In an observational investigation of 608 community-dwelling older adults across Greece, researchers found that subjective loneliness and multidimensional self-care practices demonstrated substantial independent associations with both mental and physical health-related quality of life, whereas widowed status retained no independent explanatory value once these psychological and behavioral dimensions were taken into account. The study, conducted by a Greek research team and published in Discover Public Health, examined how personal self-regulation and emotional experiences relate to functional well-being in later life. While bereavement and marital transitions have long served as primary administrative benchmarks for social vulnerability among aging populations, the findings indicate that structural demographic categories fail to capture the day-to-day mechanisms that sustain vitality. Instead, an individual’s subjective sense of relational deficit and their concurrent capacity to engage in proactive, multi-domain self-care emerged as the primary statistical correlates of perceived health status, explaining more than half of the total variance across both aggregated physical and psychological health measures.
The conceptual underpinning of the investigation relied upon the biopsychosocial framework alongside modern self-regulation theories, which emphasize that health in older adulthood reflects dynamic, continuous interactions among objective social circumstances, emotional perceptions, and daily self-management behaviors. Within this framework, loneliness is strictly defined as the subjective perception that personal social relationships are deficient in either quantity, quality, or emotional intimacy, rendering it conceptually distinct from objective social isolation or the structural reality of living alone. Widowhood frequently represents one of the most disruptive social transitions encountered in later life, typically accompanied by acute emotional distress and altered household routines. However, the authors emphasized that marital status represents a static structural condition that obscures profound interpersonal heterogeneity. While some widowed individuals encounter persistent isolation and functional decline, others maintain robust social networks, emotional agency, and consistent daily practices. Consequently, treating widowhood as an automatic proxy for vulnerability overlooks the psychological resources and behavioral self-regulation that actually govern how older individuals adapt to significant life losses.
Fieldwork was conducted between July 2024 and July 2025 across multiple geographic regions in Greece, recruiting community-dwelling participants through municipal open care centers for the elderly, primary healthcare clinics, municipal social services, and local community associations. The sampling protocol utilized convenience-based community sampling, requiring participants to be at least 60 years old, residing in non-institutional community settings, capable of communicating fluently in Greek, and possessing the cognitive capacity to provide informed consent and complete assessment batteries independently or with trained interviewer support. Individuals presenting with acute psychiatric disturbances or severe cognitive impairment that precluded meaningful participation were excluded. The principal analytical cohort comprised 608 individuals, with complete age and demographic data recorded for 607 participants aged between 62 and 99 years, exhibiting a mean age of 77.61 years and a standard deviation of 8.53 years. Within this analytical sample, 63.5 percent were women, 58.4 percent were married, 32.7 percent were widowed, and 30.7 percent lived alone in single-person households. Furthermore, 72.9 percent engaged in regular walking, while 35.7 percent reported taking recreational vacations.
Health-related quality of life was evaluated using the standardized 36-Item Short Form Health Survey, which measures functional well-being across eight distinct dimensions: physical functioning, role limitations caused by physical health problems, bodily pain, general health perceptions, vitality, social functioning, role limitations caused by emotional problems, and mental health. Rather than computing norm-based physical and mental component summary scores, the researchers constructed two transparent, study-specific aggregated composite indices ranging theoretically from 0 to 400. The aggregated physical health index combined the scores of physical functioning, role physical, bodily pain, and general health perceptions, while the aggregated mental health index combined vitality, social functioning, role emotional, and mental health. Across the study cohort, the survey demonstrated robust internal reliability with an overall Cronbach’s alpha of 0.89. Concurrently, subjective loneliness was measured using the 11-item De Jong Gierveld Loneliness Scale, which yields an overall loneliness score from 0 to 11 while differentiating between emotional loneliness, characterized by the absence of an intimate attachment figure, and social loneliness, reflecting the lack of a broader social network, achieving an internal consistency alpha of 0.80.
A central feature of the investigation was the operationalization of self-care through the newly developed Multidimensional Self-Care Scale for Older Adults, a 37-item instrument specifically designed to evaluate self-care as an active, person-centered construct rather than a simple checklist of medical compliance tasks. Prior to the main analytical modeling, the instrument’s structural validity was evaluated in a broader sample split randomly into an exploratory factor analysis subset of 202 participants and a confirmatory factor analysis subset of 406 participants. The confirmatory model confirmed a robust seven-factor architecture encompassing functional independence, daily well-being and positive resources, psychological distress and symptoms, prevention and healthcare use, oral hygiene, spirituality and religiosity, and meaning and life purpose, demonstrating strong goodness of fit metrics. Individual items were scored on a five-point Likert scale. To capture overall self-care capacity, the investigators calculated a global composite index by summing the scores of the six positively oriented domains, subtracting the psychological distress and symptoms domain, and dividing the result by the 37 retained items, yielding an index with acceptable reliability and documented three-month test-retest stability.
Preliminary bivariate analyses identified pronounced linear correlations among all principal study variables. Total loneliness exhibited significant negative associations with both aggregated mental health and aggregated physical health, with Pearson correlation coefficients of negative 0.57 and negative 0.50, respectively. Conversely, higher composite multidimensional self-care correlated positively and robustly with mental health at 0.69 and physical health at 0.70, while showing an inverse correlation with total loneliness at negative 0.62, all reaching statistical significance at the p less than 0.001 threshold. Exploratory comparisons between widowed and non-widowed older adults revealed that widowed individuals reported significantly higher total and emotional loneliness, lower physical and mental quality of life, and reduced scores across functional independence, daily well-being, and oral hygiene domains. Interestingly, widowed participants scored significantly higher in spirituality and religiosity. No significant differences emerged regarding social loneliness, preventive healthcare utilization, or life purpose, while the unadjusted elevation in psychological distress among widowed adults became non-significant following Holm adjustment for multiple testing, underscoring that structural bereavement does not produce uniform psychological deficits across all functional domains.
To rigorously examine whether subjective loneliness and self-care accounted for incremental variance beyond demographic and lifestyle attributes, the researchers executed hierarchical multiple regression models utilizing heteroskedasticity-consistent HC3 standard errors across 607 complete cases. In the model predicting aggregated mental health, demographic variables entered at Step 1, comprising age, gender, and widowed status, accounted for 8.3 percent of the variance. The addition of lifestyle factors at Step 2, specifically habitual walking and vacation-taking, explained an additional 14.3 percent. Introducing total loneliness at Step 3 contributed an incremental 18.0 percent of explained variance. In the final step, adding the composite multidimensional self-care index explained an additional 12.0 percent, bringing the total model explained variance to 52.7 percent, with an adjusted R-squared of 0.522. In this fully adjusted model, multidimensional self-care demonstrated the strongest independent relationship with mental health, displaying an unstandardized coefficient of 73.62 and a standardized beta of 0.494, while loneliness remained negatively associated with an unstandardized coefficient of negative 6.17 and a standardized beta of negative 0.216. Demographic indicators, including widowed status, showed no independent statistical associations with mental health outcomes.
A parallel hierarchical regression was performed to evaluate predictors of aggregated physical health-related quality of life. The initial demographic block entered at Step 1 accounted for 15.8 percent of outcome variance, while lifestyle variables entered at Step 2 added another 16.3 percent. The entry of total loneliness at Step 3 explained an additional 9.7 percent of variance, and the final inclusion of the composite multidimensional self-care index at Step 4 explained an additional 12.9 percent. The final physical health model explained 54.7 percent of the total variance, yielding an adjusted R-squared of 0.541 and a robust F-statistic of 172.95. Within this fully adjusted specification, multidimensional self-care maintained the largest standardized association with physical well-being, showing an unstandardized coefficient of 97.29 and a standardized beta of 0.511. Total loneliness retained a statistically significant, albeit smaller, negative association, exhibiting an unstandardized coefficient of negative 3.10 and a standardized beta of negative 0.085. Advanced age was independently associated with poorer physical health, while regular walking and vacation habits remained positively associated. Mirroring the mental health findings, neither gender nor widowed status retained an independent association with physical functioning after accounting for loneliness and self-care.
To investigate the structural interrelationships among these variables, the authors conducted cross-sectional indirect-effect analyses using 10,000 non-parametric case-resampling bootstrap iterations with a fixed random seed. Higher loneliness was negatively associated with composite self-care, yielding an unstandardized path coefficient of negative 0.094. The estimated indirect statistical association linking loneliness to mental health through self-care was negative 6.91, supported by a 95 percent bootstrap confidence interval spanning negative 8.33 to negative 5.59. For physical health, the indirect association was negative 9.14, with a bootstrap confidence interval from negative 10.94 to negative 7.38. Because the Multidimensional Self-Care Scale contains items that share conceptual proximity with SF-36 health domains, post hoc sensitivity analyses were conducted using restricted self-care scores. Excluding daily well-being and distress from the mental health predictor and functional independence from the physical health predictor attenuated the standardized self-care coefficients to 0.311 and 0.310, respectively, while incremental variance contributions dropped to approximately 5.2 and 5.8 percent. Crucially, both restricted direct and indirect associations remained statistically significant, confirming that conceptual overlap partially inflated the primary estimates but did not wholly account for the observed relationships.
The study’s authors highlighted several critical methodological constraints that govern how these findings must be interpreted. Foremost, the cross-sectional observational design precludes any inference regarding temporal ordering, directional causality, or true biological mediation. While statistical indirect associations were formally supported, identical data structures could reflect reverse causation, wherein deteriorating physical or emotional health undermines an individual’s ability to maintain self-care routines or precipitates social withdrawal. Additionally, all measures relied exclusively upon subjective self-reports, introducing potential vulnerability to common-method variance, positive or negative affectivity, and recall bias in the absence of objective clinical records or physical performance metrics. The community convenience sample may also have underrepresented severely impaired, cognitively compromised, or institutionalized older adults who could not access municipal centers. Nevertheless, the research provides compelling evidence that clinical assessments and public health initiatives must move beyond crude marital categories. Supporting healthy aging requires nuanced evaluations of subjective relational loneliness and comprehensive, person-centered support that bolsters an individual’s functional, psychological, and preventive self-care practices in their everyday environment.
Subject of Research: Associations of loneliness, multidimensional self-care, and widowhood with health-related quality of life among older adults
Article Title: Associations of loneliness, multidimensional self-care, and widowhood with health-related quality of life among older adults in Greece
Article References: Nakou, A., Dragioti, E., Mantzoukas, S., Tsironis, C., Bartzou, E., Zagorianakou, N., & Gouva, M. (2026). Associations of loneliness, multidimensional self-care, and widowhood with health-related quality of life among older adults in Greece. Discover Public Health, 23(1), Article 1691. https://doi.org/10.1186/s12982-026-02809-x
Image Credits: AI Generated
DOI: 10.1186/s12982-026-02809-x
Keywords: loneliness, multidimensional self-care, widowhood, health-related quality of life, older adults, Greece, SF-36, De Jong Gierveld Loneliness Scale, MSCS-OA, gerontology, cross-sectional study, hierarchical regression, bootstrap mediation, mental health, physical functioning, biopsychosocial model
