairment. J Kor Gerontol Nurs. 2021;23(2):164–75. [Google Scholar]
- 29.Ware J Jr., Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34(3):220–33. [DOI] [PubMed] [Google Scholar]
- 30.AshaRani PV, Lai D, Koh J, Subramaniam M. Purpose in Life in older adults: a systematic review on conceptualization, measures, and determinants. Int J Environ Res Public Health. 2022;19(10):5860. 10.3390/ijerph19105860. [DOI] [PMC free article] [PubMed]
- 31.Reich AJ, Claunch KD, Verdeja MA, Dungan MT, Anderson S, Clayton CK, Goates MC, Thacker EL. What does successful aging mean to you? - Systematic review and cross-cultural comparison of lay perspectives of older adults in 13 Countries, 2010–2020. J Cross Cult Gerontol. 2020;35(4):455–78. [DOI] [PubMed] [Google Scholar]
- 32.Garcia EL, Banegas JR, Perez-Regadera AG, Cabrera RH, Rodriguez-Artalejo F. Social network and health-related quality of life in older adults: a population-based study in Spain. Qual Life Res. 2005;14(2):511–20. [DOI] [PubMed] [Google Scholar]
- 33.Belanger E, Ahmed T, Vafaei A, Curcio CL, Phillips SP, Zunzunegui MV. Sources of social support associated with health and quality of life: a cross-sectional study among Canadian and Latin American older adults. BMJ Open. 2016;6(6):e011503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Harasemiw O, Newall N, Mackenzie CS, Shooshtari S, Menec V. Is the association between social network types, depressive symptoms and life satisfaction mediated by the perceived availability of social support? A cross-sectional analysis using the Canadian Longitudinal Study on Aging. Aging Ment Health. 2019;23(10):1413–22. [DOI] [PubMed] [Google Scholar]
- 35.Han S, Gao Y, Gan D. The combined associations of depression and cognitive impairment with functional disability and mortality in older adults: a population-based study from the NHANES 2011–2014. Front Aging Neurosci. 2023;15:1121190. 10.3389/fnagi.2023.1121190. [DOI] [PMC free article] [PubMed]
- 36.Lee SH, Lee H, Yu S. Effectiveness of social support for community-dwelling elderly with depression: a systematic review and meta-analysis. Healthcare. 2022;10(9):1598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Marquez DX, Aguiñaga S, Vásquez PM, Conroy DE, Erickson KI, Hillman C, Stillman CM, Ballard RM, Sheppard BB, Petruzzello SJ, et al. A systematic review of physical activity and quality of life and well-being. Translational Behav Med. 2020;10(5):1098–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Monteiro JM, Gonçalves R, Bastos A, Barbosa MR. Social engagement and wellbeing in late life: a systematic review. Aging Soc. 2025;45(7):1327–54. [Google Scholar]
- 39.Townsend BG, Chen JA-O, Wuthrich VA-OX. Barriers and facilitators to social participation in older adults: a systematic literature review. Clin Gerontol. 2021;44(4):359–80. 10.1080/07317115.2020. [DOI] [PubMed]
- 40.Wassink-Vossen SA-O, Oude Voshaar RC, Naarding P, Collard RM. Effectiveness of late-life depression interventions on functional limitations: a systematic review. Int J Ment Health Nurs. 2022;31(4):823–42. 10.1111/inm.12982. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.