Chinese-language studies were searched in the China National Knowledge Infrastructure (CNKI), Wanfang Data, and VIP Database. English-language studies were searched in Web of Science, PubMed, PsycINFO, and Scopus. All databases were searched from inception to March 2026.

The Chinese databases were used to identify studies published in Chinese that examined older adults in Chinese mainland. The English-language databases were searched to identify studies of the same population that were published in English. Thus, the English-language search was intended to capture internationally published research on older adults in Chinese mainland rather than to extend the review to populations from other countries or regions.

Searches combined controlled vocabulary, where available, with free-text terms, and the search fields and syntax were adapted to the requirements of each database. The search terms were organized into three main concept groups. The first group described the target population and included terms such as “老年人”, “高龄老人”, “older adults”, “elderly”, and “aging”. The second group described the outcomes of interest and included “焦虑”, “抑郁”, “焦虑抑郁共病”, “anxiety”, “depression”, and related terms. The third group described potential correlates and included terms such as “风险因素”, “保护因素”, “社会支持”, “孤独”, “慢性病”, “衰弱”, “risk factors”, “associated factors”, “protective factors”, “social support”, and “frailty”. Terms within each concept group were combined using the Boolean operator “OR”, whereas the three concept groups were combined using “AND”.

All retrieved records were imported into Zotero and deduplicated before screening. The reference lists of included studies and relevant publications were also examined to identify potentially eligible studies not captured by the database searches. The design and reporting of the search strategy followed PRISMA-S recommendations to enhance the completeness, transparency, and reproducibility of the literature search [11].

Studies were eligible for inclusion when they met all of the following criteria. First, participants were adults aged 60 years or older residing in Chinese mainland. When a study included participants from Chinese mainland and other countries or regions, data for the Chinese mainland sample had to be reported separately or be independently extractable. Second, the study examined anxiety symptoms, depressive symptoms, or their co-occurrence using a reported screening instrument or symptom scale. Third, it reported the measurement instrument, prevalence or detection rate, or factors associated with either symptom outcome. Fourth, it was an empirical study employing a cross-sectional, longitudinal, cohort, or secondary-data design, or it reported eligible baseline data from an intervention study. Analyses based on large publicly available datasets were also eligible. Fifth, the article was published in Chinese or English.

Studies were excluded when they met any of the following criteria: 1) they were reviews, commentaries, conference abstracts, theoretical articles, or other nonempirical publications; 2) the participants were not older adults, or findings for older adults could not be separately extracted from a mixed-age sample; 3) the sample consisted exclusively of patients with a confirmed clinical diagnosis and therefore did not reflect mental health symptoms and associated factors in the broader older population; 4) the measurement instrument, classification criterion, or principal findings for anxiety or depression were insufficiently described; or 5) the study represented a duplicate publication of the same data or the full text could not be obtained. More details can be seen in Figure 1.

Study selection proceeded through title and abstract screening, full-text eligibility assessment, methodological quality appraisal, and final inclusion. Two reviewers independently screened all records against the prespecified eligibility criteria and documented the reasons for excluding articles at the full-text stage. Disagreements were resolved through discussion and consensus; when consensus could not be reached, a third reviewer adjudicated. Independent screening by two reviewers was used to reduce the risk of missed studies and subjective selection bias [12].

Methodological quality was evaluated using design-specific Joanna Briggs Institute (JBI) critical appraisal checklists: the checklists for analytical cross-sectional studies and prevalence studies for the corresponding cross-sectional designs, the cohort-study checklist for longitudinal or prospective studies, the randomized controlled trial checklist for randomized intervention studies, and the quasi-experimental checklist for nonrandomized intervention designs or eligible baseline studies [13]. Each applicable item was rated “Yes”, “No”, “Unclear”, or “Not applicable”. Studies with affirmative ratings on fewer than 50% of applicable

Figure 1.PRISMA flow diagram.

items were classified as low quality and failed appraisal; studies with 50% - 74% affirmative ratings were classified as moderate quality, and those with at least 75% were classified as high quality. Fourteen studies that failed the prespecified threshold were excluded. Among retained studies, appraisal results did not determine numerical weights, but high-quality, longitudinal, and representative studies received greater interpretive emphasis, whereas moderate-quality findings were described more cautiously.

The combined database searches yielded 1269 records. After 77 duplicates were removed, 1192 unique records underwent title, abstract, and keyword screening. At this stage, 683 records were excluded: 35 were nonempirical or secondary publications, 510 did not focus on anxiety or depression among older adults, 133 did not focus on older adults in Chinese mainland, and 5 included mixed-age samples for which findings concerning adults aged 60 years or older could not be separately extracted. This left 509 records. Eighteen full texts could not be obtained, and 491 reports underwent full-text assessment. A further 205 reports were excluded because the age criterion was ambiguous or mixed-age findings were not extractable (n = 68), the sample was restricted to clinically diagnosed cases without eligible associated-factor analyses (n = 80), or key data were insufficient (n = 57). Of the 286 reports entering quality appraisal, 14 failed the prespecified quality criterion, resulting in 272 included reports. The anxiety-specific search subset yielded 393 records, 387 after duplicate removal, and 101 included reports; these 101 reports form part of, rather than an additional set beyond, the overall total of 272. The remaining 171 reports constituted the depression evidence set.

Data from the included studies were coded using a standardized extraction form that had been developed in advance, pilot-tested on a small subset of studies, and revised accordingly. The extracted information included authorship, year of publication, study location, sample size, participant age, gender composition, urban or rural setting, sampling source, study design, instruments used to assess anxiety or depression, statistical methods, principal associated factors, direction of association, and reported effect estimates. One reviewer extracted the data, and a second reviewer verified each item against the original publication. Disagreements were resolved through discussion or, when necessary, adjudication by a third reviewer. Standardized extraction and independent verification were used to reduce transcription errors and interpretive bias [12].

Variables were coded using a four-level structure comprising specific variable, variable domain, direction of association, and model dimension. First, specific variables reported in the original studies—such as chronic illness, living alone, loneliness, social support, psychological resilience, and sleep disturbance—were extracted. Second, these variables were grouped into broader domains, including sociodemographic characteristics, physical health, psychological resources, family and social relationships, living environment, and major life events. Third, the direction and statistical evidence of each variable’s association with anxiety or depressive symptoms were recorded. Finally, each construct was assigned one primary model dimension according to its operational definition and principal theoretical role in the reported analysis, rather than according to its label alone.

The preliminary assignment rules were as follows. Risk accumulation included adverse socioeconomic, relational, environmental, and life-event exposures that represent stress burden or resource loss. Functional vulnerability included physical, cognitive, sensory, sleep-related, nutritional, and daily-functioning conditions that reduce autonomy, coping capacity, or access to support. Positive support included familial, social, community, behavioral, and psychological resources operationalized in a beneficial direction. Directionally opposite indicators were not double counted: loneliness and low social participation were coded as risk accumulation, whereas actual participation or social engagement was coded as positive support; sleep disturbance was coded as functional vulnerability, whereas a sleep-management intervention could be coded as positive support; and functional impairment was coded as functional vulnerability even when an environmental barrier contributed to it. A study could contribute distinct variables to more than one dimension, but each construct-outcome estimate was counted once for a given outcome and time point. Secondary or cross-dimensional roles were discussed qualitatively rather than entered again in frequency summaries. Two reviewers independently coded and categorized the variables, and discrepancies were resolved by consensus.

Because the included studies differed substantially in sampling sources, measurement instruments, variable definitions, and statistical indices, quantitative pooling of effect sizes was considered inappropriate. The evidence was therefore synthesized using narrative synthesis and variable mapping [14].

The synthesis involved four steps. First, factors repeatedly associated with anxiety or depression were identified across the included studies. Second, the consistency of the direction of associations was examined across mental health outcomes, study designs, and sample contexts. Third, factors common to anxiety and depression were identified by jointly considering methodological quality, replication across studies, sample representativeness, and theoretical plausibility. Fourth, the resulting factors were mapped onto three proposed dimensions—risk accumulation, functional vulnerability, and positive support—to generate an initial conceptual model.

Evidence strength was not determined solely by counting the number of statistically significant findings. Instead, the interpretation considered study design, sample size and representativeness, measurement quality, adjustment for potential confounders, magnitude and direction of associations, replication across independent samples, and consistency across studies. Evidence statements were classified as: 1) direct association evidence, when a factor was statistically associated with anxiety or depressive symptoms; 2) temporal or pathway evidence, when longitudinal ordering, mediation, moderation, or interaction was formally tested; or 3) a model-derived hypothesis, when an amplifying, buffering, or dynamic role was theoretically inferred but not directly tested. Greater interpretive weight was given to findings supported by methodologically stronger studies, longitudinal evidence, nationally or regionally representative samples, and repeated observations across different settings. Inconsistent or isolated findings were retained when theoretically relevant but were interpreted more cautiously.

Within the cited evidence base, longitudinal or prospective ordering was examined in studies of social participation, loneliness and social isolation, cognitive frailty, and psychosocial profiles [5][8][9][15], whereas formal mediation or moderation was examined for perceived isolation, social capital, and coping or aging perceptions [16]-[18]. These studies support selected temporal or pathway propositions. However, the broader claims that functional vulnerability amplifies accumulated risk and that positive support buffers this amplification remain hypotheses requiring direct interaction tests and independent validation.

The conceptual model was developed iteratively from the synthesized evidence. Factors reflecting cumulative social, economic, interpersonal, environmental, or life-event exposures were classified under risk accumulation. Physical, cognitive, and functional conditions that could reduce independence, coping capacity, or participation and thereby intensify the effects of other risks were classified under functional vulnerability. Psychological, familial, social, behavioral, and community resources that could buffer risk or provide feasible targets for intervention were classified under positive support. These dimensions were then integrated in light of age-related changes in physical capacity, cognitive functioning, social roles, and support needs to construct a model for identifying mental health problems among older adults.

The eligibility assessment, quality appraisal, variable coding, and frequency summaries in this review were restricted to studies of older adults in Chinese mainland. To explain the potential mechanisms underlying the identified factors and assess whether findings from China were consistent with the broader international literature, selected international systematic reviews, meta-analyses, and longitudinal studies were additionally cited in the model-development and discussion sections. These international publications were not included in the number of studies reviewed, the methodological quality appraisal, or the frequency counts of coded variables. They were used solely to support theoretical interpretation and provide external corroboration.

A total of 272 study reports were included, comprising 101 reports in the anxiety evidence set and 171 reports in the depression evidence set. The studies were published between 2010 and 2025, with sample sizes ranging from 86 to 67,817 participants. Of the included studies, 173 employed cross-sectional designs (63.60%), 45 used longitudinal or prospective cohort designs (16.54%), 42 analyzed baseline data from intervention studies (15.44%), and 12 adopted other empirical designs (4.41%). Thus, the available evidence was predominantly cross-sectional. Regarding sample sources, 57 studies recruited community-dwelling older adults, 43 focused on residents of long-term care institutions, and 172 recruited participants from hospitals or primary healthcare settings.

Anxiety was most commonly assessed using the Generalized Anxiety Disorder scale, the Self-Rating Anxiety Scale, the anxiety subscale of the Hospital Anxiety and Depression Scale, and other screening instruments, with the Generalized Anxiety Disorder scale being used relatively frequently. Depression was primarily assessed using the Geriatric Depression Scale, the Center for Epidemiologic Studies Depression Scale, the Patient Health Questionnaire, and related instruments, among which the Geriatric Depression Scale was particularly common.

Variable mapping indicated that the factors examined in the included studies were distributed across several broad domains: sociodemographic characteristics, physical health and functional status, psychological resources, family and social relationships, living environments, and major life events. Multimorbidity, functional limitations, and sleep disturbances were reported in 67 studies; loneliness and low social support were examined in 55 studies; and social participation and family support were reported in 87 studies. Overall, multimorbidity, functional impairment, loneliness, and inadequate social support showed relatively consistent associations with anxiety or depression. In contrast, findings concerning age, urban-rural residence, marital status, and several lifestyle factors were less consistent, potentially because of differences in sample sources, measurement instruments, regional contexts, and adjustment for confounding variables.

Risk accumulation encompassed socioeconomic disadvantages, loss of relational resources, subjective psychological distress, adverse living environments, and major life events. Studies of older adults in China repeatedly identified female gender, lower educational attainment, financial hardship, widowhood or unmarried status, living alone, and limited social resources as correlates of elevated anxiety, depression, or anxiety-depression comorbidity. Gender, marital status, economic circumstances, chronic illness, and lifestyle factors often jointly constituted the broader risk context in which mental health problems emerged [6][19].

International evidence has produced similar findings. A systematic review and meta-analysis by Cole and Dendukuri [20] identified female gender, bereavement, sleep disturbance, and functional disability as important predictors of depression among community-dwelling older adults. A Dutch cohort study further suggested that the association between low income and depression may be stronger in later life than in younger age groups [21]. These findings indicate that sociodemographic characteristics are not merely descriptive variables but may also reflect prolonged resource deprivation and cumulative exposure to stress.

Psychological factors such as loneliness, negative attitudes toward aging, health-related concerns, and low life satisfaction may represent proximal signals through which accumulated adversity is translated into psychological symptoms. Longitudinal research has shown that older adults experiencing persistent loneliness or sustained social isolation tend to maintain higher levels of depressive symptoms over time [8]. A longitudinal study of community-dwelling older adults in Shanghai similarly found that persistently high loneliness, social isolation, low social support, and low social capital—as well as transitions from favorable to adverse psychosocial conditions—were associated with more severe subsequent symptoms of depression, anxiety, and stress [15].

International longitudinal evidence further indicates that loneliness predicts depressive symptoms even after adjustment for age, income, marital status, and physical health [22]. Loneliness may also have reciprocal relationships with major depression and generalized anxiety, although its predictive association with subsequent mental health problems appears to be comparatively stronger [23]. Loneliness may therefore function not only as a correlate of poor mental health but also as an early indicator that accumulated social and functional risks are beginning to manifest psychologically.

Deficits in family and social relationships constitute another major source of accumulated risk. The quality of spousal and parent-child relationships, frequency of contact with adult children, interaction with friends, neighborhood reciprocity, social capital, and participation in social activities have all shown direct associations with anxiety or depressive symptoms in later life. Chinese longitudinal studies further indicate that older adults with limited social participation are more likely to experience subsequent depressive symptoms than those who participate regularly in social activities [9]. Insufficient family and social support is therefore a plausible moderator of the psychological consequences of chronic illness, pain, and daily-life difficulties, but direct interaction tests of this buffering or amplifying proposition remain limited in the reviewed evidence [19].

International studies similarly suggest that reduced social connectedness may increase anxiety and depressive symptoms by heightening perceived social isolation [18]. Smaller social networks and greater loneliness have also been prospectively associated with anxiety and depression [23]. Importantly, objective social isolation and subjective loneliness are related but conceptually distinct. An older adult may maintain a relatively large number of social contacts yet remain psychologically vulnerable when those relationships lack emotional closeness, stability, or accessible support.

Residential and neighborhood conditions provide the external context in which risks accumulate. Poor housing, neighborhood disorder, transportation barriers, insufficient public activity spaces, and limited access to healthcare and long-term care services may adversely affect mental health by restricting physical activity, reducing social interaction, and weakening community belonging. Conversely, supportive neighborhood environments and higher community social capital may promote psychological well-being [17][24].

An international systematic review and meta-analysis likewise found that neighborhood safety, social cohesion, service accessibility, walkability, and green space may be associated with depression in later life. However, these environmental effects may vary according to physical mobility, socioeconomic position, and urban-rural context [25]. Environmental disadvantage should therefore be understood not as an isolated exposure but as a contextual condition that may interact with individual functional capacity and access to social resources.

Major life events may trigger the transition from latent vulnerability to more overt psychological problems. Relevant events include bereavement, the loss of an only child, diagnosis of a serious illness, hospitalization or surgery, residential relocation, substantial financial loss, changes in caregiving relationships, and public health emergencies. Evidence from a nationally representative sample of Chinese adults indicates that widowhood is associated with higher depressive symptoms and that this association may be mediated or moderated by coping styles and perceptions of aging [16].

International research has similarly identified bereavement as a relatively consistent risk factor for late-life depression [20]. The psychological consequences of widowhood may nevertheless differ according to gender, pre-bereavement marital quality, caregiving experiences, and regional welfare conditions [26]. The effects of major life events therefore depend not only on the events themselves but also on older adults’ pre-existing risk burden, functional status, coping capacity, and available support.

Physical illness and functional decline represent the most age-specific group of factors associated with mental health problems in later life. The included studies examined chronic illness, multimorbidity, pain, sleep disturbance, frailty, sarcopenia, disability, limitations in activities of daily living, cognitive decline, visual and hearing impairment, malnutrition, and dependence on care.

Research among older adults in China consistently showed that multimorbidity, impaired daily functioning, and cognitive frailty were associated with elevated anxiety and depression. Analyses based on the China Health and Retirement Longitudinal Study indicated that multimorbidity was associated with an increased likelihood of depression [2]. Physical impairment and limitations in activities of daily living were also related to more severe depressive symptoms [4]. Prospective evidence further showed that older adults with cognitive frailty had a higher subsequent risk of depression than those with physical frailty or cognitive impairment alone [5].

International evidence supports a close relationship between physical and cognitive decline and mental health problems. A systematic review and meta-analysis by Soysal et al. [27] found a bidirectional association between depression and frailty: frailty predicted subsequent depression, whereas depression was also associated with the later development of frailty. A systematic review by Tan et al. [28] further showed that anxiety was more prevalent or severe among frail older adults than among their physically robust counterparts. Earlier meta-analytic evidence also identified disability, sleep disturbance, and chronic health problems as important predictors of late-life depression [20].

Functional limitations may provide a key pathway through which physical illness is translated into psychological distress. Chronic disease, pain, sleep disturbance, and sensory impairment may directly increase worry, fatigue, helplessness, and loss of control. They may also reduce independence, physical activity, social participation, and self-care, thereby weakening self-efficacy and coping capacity [4]. The latter sequence is theoretically plausible, but it should be treated as a pathway hypothesis unless mediation, moderation, or interaction is formally tested. The coexistence of physical frailty and cognitive decline may create a particularly pronounced state of vulnerability [5].

Sleep disturbance may occupy a similarly complex position. It can arise from illness, pain, medication use, caregiving burden, or emotional distress, while also exacerbating anxiety and depression through fatigue, impaired emotion regulation, and reduced cognitive efficiency. Functional conditions may therefore both directly contribute to mental health problems and increase sensitivity to social, economic, and environmental stressors.

On the basis of this evidence, chronic illness, multimorbidity, frailty, disability, cognitive decline, sleep disturbance, pain, malnutrition, sensory impairment, and care dependence were classified as indicators of functional vulnerability. Their direct associations with anxiety or depressive symptoms are better supported than the proposed amplification mechanism. The model therefore hypothesizes that these conditions may intensify the effects of other risks by reducing autonomy, social participation, coping capacity, and the ability to access and use supportive resources; this hypothesis requires direct interaction and longitudinal testing.

Positive support included family support, broader social support, interaction with friends and neighbors, community services, social participation, regular physical activity, health-promoting behaviors, psychological resilience, positive attitudes toward aging, subjective well-being, life satisfaction, and adaptive coping. Compared with relatively fixed or difficult-to-modify risk factors, these resources may both reduce the likelihood or persistence of mental health problems and provide feasible targets for intervention.

Family and social support can provide emotional companionship, practical care, financial assistance, health information, and early recognition of psychological crises. Lack of social participation and support has been associated with greater anxiety and depression, whereas stable social relationships and regular participation in social activities appear to be protective [9][18][19].

The protective value of a social network depends not only on its size but also on its emotional closeness, stability, reciprocity, and practical availability. Social and community activities may strengthen connectedness and reduce both objective isolation and subjective loneliness. Community health management, day-care programs, accessible medical services, and neighborhood mutual assistance may be especially important for older adults with functional limitations, for whom informal social participation is otherwise difficult.

Residential environments and community resources may also function as positive supports. Adequate housing, safe and accessible walking environments, public activity spaces, green areas, healthcare and long-term care services, and strong neighborhood social capital may promote physical activity, social interaction, and community belonging [17][24]. However, the psychological benefits of neighborhood resources may vary according to physical functioning, socioeconomic status, and urban-rural context [25]. Environmental resources can produce meaningful protection only when older adults are able to access, use, and remain engaged with them.

Psychological resilience, positive attitudes toward aging, life satisfaction, subjective well-being, and adaptive coping constitute internal supportive resources. These factors may help older adults maintain a sense of control, hope, continuity, and meaning when facing illness, functional decline, or major life events. They may thereby reduce the likelihood that stress is translated into clinically relevant anxiety or depressive symptoms.

Positive support should therefore be conceptualized not only as a group of protective correlates but also as a set of actionable intervention resources. Potential strategies include psychological education, cognitive reframing, meaning-centered approaches, resilience enhancement, promotion of physical and social activity, strengthening of family relationships, and group- or community-based interventions.

The reviewed evidence indicates that elevated anxiety and depressive symptoms in later life are associated with factors from multiple domains rather than any single source (More details can be seen in Table 1). The strongest evidence concerns direct associations among socioeconomic disadvantage, relational and environmental adversity, physical and cognitive decline, supportive resources, and symptom burden. Claims about cumulative interaction, amplification, and buffering are less directly established and are therefore specified as hypotheses within the proposed model.

Table 1.Evidence mapping of factors associated with mental health problems in older adults.

The dimensions are analytically distinct, and each construct received one primary assignment for frequency mapping. Overlapping constructs were handled according to operationalization and direction. Loneliness was coded as an adverse subjective social state under risk accumulation; social participation was coded as positive support when measured as actual engagement, but low or absent participation was coded as accumulated relational risk. Sleep disturbance and functional impairment were coded under functional vulnerability because they reflect reduced regulatory or daily-functioning capacity. Environmental barriers that contributed to impairment remained environmental risks, and interventions designed to improve sleep or functioning could be discussed as positive supports without recoding the original adverse construct. This rule prevented the same construct-outcome estimate from being counted twice while allowing secondary roles to be discussed conceptually.

Across the included evidence, multimorbidity, functional limitations, sleep disturbance, loneliness, and inadequate social support showed relatively consistent associations with anxiety or depression. In contrast, the effects of age, urban-rural residence, marital status, and some lifestyle factors varied across contexts. These inconsistencies may reflect differences in population characteristics, measurement methods, regional resources, model specification, and confounder adjustment rather than the complete absence of an association.

Considering the frequency with which variables were examined, the direction and consistency of their direct associations, study design, sample representativeness, methodological quality, and theoretical relevance, the identified factors were mapped onto three overarching dimensions: risk accumulation, functional vulnerability, and positive support. The three-dimensional structure provides the empirical basis for the model, whereas the proposed amplification, buffering, and transition processes constitute hypotheses for subsequent validation.

Risk accumulation refers to the progressive clustering and mutual reinforcement of adverse conditions across social, relational, psychological, and environmental domains. Key factors include financial hardship, living alone or in an empty-nest household, widowhood, loss of an only child, social isolation, loneliness, limited social participation, adverse living environments, and major life events. Female gender, bereavement, low income, sleep disturbance, functional limitations, and chronic illness frequently form the broader context of risk for late-life depression [20][21]. Socioeconomic disadvantage, multimorbidity, and limited social participation are likewise closely associated with anxiety, depression, and their comorbidity among older adults in China [2][6][9].

Risk accumulation is not simply the additive effect of multiple risk variables. Financial hardship may limit healthcare access and social participation; chronic illness may reduce mobility and interpersonal contact; living alone may increase loneliness; and loneliness may worsen sleep and emotional functioning. Longitudinal studies show that loneliness, social isolation, and shrinking social networks precede or accompany later anxiety and depressive symptoms [8][22][23]. These findings support temporal ordering for selected factors, but synergistic accumulation across domains remains a hypothesis until additive and interactive risk models are directly compared.

Major life events may act as triggers within this cumulative process. Widowhood, serious illness, hospitalization, residential relocation, and caregiving changes can disrupt established routines and social structures. Formal mediation and moderation evidence indicates that coping strategies and perceptions of aging may help explain heterogeneity in the association between widowhood and depression [16]. Nevertheless, the broader triggering proposition should be tested prospectively by examining event timing, pre-existing risk burden, functional status, and available support.

Functional vulnerability refers to reduced physical and psychological adaptive capacity resulting from aging, illness, and functional decline. It includes chronic illness, multimorbidity, frailty, disability, cognitive decline, sleep disturbance, pain, malnutrition, sensory impairment, and dependence on care. This dimension distinguishes the present model from more general frameworks for identifying psychological crises. Existing evidence indicates that multimorbidity, limitations in activities of daily living, and cognitive frailty are associated with a higher risk of depression [2][4][5].

Functional vulnerability may operate through both direct and amplifying pathways. Chronic illness, pain, insomnia, and mobility limitations show direct associations with anxiety or depressive symptoms and may increase worry, fatigue, helplessness, and perceived loss of control. Functional decline may also restrict mobility, physical activity, social participation, and self-care, making it more difficult to obtain or use external support. Prospective and meta-analytic evidence supports temporal and reciprocal links between frailty, cognitive frailty, anxiety, and depression [5][27][28], but these studies do not by themselves establish that vulnerability statistically amplifies external risks.

Functional vulnerability is therefore treated as an independent model dimension with a testable amplifying hypothesis. Similar levels of financial hardship, bereavement, or social isolation may have different psychological consequences for older adults with preserved versus impaired functioning. This proposition requires direct tests of interaction, moderated mediation, and longitudinal change; until such evidence is available, the model should not be interpreted as demonstrating causal amplification.

Positive support refers to the system of resources that helps maintain mental health, buffer the effects of risk, and promote adaptive functioning. It includes family support, social support, community services, social participation, regular physical activity, health-promoting behaviors, psychological resilience, positive attitudes toward aging, and life satisfaction. Most reviewed evidence demonstrates direct protective associations; a buffering effect should be inferred only when a study formally tests interaction, mediation, moderation, or intervention-related change.

Three candidate pathways are proposed. First, family and social support may provide companionship, caregiving assistance, and practical help. Second, participation and community services may strengthen connectedness and reduce objective isolation and subjective loneliness. Formal mediation evidence shows that perceived isolation can link social disconnectedness with later anxiety and depressive symptoms [18], and social capital has been examined as a mediator between neighborhood environment and mental well-being [17]. Third, resilience, positive attitudes toward aging, life satisfaction, and adaptive coping may help preserve control, hope, and meaning. These findings support selected pathways but do not establish a universal buffering mechanism across all support variables.

The model hypothesizes that supportive resources may weaken links between risk accumulation, functional vulnerability, and symptom burden. Family care and community services may reduce the daily burden associated with functional limitations; social participation may interrupt the progression from living alone to loneliness; and adaptive coping may lessen the impact of major life events. These are intervention-relevant hypotheses, not established causal conclusions, and should be tested using interaction models, longitudinal mediation, and randomized or quasi-experimental designs.

Risk accumulation, functional vulnerability, and positive support are expected to vary over time as diseases progress, physical and cognitive functioning changes, major life events occur, and supportive resources increase or decline. Longitudinal studies show that changes in loneliness, social isolation, social networks, and functional status are associated with subsequent anxiety or depressive symptoms [5][8][23]. However, the reviewed studies did not directly validate transitions among the three proposed dimensions; the following stage-transition account is therefore conceptual and requires prospective testing.

Conceptually, older adults may be more likely to remain psychologically stable when cumulative risk is limited, functioning is relatively preserved, and support is available. As financial hardship, living alone, social isolation, and stressful life events cluster while support declines, loneliness, sleep disturbance, and mild anxiety or depressive symptoms may emerge. Additional functional losses or acute events may correspond to a higher-risk state, whereas stronger family support, timely community services, renewed participation, or improved functioning may correspond to movement toward lower risk. These proposed states and transitions should be empirically derived rather than assumed.

The model therefore recommends repeated, multidimensional risk assessment rather than reliance on a one-time symptom screen. Assessment may jointly consider current anxiety and depressive symptoms, cumulative risk exposure, functional vulnerability, supportive resources, and recent major life events. Such information may support risk stratification and referral decisions, but it does not constitute a clinical diagnosis and should not be used to assign deterministic risk stages before validation.

The practical value of the risk accumulation-functional vulnerability-positive support model lies in organizing multidimensional information for stratified services. Symptom-based mental health identification in later life should not rely solely on one-time self-report screening. Instead, validated symptom measures may be considered alongside community health records, chronic disease management data, family caregiving information, social participation, and primary-care service records. This integrated assessment is intended for screening and service planning, not for replacing diagnostic evaluation by qualified professionals.

Such an approach requires collaboration among primary care physicians, mental health professionals, social workers, and family caregivers. In a review of 38 studies covering 13 integrated care models for older adults with depression and physical multimorbidity, Tops et al. [30] identified multidisciplinary teams, comprehensive assessment, individualized care plans, proactive follow-up, continuity of care, and cross-sector coordination as recurring components of effective service delivery. These findings support the development of information-sharing and coordinated care mechanisms for older adults at psychological risk.

On the basis of multisource data, provisional indices of risk accumulation, functional vulnerability, and positive support could be developed, while latent profile analysis could explore distinct combinations of risk and resources. Machine-learning methods such as random forests and XGBoost may estimate variable importance, nonlinear associations, and potential thresholds, with SHAP values improving interpretability. Using longitudinal data from 2331 middle-aged and older adults, Xia et al. [31] found that XGBoost performed relatively well in predicting depression two years later. Nevertheless, such models should be externally validated and used for symptom-risk prediction and service decision support rather than diagnosis or causal inference.

The ultimate purpose of the model is not to label or diagnose individuals but to match provisional risk profiles with proportionate support. Older adults with low and stable symptom risk may benefit from mental health education, community activities, and healthy-lifestyle guidance. Those experiencing loneliness, sleep problems, or mild anxiety or depressive symptoms may require social-connection programs, sleep management, group activities, and low-intensity psychological support [32]. Individuals with substantial cumulative risk, pronounced functional vulnerability, and inadequate support may need coordinated follow-up by primary care, social work, and mental health professionals. Severe, persistent, or co-occurring symptoms, particularly when accompanied by disability, cognitive decline, major life events, or extreme isolation, should prompt professional assessment and appropriate medical or mental health referral.

Group-based arts interventions may provide another feasible pathway for translating the model into practice. The term group art therapy should be reserved for interventions delivered by professionally trained therapists according to explicit therapeutic goals, whereas broader programs involving music, dance, painting, crafts, drama, or creative writing are more appropriately described as group-based arts interventions. Such activities provide opportunities for emotional expression, shared participation, and social connection. Because social disengagement may exacerbate anxiety and depression through increased perceived isolation, strengthening group belonging and interpersonal interaction has a clear theoretical rationale [18].

A systematic review and meta-analysis of 39 controlled studies found that group-based arts interventions were associated with moderate reductions in anxiety and depressive symptoms among older adults [33]. The effect on depression remained after adjustment for publication bias and was larger in residential care settings than in community settings. Both professionally delivered art therapy and general participatory arts activities appeared beneficial, with no consistent evidence that music, dance, or visual arts was superior to the others. Group-based arts interventions may therefore simultaneously provide psychological support, social participation, and community engagement, making them especially suitable for older adults with emerging or moderately elevated risk. However, substantial heterogeneity and possible publication bias indicate the need for further randomized controlled trials examining intervention format, dosage, mechanisms, long-term effects, and differential effectiveness across risk profiles.

The principal contribution of the proposed model is its extension of mental health identification from symptom-based screening to an integrative assessment of risk accumulation, functional vulnerability, and positive support. Conventional screening can identify current anxiety or depressive symptoms but is less informative about the contexts in which symptoms arise or the resources that may support intervention. The present model therefore asks where associated risks originate, which aging-related conditions may increase susceptibility, and which resources may be actionable. These questions extend screening and service planning but do not convert correlational evidence into clinical diagnosis or causal explanation.

Theoretically, the model integrates factors previously examined separately across medicine, psychology, sociology, and public health. The risk accumulation dimension captures adverse financial, relational, environmental, and life-event exposures; the positive support dimension captures family relationships, social participation, community services, health-promoting behaviors, psychological resilience, and other adaptive resources. The primary extension is the inclusion of functional vulnerability as an independent age-specific dimension. The review supports direct associations between many functional conditions and anxiety or depressive symptoms. It does not yet establish that these conditions amplify external adversity or weaken protective resources; those cross-dimensional mechanisms are explicit hypotheses derived from the integrated framework.

Practically, the model may facilitate a shift from reactive treatment after symptom onset toward earlier identification, warning, and dynamic monitoring. It also supports movement from uniform service provision to stratified intervention and from an exclusive focus on individual symptoms to the simultaneous improvement of family relationships, community environments, and opportunities for social participation. Primary care and community practitioners could assess whether an older adult’s principal difficulty involves excessive risk exposure, pronounced functional vulnerability, insufficient support, or a combination of these conditions, and then select appropriate pathways involving health management, family care, social support, psychological intervention, or medical referral. The framework may thus inform community screening, priority-population management, service allocation, and healthy-aging policy.

Several limitations should be acknowledged. First, the model was developed through narrative synthesis and theoretical integration and has not been validated in an independent sample. Its construct validity, predictive accuracy, classification thresholds, and added value beyond symptom screening remain uncertain. Second, the evidence base was predominantly cross-sectional. Direct associations are therefore better supported than causal claims about accumulation, amplification, buffering, or dynamic transitions; future studies should test these mechanisms using prospective interaction models, longitudinal mediation, and experimental or intervention designs. Third, substantial heterogeneity in samples, measures, variable definitions, and statistical methods limits conclusions about temporal ordering and effect magnitude. Fourth, anxiety and depressive symptoms were the primary outcomes, while suicide risk, sleep problems, cognitive impairment, and quality of life were not fully incorporated as outcomes. Fifth, because the review focused on Chinese mainland, cross-cultural generalizability remains unknown.