Multidimensional Factors and Predictors of Vulnerability and Frailty in Older Adults: A Systematic Review.

Multidimensional Factors and Predictors of Vulnerability and Frailty in Older Adults: A Systematic Review

Abstract

Highlights

What are the main findings?

Mapped candidate vulnerability determinants across five interrelated biopsychosocial and self-care domains.

Addressed predictor–outcome construct overlap (incorporation bias) to distinguish operational diagnostic criteria from independent prognostic drivers.

What are the implications of the main findings?

Identified candidate variables to inform future predictive modeling and risk screening.

Abstract

Background: Population ageing is characterized by rising multimorbidity and multifaceted vulnerability. While vulnerability is inherently a biopsychosocial state, clinical research predominantly operationalizes risk through physical frailty phenotypes, constraining holistic geriatric care. Identifying the interplay across biological, functional, and psychosocial domains is essential for comprehensive risk evaluation. Objectives: To synthesize multidimensional factors associated with vulnerability and frailty in adults aged ≥65, categorizing cross-sectional correlates and prospective longitudinal predictors to inform multidimensional assessment. Methods: Following PRISMA 2020 guidelines, systematic searches were conducted across PubMed, CINAHL, Scopus, Web of Science, and Google Scholar (2020–2025). Independent reviewers performed screening, selection, extraction, and Joanna Briggs Institute (JBI) critical appraisal (PROSPERO: CRD420251039796). Results: Forty studies (22 cross-sectional, 17 cohort, 1 quasi-experimental) comprising 77,229 cumulative participant observations across community, outpatient, inpatient, and residential care settings were synthesized. Thirty-seven candidate determinants were mapped across five domains: biological/clinical, psychological, socioeconomic/relational, lifestyle, and self-care profile. Advanced age, multimorbidity, and polypharmacy were the most frequent clinical factors; prospective cohorts specifically linked high anticholinergic burden and functional dependence in complex activities to incident frailty. Depressive symptoms, cognitive decline, low education, financial strain, and social isolation consistently delineated vulnerable states. Conclusions: Vulnerability in older adults is a multifaceted construct requiring assessment tools that extend beyond isolated physical markers. Rather than a definitive feature set, these identified factors provide an evidence-based pool of candidate variables that may inform future risk-stratification and predictive modeling initiatives, supporting integrated, person-centered care.

Keywords: frailty, vulnerability, predictors, elderly, geriatric assessment, systematic review

1. Introduction

Global demographic shifts indicate that by 2050, one in six people worldwide will be aged 65 or older, surpassing 2 billion, an unprecedented demographic transition reaching critical thresholds in more developed regions [1]. However, increased longevity has not translated into proportional improvements in health spans; instead, ageing populations face a rising burden of multimorbidity, chronic disease, and complex health demands [2]. Addressing this shift requires healthcare systems and social policies to transition from fragmented disease management toward integrated, person-centered models of care [1].

Central to this transition is the conceptual differentiation among related geriatric constructs: intrinsic capacity, frailty, functional disability, and vulnerability. In this review, vulnerability is conceptualized as the overarching, dynamic biopsychosocial state of susceptibility to adverse health-related outcomes, shaped by the cumulative interaction of biological, psychological, socio-economic, and environmental determinants over the life course [3,4,5,6]. Within this hierarchy, intrinsic capacity represents the composite of an individual’s physical and mental reserves [7], whereas frailty denotes a specific biological syndrome characterized by cumulative physiological decline and diminished resistance to stressors [8]. In clinical research, frailty and clinical-functional disability are frequently operationalized as clinical proxies or manifestations of this broader vulnerability state [9,10,11]. Establishing this distinction is vital: while primary studies often focus narrowly on physical frailty scales, a comprehensive understanding of vulnerability demands an evaluation across all underlying domains.

Despite this multidimensional conception, existing clinical practice and epidemiological research remain disproportionately anchored in physical frailty phenotypes [9,10]. Yet, vulnerability extends far beyond biological parameters. Susceptibility to decline is heavily driven by social determinants, including lower educational attainment, financial strain, and reduced social engagement [7,11]. As highlighted by OECD data, high out-of-pocket costs for long-term care place up to 95% of care-dependent older adults at risk of poverty, illustrating how economic precarity directly compounds clinical vulnerability [12]. Conversely, robust family networks and psychosocial resources serve as vital protective factors [13].

Functionality and self-care capacities are inextricably linked to this continuum. Unmet care needs often accelerate functional decline, transitioning individuals from independence into states of severe vulnerability [14,15,16]. Self-care—defined as the intrinsic ability of individuals, families, and communities to promote health, manage illness, and maintain autonomy [17]—encompasses cognitive, physical, psychosocial, demographic, and sociocultural dimensions [18,19,20]. Preserving functional autonomy and self-care performance represents a critical barrier against adverse events such as falls, preventable hospitalizations, and institutionalization [8,21,22].

Currently, there is no universally accepted gold standard for screening or quantifying multidimensional vulnerability, resulting in substantial variation across healthcare settings and scientific disciplines [5,23]. While several systematic reviews have examined predictors of physical frailty [24,25], there is a marked scarcity of syntheses examining the comprehensive spectrum of non-biological, social, and functional factors that underpin global vulnerability across both community and acute healthcare contexts. Moreover, the integration of electronic health records and emergent analytical technologies highlights the need to systematically identify which multi-domain variables warrant clinical prioritization [25].

To address these gaps, this systematic review synthesizes the multidimensional factors associated with vulnerability and frailty in adults aged 65 and older. By examining both cross-sectional correlates and longitudinal predictors across hospital and community settings, this review aims to clarify domain-specific determinants and outline an evidence-based pool of candidate variables to inform comprehensive clinical assessment and future risk-stratification models.

2. Materials and Methods

2.1. Protocol Registration, Published Protocol, and Operational Refinements

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines. The review was prospectively registered in PROSPERO on 02 June 2025 (registration number: CRD420251039796; first submitted 31 May 2025), and the foundational study protocol was subsequently published in full [26].

The core research question, target population (≥65 years), primary focus on multidimensional vulnerability/frailty, and the a priori exclusion of acute environmental shocks (COVID-19, natural disasters) strictly adhered to both the registered PROSPERO protocol and the published conference protocol. In accordance with PRISMA 2020 guidelines (Item 24c), minor operational refinements implemented during the analytical execution of the review—specifically: standardizing eligibility reporting strictly to the PECO framework; establishing explicit numerical JBI quality cut-offs (≥80% High, 60–79% Moderate, <60% Low); and performing an inductive narrative thematic synthesis rather than quantitative meta-analysis due to severe operational heterogeneity across frailty definitions—are transparently reported and justified in the subsequent methodology subsections.

2.2. Research Question

To guide the search strategy, eligibility screening, and narrative synthesis, the research question was structured using the PECO (Population, Exposure, Comparator, Outcome) framework, recommended for observational and prognostic reviews of epidemiological evidence [27,28].

Conceptually, this review distinguishes between multidimensional vulnerability—an overarching biopsychosocial state of diminished physiological reserve and heightened susceptibility to stressors across biological, psychological, and social domains—and frailty, which predominantly serves in the published empirical literature as an operational clinical proxy measured via physical phenotypes, deficit accumulation indices, or multidimensional assessment scales.

The PECO parameters were operationalized across three interconnected evidence streams:

Population (P): Older adults aged 65 years and older residing in community, acute-care (hospital), or institutional/residential aged care settings.

Exposure (E): Multidimensional candidate determinants and risk factors (biological/clinical, psychological, socioeconomic/relational, behavioral/lifestyle, and self-care profile) and, for downstream prognostic analyses, established baseline vulnerability or frailty states.

Comparator (C): Robust, non-frail, or non-vulnerable reference groups; individuals unexposed to specific risk factors; or lower vulnerability/frailty exposure tiers.

Outcomes (O): Defined according to the epidemiological focus: (i) Concurrent frailty/vulnerability states assessed cross-sectionally via validated instruments; (ii) Longitudinal trajectories, including incident frailty, frailty progression, or bidirectional state transitions; and (iii) Downstream adverse clinical endpoints resulting from established frailty/vulnerability (specifically, hospitalization, institutionalization, and mortality).

Primary Research Questions:

“Which multidimensional factors are cross-sectionally associated with, or prospectively predict, vulnerability and frailty onset and progression in older adults aged ≥65 years across community, hospital, and residential settings?”

“To what extent do established vulnerability and frailty states prospectively predict downstream adverse clinical outcomes (hospitalization, institutionalization, and mortality) in this population?”

2.3. Search Strategy

Between 2 and 5 June 2025, a comprehensive and systematic bibliographic search was executed and records were exported across scientific databases: PubMed/MEDLINE (2 June 2025), Scopus (3 June 2025), Web of Science (4 June 2025), and CINAHL (4 June 2025), followed by Google Scholar on 5 June 2025. These execution dates were strictly distinguished from the publication date filters applied to the records: all database searches were restricted to publications from 2020 onwards, with CINAHL and Web of Science applying an exact publication cutoff up to 29 May 2025 (reflecting the date of protocol finalization), while PubMed and Scopus applied standard 2020–2025 annual limits (full search syntax details and interface parameters are provided in Table A1, Appendix A). To achieve a targeted and precise search, we employed a combination of specific keywords and Mesh Terms (Medical Subject Headings), using the keywords “Vulnerable Populations”, “Frail Elderly”, “Functionally Impaired Elderly”, “Geriatric Assessment”, “International Classification of Functioning, Disability and Health”, “Health Status Indicators”, “Quality of Life”, “prognosis” and “predictor” and their variations, opportunely combined by Boolean operators. Additionally, a targeted search was conducted in Google Scholar to retrieve non-indexed grey literature, specifically academic dissertations, theses, and institutional reports. Using an advanced title query (allintitle: ((Vulnerable OR FRAIL OR risk) AND aged 65)), the first 100 consecutive records sorted by default algorithmic relevance were screened, in accordance with established methodological guidance for evidence syntheses [29] (full search syntax details are provided in Table A1, Appendix A). Only articles published in English, French, Portuguese, or Spanish were included.

In the initial screening phase, two expert reviewers (HP and AR) independently assessed all titles and abstracts extracted from the electronic database searches. Using the software (https://new.rayyan.ai/) (accessed on 5 June 2025), duplicate and irrelevant records were systematically eliminated. In instances of discordance, a third expert reviewer (AC) was consulted to facilitate consensus. For studies deemed potentially relevant during initial screening, full-text articles were obtained. Each of these was subjected to a rigorous independent assessment by the reviewers (HP and AR) in alignment with predetermined eligibility criteria. In situations where consensus was challenging, dialogues were initiated between the primary reviewers. If no agreement was reached, the adjudication of the third reviewer (CF), previously uninvolved, was sought to ensure an unbiased decision-making process.

2.4. Eligibility Criteria

To ensure methodological transparency and rigor, eligibility criteria were defined a priori using the PECO framework [28], structured into explicit inclusion and exclusion parameters. Inclusion Criteria: Population: Older adults aged ≥65 years residing in community, acute-care, long-term care, or institutional settings. Exposures/Determinants: Studies evaluating at least one multidimensional determinant (e.g., biological, clinical, functional, psychological, cognitive, behavioral, or socioeconomic/relational factors). Outcomes: Vulnerability or frailty assessed using validated clinical instruments, multidimensional indices, or deficit accumulation models. Study Design: Primary quantitative observational and longitudinal empirical studies (prospective or retrospective cohort studies, cross-sectional investigations, and quasi-experimental studies). Language and Period: Peer-reviewed articles published between 2020 and 2025 in English, French, Portuguese, or Spanish. The search was restricted to publications from 2020 onwards to capture the most contemporary epidemiological profiles, reflect recent advancements in multidimensional frailty screening, and align with current clinical practice guidelines.

Exclusion Criteria: Population: Studies focusing exclusively on populations aged <65 years. Furthermore, studies involving mixed-age populations were completely excluded if the data for the older adult cohort (≥65 years) were not disaggregated and reported separately. Operationalization: Studies relying on non-validated screening scales, unstandardized proxy self-reports, or those failing to specify the measured dimension of vulnerability or frailty. Contextual and Stressor Confounders (Acute Disruptions): Studies investigating the transient impact of acute, catastrophic environmental or societal shocks—specifically the SARS-CoV-2 pandemic, natural disasters, or acute armed conflict. While our review conceptualizes vulnerability as encompassing environmental and social contexts, these catastrophic shocks represent acute, non-endemic disruptors that confound baseline, chronic biopsychosocial vulnerability. Excluding acute shock studies isolates persistent, everyday risk profiles relevant to routine clinical practice. Methodological Designs: Secondary syntheses (systematic reviews, meta-analyses), study protocols, commentaries, conference abstracts, and purely psychometric or linguistic validation studies. While studies were required to apply validated instruments to quantify vulnerability, pure psychometric validation studies (e.g., focusing solely on factor structure, test–retest reliability, or translation without analyzing substantive clinical/social predictors or longitudinal outcomes) were excluded.

2.5. Evaluation of Risk of Bias and Methodological Quality of Studies

To maintain methodological rigor, two reviewers (H.P. and A.R.) independently assessed the methodological quality and risk of bias of all eligible studies using the official Joanna Briggs Institute (JBI) Critical Appraisal Checklists: the JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies (8 items), the JBI Critical Appraisal Checklist for Cohort Studies (11 items), and the JBI Critical Appraisal Checklist for Quasi-Experimental Studies (non-randomized experimental studies) (9 items). Discrepancies were resolved through structured consensus discussions, with adjudication by a third senior reviewer (A.C. or C.F.) when required.

Because official Joanna Briggs Institute (JBI) methodological guidance does not prescribe standardized numerical cut-off scores or mandatory exclusion rules [30]—leaving decisions regarding scoring rubrics and threshold criteria to the review team’s discretion—methodological quality was evaluated through a two-step operational refinement to the protocol (as detailed in Section 2.1): 1. Quality Tier Classification: Studies were categorized into explicit quality strata based on the proportion of affirmative (‘yes’) responses relative to applicable criteria: High quality (≥80%), Moderate quality (60–79%), and Low quality (<60%). These tiers were established to weight and contextualize the robustness of narrative findings in the synthesis. 2. Synthesis Eligibility Threshold: To prevent critical reporting deficiencies or high risk of bias from compromising the qualitative synthesis, a minimum retention threshold of ≥60% was defined, establishing that studies categorized as Low quality (<60%) would be excluded from evidence synthesis.

All 40 eligible primary studies achieved scores ≥ 60% (32 classified as High quality and 8 as Moderate quality; no studies scored <60%), allowing the entire eligible corpus to inform the systematic synthesis (Table 1).

Table 1Summary of methodological quality and risk of bias across included studies.

Study JBI Appraisal Tool Final Score Quality Level

Huang et al. (2023) [31] Quasi-Experimental 6/9 (67%) Moderate

Steinmeyer et al. (2020) [32] Cross-sectional 8/8 (100%) High

Venturini et al. (2021) [33] Cross-sectional 6/8 (75%) Moderate

Sharma et al. (2021) [34] Cross-sectional 8/8 (100%) High

Jung et al. (2021) [35] Cross-sectional 8/8 (100%) High

Sella-Weiss et al. (2021) [36] Cross-sectional 8/8 (100%) High

Tavares et al. (2022) [37] Cross-sectional 8/8 (100%) High

Mangin et al. (2023) [38] Cross-sectional 6/8 (75%) Moderate

Yan et al. (2023) [39] Cross-sectional 8/8 (100%) High

Zhu et al. (2023) [40] Cross-sectional 8/8 (100%) High

Shakya et al. (2023) [41] Cross-sectional 8/8 (100%) High

Cömertoğlu et al. (2023) [42] Cross-sectional 7/8 (88%) High

Henry et al. (2023) [43] Cross-sectional 8/8 (100%) High

Alamri & Mealif (2024) [44] Cross-sectional 8/8 (100%) High

Johnson et al. (2024) [45] Cross-sectional 8/8 (100%) High

Sepúlveda et al. (2024) [46] Cross-sectional 8/8 (100%) High

Güner et al. (2024) [47] Cross-sectional 8/8 (100%) High

Tang et al. (2024) [48] Cross-sectional 8/8 (100%) High

Savci et al. (2025) [49] Cross-sectional 8/8 (100%) High

Ceylan et al. (2025) [50] Cross-sectional 8/8 (100%) High

Ortiz-Navarro et al. (2025) [51] Cross-sectional 8/8 (100%) High

Park et al. (2025) [52] Cross-sectional 8/8 (100%) High

Tzeng et al. (2025) [53] Cross-sectional 8/8 (100%) High

Hajek et al. (2020) [54] Cohort 8/11 (73%) Moderate

Lewis (2020) [55] Cohort 7/11 (64%) Moderate

Breij et al. (2020) [56] Cohort 8/11 (73%) Moderate

Varan et al. (2021) [57] Cohort 10/11 (91%) High

Rodríguez-Laso et al. (2021) [58] Cohort 8/11 (73%) Moderate

Tsay et al. (2021) [59] Cohort 9/11 (82%) High

Mielke et al. (2022) [60] Cohort 9/11 (82%) High

Lee et al. (2022) [61] Cohort 11/11 (100%) High

Thompson et al. (2022) [62] Cohort 9/11 (82%) High

Yoshizawa (2022) [63] Cohort 9/11 (82%) High

Rodríguez-Laso et al. (2023) [64] Cohort 9/11 (82%) High

Costenoble et al. (2023) [65] Cohort 9/11 (82%) High

Tchalla et al. (2023) [66] Cohort 9/11 (82%) High

Van der Ploeg et al. (2024) [67] Cohort 8/11 (73%) Moderate

Suprawesta et al. (2024) [68] Cohort 9/11 (82%) High

Tanaka et al. (2025) [69] Cohort 10/11 (91%) High

Dong et al. (2025) [70] Cohort 10/11 (91%) High

2.6. Data Extraction and Synthesis

Data extraction was conducted independently and in duplicate by two reviewers (H.P. and A.R.) using a standardized, pre-piloted electronic data extraction form designed in Microsoft Excel. The extraction template was initially piloted on a random sample of five studies (12.5% of the corpus) to calibrate extraction consistency across reviewers. Extracted variables comprised: study details (first author, publication year, country, healthcare setting); study design; sample size and demographic characteristics; operational definitions and validated tools used to measure vulnerability/frailty; investigated predictor domains; and primary statistical findings.

Discrepancies during extraction were resolved through discussion and re-examination of the primary papers, with persistent disagreements settled by a third senior investigator (C.F. or A.C.). When multiple publications arose from the same longitudinal cohort (e.g., SHARE, HRS, ELSA), we cross-referenced sample sizes, follow-up waves, and reported variables to prevent redundant data entry.

Extracted findings were categorized across three distinct epidemiological roles: (i) factors cross-sectionally associated with frailty or vulnerability states; (ii) prospective predictors of frailty onset, progression, or state transitions in longitudinal cohorts; and (iii) frailty or vulnerability as a baseline predictor of subsequent adverse clinical events (e.g., hospitalization, institutionalization, and mortality).

A narrative, inductive thematic synthesis was conducted to group individual variables into overarching descriptive categories and five analytical domains: biological/clinical, self-care/functional profile, psychological, behavioral, and socioeconomic/relational.

3. Results

3.1. Study Selection

The selection of studies followed a systematic three-stage process: identification, screening, and inclusion. A total of 1281 records were initially identified. This included 1181 records from databases (Scopus: 599; PubMed: 466; CINAHL: 77; Web of Science: 39) and 100 records from grey literature via Google Scholar. Before screening began, 560 duplicate records were removed. The remaining 721 records were screened for relevance. During this phase, 678 records were excluded based on title and abstract, or because the study population was under 65 years of age. This left 43 reports sought for retrieval. Of these, 3 reports could not be retrieved due to a lack of access to the full text. The remaining 40 reports were assessed for eligibility and all met predetermined criteria. Consequently, a total of 40 studies were included in the final systematic review (Figure 1).

3.2. Study Characteristics

All 40 studies were conducted in the following countries: Turkey (n = 5), China (n = 4), Taiwan (n = 3), Spain (n = 3), Australia (n = 3), South Korea (n = 3), Germany (n = 2), France (n = 2), the Netherlands (n = 2), the USA (n = 2), and Japan (n = 2), with additional single-study contributions from Belgium, Brazil, Canada, Chile, Ireland, Israel, Portugal, Saudi Arabia and the United Kingdom. The included studies were geographically diverse, with a notable concentration from Asia (n = 19) and Europe (n = 13). America (n = 5) and Oceania (n = 3) were also represented. Figure 2 shows the geographical distribution of the included studies by country.

The primary characteristics of the included studies—such as study design, setting, participant demographics, outcome measures, and main findings—are summarized in Table 2. Various study designs were used, with cross-sectional studies (n = 22) being the most numerous. Longitudinal/cohort studies (n = 17) were also frequently observed, enabling analysis of vulnerability trajectories and predictors of change over time [54,56,58,60,62,69]. Another [31] was a quasi-experimental observational study with a multidomain intervention (n = 1).

Table 2Characteristics of included studies.

Authors (Year) Country Study Design Study Aim Setting Participants and Sample Outcome Measures Main Findings

Hajek et al. (2020) [54] Germany Cohort Predictors of frailty among individuals aged 85+ Community (Primary care) 486 individuals aged ≥85 (88.7 ± 2.9 years) Clinical Frailty Scale (CFS) Age ↑, dementia ↑ (significant), chronic conditions ↑ (marginally significant)

Lewis (2020) [55] Ireland Cohort Health trajectories and patient-specific factors influence the risk of adverse outcomes (hospitalization, institutionalization, death). Community Virtual Ward 88 older adults (82.8 ± 6.4 years). Health status indicators, disability, and adverse clinical outcomes Cognition and function were key predictors; delirium and multiple care episodes predicted institutionalization

Steinmeyer et al. (2020) [32] France Cross-sectional study To evaluate whether haemoglobin concentration is an independent predictor of frailty, and to investigate causal pathways Hospital 1829 individuals aged ≥65 (82.4± 6.5 years) Fried’s frailty phenotype ↓ haemoglobin concentration, inflammation (CRP), cognitive impairment, malnutrition, and ↓ education level

Breij et al. (2020) [56] Netherlands Cohort study To identify predictors of frailty and vitality in older adults aged 75 years and over, the Longitudinal Ageing Study Amsterdam (LASA) Community-dwelling 569 individuals aged ≥75 (79.4 ± 5.4 years) Frailty and vitality (LASA Frailty Index) Age, depressive symptoms, ↑ chronic diseases, poor self-rated health, cognitive functioning, polypharmacy, and pain

Venturini et al. (2021)[33] Brazil Cross-sectional study To create a multidimensional model to evaluate frailty in older Brazilian adults and to compare the dimensions of the model created among the categories of the physical frailty phenotype (FIBRA-BR). Community-dwelling 3569 participants (73.7 ± 6.6 years). Frailty (Physical frailty phenotype) Poor self-rated health, urinary incontinence, not visiting/receiving visitors, depressive symptoms, falls, feelings of sadness, and cognitive impairment

Varan et al. (2021) [57] Turkey Cohort study (prospective) To compare the predictive value of all comprehensive geriatric assessments Community-dwelling 967 patients aged ≥65 (median 73 years) Edmonton Frail Scale and Fried Frailty Index/Long-term mortality/ Serum albumin level, ↓ functional status (IADL) score, and walking time were identified as the best predictors

Sharma et al. (2021) [34] Australia Cross-sectional To determine whether the Malnutrition Universal Screening Tool (MUST) can predict frailty in older hospitalized patients Hospital 243 patients ≥ 65 (83.9 ± 6.5 years) Frailty (Edmonton Frail Scale) limitations in (ADL), sex, social participation, age, cognitive impairment, ↓ independence in mobility, ↑ comorbidity and polypharmacy

Jung et al. (2021) [35] South Korea Cross-sectional study To assess frailty using gait parameters and LSTM networks Community setting with wearable sensors 74 elderly adults aged ≥65 (76 ± 4 years) Frailty classification (FRAIL scale & Machine Learning) Gait variability, asymmetry, irregularity in step/stride/double-limb support times; demographic factors (age, sex)

Sella-Weiss et al. (2021) [36] Israel Cross-sectional study To explore associations between swallowing function, dysphagia, malnutrition, depression and frailty Community-dwelling 180 individuals aged ≥65 (75.9 ± 7.9 years) Frailty (FRAIL Questionnaire) Age ↑, suspected dysphagia ↑, depression ↑, poor nutrition ↑; worse swallowing performance

Rodríguez-Laso et al. (2021) [58] Spain Cohort study (prospective) To examine the frequency and predictors of transitions between frailty states in community-dwelling older adults Community-dwelling (Toledo Study on Healthy Ageing—TSHA) 1748 individuals aged ≥65 (75.0 ± 5.9 years) Fried phenotype age, smoking, depressive symptoms, hearing/cognitive impairment, and ↓ physical activity

Tsay et al. (2021) [59] USA Cohort study To identify predictors of change in frailty status. Community-based Home and Community-Based Services (HCBS) 16,309 older adults ≥ 65 (80 years) Clinically meaningful changes in Frailty Index (FI) score age ↑, ↓ ADL and physical activity, service use (in-home care, adult day), living alone, social interaction

Tavares et al. (2022) [37] Portugal Cross-sectional study To identify the prevalence and predictors of frailty (Primary Health Care) Community-dwelling 136 older adults aged ≥65 (74.0 ± 6.2 years) Frailty (Fried frailty phenotype) Age ↑, self-rated health, physical capacity

Mielke et al. (2022) [60] Germany Cohort study To analyze gender differences in frailty transition patterns and develop prediction models for improvement/worsening Community-dwelling older 1158 adults aged ≥70 (84.4 ± 5.6 years) Frailty (Fried frailty phenotype) Gender, age, cognitive impairment, education, self-rated health, diabetes, stroke, congestive heart failure (CHF), chronic kidney disease (CKD), osteoarthritis, hospitalization, living alone

Lee et al. (2022) [61] South Korea Cohort study To investigate how multiple frailty domains and their combinations predict institutionalization and mortality (Living Profiles of Older People Survey—LPOPS) Community-based sample 9171 adults aged ≥65 (73.1 ± 6.8 years) Multiple dimensions (physical, cognitive, psychological, and social) profiles; Institutionalization and all-cause mortality age, gender, education, income, marital status, comorbidities, health behaviors; multiple dimensions predict institutionalization/mortality

Thompson et al. (2022) [62] Australia Cohort study To estimate the minimally important difference (MID) in frailty, then describe frailty change over 12 months, and identify factors associated with worsening frailty. Long-term care/Residential Aged Care Services (RACS) 548 residents aged ≥65 (87.7 ± 7.2 years). Frailty change (Frailty Index) age, malnutrition, risk of malnutrition, and diabetes

Yoshizawa (2022) [63] Japan Cohort study To examine whether health literacy (HL) is a predictor of frailty progression among community-dwelling older adults. Community (12 local centers and health institutions). 621 participants (71.6 ± 4.6 years). Frailty progression (Kihon Checklist) Age ↑, ↓ HL, poor mental health, lack of social support, negative views toward mutual aid of social capital

Rodríguez-Laso et al. (2023) [64] Spain Cohort study To explore predictors of sustained transitions between robustness and prefrailty in both directions among older adults Community-dwelling (Toledo Study on Healthy Ageing—TSHA) 949 individuals aged ≥65 (72.3 ± 7.1 years) Fried phenotype age, ↑ medications, not declaring alcohol consumption, and ↓ grip strength

Mangin et al. (2023) [38] Canada Cross-sectional study To explore the association between frailty, chronic conditions, and socioeconomic status (SES) among older adults attending primary care, McMaster University Sentinel and Information Collaboration (MUSIC). Community Primary care 2043 adults aged ≥65 (76.0 ± 7.8 years) Frailty (Clinical Frailty Scale) age, female sex, multimorbidity, ↓ neighborhood income, and ↑ neighborhood material deprivation

Yan et al. (2023) [39] China Cross-sectional study To explore the relationship between frailty syndrome and blood pressure variability in the elderly. Hospital (Neurology department) 150 patients aged ≥65 (69.79 ± 3.23 years) Senile debilitating syndrome/Frailty (Fried scale) Risk factor: Age. Protective factors: Serum albumin (ALB) and high-density lipoprotein (HDL-C)

Huang et al. (2023) [31] Taiwan Quasi-experimental study To investigate the associations between physical activity, frailty, and the effects of multidomain intervention. Community 106 older adults aged 65–96 (77.47 ± 7.19 years) Frailty (Fried’s frailty phenotype) ↓ physical activity (assessed via PASE scores), ↓ mobility, and more depressive symptoms

Zhu et al. (2023) [40] China Cross-sectional population-based study To investigate the association between haemoglobin-to-red blood cell distribution width ratio (HRR) and frailty in older adults. Community-dwelling 1296 older adults aged ≥65 (70.89 ± 4.85 years) Frailty (Fried Frailty Phenotype) The haemoglobin-to-red blood cell distribution width ratio (HRR)is a good predictor.

Costenoble et al. (2023) [65] Belgium Cohort study To examine whether changes in basic, instrumental, and advanced activities of daily living (b-, i-, a-ADLs), social participation, and psychological resilience predict both a transition from robustness to prefrailty or frailty and vice versa Community-dwelling 322 octogenarians (83.04 ± 2.78 years) Worsened and improved frailty transitions (adapted Fried phenotype) Worse: limitations (a-ADLs); improved: Female gender; ↑ social participation

Shakya et al. (2023) [41] USA Cross-sectional To examine the mediating role of cumulative psychosocial stress on frailty disparities among older adults Community-dwelling 7679 older adults aged ≥65 Frailty (Fried frailty phenotype) Female gender, ethnicity, ↓ education, and ↑ cumulative psychosocial stress (e.g., loneliness, discrimination, financial strain)

Tchalla et al. (2023) [66] France Cohort study To identify frailty trajectories and their baseline predictors and clinical outcomes, FREEDOM Community-dwelling 497 individuals aged >75 or >65 with ≥2 comorbidities (82.7 ± 5.5 years) Frailty trajectories (Fried’s criteria) Age, cognitive deficit/dementia, depressive symptoms, and hypertension

Cömertoğlu et al. (2023) [42] Turkey Cross-sectional study To examine the effect of social frailty on mental health and quality of life in older people. Community-dwelling 136 participants over 65 years old (72.2 years median) Social frailty (Tilburg Frailty Indicator); Quality of life Being widowed, having anxiety symptoms, and increased physical limitations

Henry et al. (2023) [43] Australia Cross-sectional study To provide initial insights into whether social cognitive difficulties in older age are associated with social frailty, and how social frailty is linked to psychological characteristics important for health and well-being. Community-dwelling 90 participants aged between 65 and 91 (73.78 ± 5.92 years) Social frailty (Social Frailty Scale) Poorer (inappropriate) social behavior, ↓affective empathy, and depression

Van der Ploeg et al. (2024) [67] Netherlands Cohort study To examine transitions in disability over nine years among Dutch community-dwelling older people aged ≥75 and determine associated predictors Community-dwelling 479 participants at baseline, 69 with complete data across 9 years Disability transitions, Tilburg Frailty Indicator Lifestyle (‘not healthy’), diseases or chronic disorders (‘two or more’), age, income, marital status, education, life events

Suprawesta et al. (2024) [68] Taiwan Longitudinal cohort study To identify baseline predictors associated with cognitive frailty (CF) improvement or progression in community-dwelling older adults Community-dwelling 832 adults aged ≥65 (70.9 ± 5.1 years) (482 completed 2-year follow-up) Cognitive frailty (CF) progression Age, sex, Tinetti balance score, gait velocity, number of comorbidities, depression, regular exercise

Alamri & Mealif (2024) [44] Saudi Arabia Cross-sectional To estimate frailty prevalence and identify its predictors among hospitalized older adults Hospital 147 inpatients aged ≥65 years Frailty status (Clinical Frailty Scale) Higher comorbidity index, polypharmacy, heart failure, and shorter height.

Johnson et al. (2024) [45] United Kingdom Cross-sectional To explore the subdimensions of frailty index using factor analysis and their relationship to quality of life, English Longitudinal Study of Ageing (ELSA) Community-based 4971 participants ages ≥ 65 (74.9 ± 7 years) Quality of life/Frailty Index subdimensions mobility impairment & physical morbidity, difficulty in daily activities, mental health, disorientation

Sepúlveda et al. (2024) [46] Chile Cross-sectional study To identify oxylipins as potential plasma biomarkers for frailty syndrome in ageing, especially in women Community centers and health clinics 71 older adults (F = 50, M = 21), aged ≥65 (F = 73.6 ± 5.9, M = 76.2 ± 6.2 years) Frailty (Fried Phenotype & FTS-5) ↑ plasma oxylipin levels—specifically thromboxane B2 (TxB2) and 7-Hydroxydocosahexaenoic acid (HDoHE) in women—correlated with poor grip strength and slow gait.

Güner et al. (2024) [47] Turkey Cross-sectional To evaluate whether handgrip strength (HGS) asymmetry and weakness are independent predictors of frailty in older adults Community-dwelling 512 older adults aged ≥65 (73.2 ± 6.1 years) Frailty (Clinical Frailty Scale) Handgrip strength (HGS) asymmetry and overall muscle weakness.

Tang et al. (2024) [48] China Cross-sectional To identify frailty risk factors and construct a predictive nomogram for older adults in the Chinese Longitudinal Healthy Longevity Survey (CLHLS) Community-dwelling, 9006 adults aged ≥65 (83.2 years) Frailty risk (Frailty Index) Risk factors: age, ↓ weight or obesity, ↓ self-assessed economic status, and poor sleep quality. Protective factors: education, having a spouse, basic living security, exercising, and social activities.

Savci et al. (2025) [49] Turkey Cross-sectional To examine relationships between dysphagia risk, frailty, and depression in older adults in nursing homes Long-term care (Nursing home) 205 older adults aged ≥65 (70.65 ± 6.36 years) Frailty (Edmonton Frailty Scale) Risk of dysphagia (swallowing difficulties) and advanced age

Tanaka et al. (2025) [69] Japan Longitudinal cohort study To investigate whether anticholinergic drug burden predicts frailty and sarcopenia in community-dwelling older adults, 9-year follow-up Community-dwelling 1549 adults aged ≥65 (72.5 ± 5.5 years) Frailty and sarcopenia Anticholinergic drug burden

Ceylan et al. (2025) [50] Turkey Cross-sectional study To investigate whether handgrip strength (HGS) can alone detect individuals living with frailty (LWF) using the Clinical Frailty Scale (CFS) Community-dwelling 742 patients aged ≥65; 49.3% LWF (median 72 years) Frailty (Clinical Frailty Scale) Older age, female sex, and lower handgrip strength.

Ortiz-Navarro et al. (2025) [51] Spain Cross-sectional study To identify body composition predictors of frailty using DXA, BIA, and US and assess their feasibility in primary care with a mobile app (PowerFrail App) Community-dwelling 94 older adults aged 70–80 classified as frail, prefrail, or robust (76.46 ± 2.92, 75.34 ± 2.86, 73.99 ± 2.59 years) Frailty Body composition variables: whole-body phase angle, right leg lean mass, and ultrasound-measured muscle thickness

Park et al. (2025) [52] South Korea Cross-sectional study To identify predictors of cognitive frailty (CF) and develop a predictive model using multidomain risk factors, Korean Frailty and Aging Cohort Study (KFACS) Community-dwelling 2404 older adults aged 70–84 (CF 77.4 ± 3.8 vs. 75.4 ± 3.8 years) Cognitive frailty (CF) education, physical limitations, malnutritional status, balance

Dong et al. (2025) [70] China Longitudinal cohort study (based on public databases) To develop and validate a risk prediction model for frailty in the elderly using a database from the Chinese Longitudinal Healthy Longevity Survey (CLHLS), 5-year follow-up Community-dwelling 4878 participants aged ≥65 (87 years median) Frailty (Frailty Index) Age, BMI, cognitive function, gender, ethnicity, education.

Tzeng et al. (2025) [53] Taiwan Cross-sectional study To explore how different components of the Falls Risk for Older People in the Community—Taiwan version (Tw-FROP-Com) relate to the five frailty phenotypes (unintentional weight loss, weakness, exhaustion, slow walking speed, and low physical activity) of Fried’s frailty criteria Community-dwelling 375 participants aged ≥65 (75.39 ± 6.94 years) Fried’s frailty: Fall risk Age, education, BMI, comorbidities, ↓ scores on physical activity and gait speed, depression (exhaustion).

Note: ↑ indicates increased risk/value; ↓ indicates decreased risk/value.

Most of the studies selected for this review were set in community environments (n = 34; 85.0%), drawing samples from independent community-dwelling older adults, home-based services, or primary healthcare centers. A smaller proportion of studies (n = 4; 10.0%) were conducted in hospital or specialized outpatient clinic settings. The remaining 2 studies (5.0%) investigated frailty within long-term care settings, including residential aged care facilities and nursing homes.

Study cohorts ranged from small, focused groups (e.g., 71 to 74 participants) [35,46] to massive population-based datasets, such as those reported by Tsay et al. [59] (n = 16,309), Lee et al. [61] (n = 9171), and Tang et al. [48] (n = 9006).

The 40 included primary studies represent a cumulative total of 77,229 participant observations. This figure represents the aggregate sample sizes across published reports rather than unique individuals, as several investigations conducted distinct secondary analyses within shared national cohorts. Specifically, two studies derived from the Chinese Longitudinal Healthy Longevity Survey (CLHLS): Dong et al. [70] conducted a prospective longitudinal cohort study evaluating 4878 older adults newly enrolled in 2008 and followed across four survey waves (median follow-up: 65 months) to model incident frailty; whereas Tang et al. [48] performed a cross-sectional risk-modeling analysis of 9006 participants from the 2017–2018 wave using multivariable logistic regression. Similarly, two publications originated from the Toledo Study on Healthy Aging (TSHA) in Spain: Rodríguez-Laso et al. [58] examining multistate transitions across 1748 participants and Rodríguez-Laso et al. [64] focusing specifically on robustness-to-prefrailty transitions in a subset of 949 participants. In our synthesis, multiple publications originating from a single parent cohort were treated as complementary methodological evidence streams rather than independent population replications.

Regarding recruitment settings, studies were categorized into four distinct environments: (i) population-based community-dwelling independently sampled (n = 34); (ii) acute-care hospital inpatients or emergency department admissions (n = 4); and (iii) institutionalized residents in long-term care or nursing home facilities (n = 2).

As illustrated in Figure 3, the analysis highlights the most frequently reported vulnerability-related factors across the included literature. Importantly, this ranking depicts reporting frequency across studies rather than the statistical magnitude, independence, or certainty of association. Advanced age was the most frequently reported factor (n = 26), followed by reduced functional status (n = 21). Psychological, cognitive, and clinical factors were also prominently represented, including depressive symptoms (n = 15), dementia or memory impairment (n = 15), and co-morbidity (n = 15). Gender-related differences were highlighted in 14 studies (consistently indicating a higher vulnerability prevalence among women). Other notable reported factors included objective mobility limitations, namely longer walking time (n = 10), polypharmacy (n = 9), malnutrition (n = 9), poor social interaction (n = 7), and lower educational attainment (n = 7).

Due to substantial heterogeneity across primary studies—including divergent definitions of vulnerability, differing covariates in multivariable models, and inconsistent reporting of metric effect sizes—pooled quantification of effect estimates (e.g., OR, HR) was not feasible. To address operational heterogeneity and distinguish concurrent correlates from prospective determinants, study-specific exposure metrics, outcome instruments, follow-up durations, multivariable-adjusted effect estimates (OR, HR, β) with 95% confidence intervals, and controlled covariates are systematically detailed across the three evidence streams in Table A5.

The studies analyzed in this review demonstrate substantial methodological diversity in the instruments used to operationalize vulnerability, reflecting the concept’s inherent complexity across varied care settings.

A comprehensive list of all other identified predictors is detailed in Table A4 (Appendix C).

4. Discussion

The findings of this systematic review synthesize the evidence on multidimensional factors associated with vulnerability and frailty in adults aged 65 and older across community, residential, and acute-care settings. To ensure rigorous interpretation in accordance with the methodological framework outlined in Section 2.6, extracted evidence was evaluated across three distinct epidemiological roles: (i) concurrent correlates identified in cross-sectional surveys; (ii) prospective predictors of frailty onset, progression, or state transitions in longitudinal cohorts; and (iii) baseline frailty as an exposure predicting downstream clinical adverse outcomes (institutionalization and mortality). Identified factors were organized across five overarching analytical domains (Table 3).

Table 3Categorization of identified factors associated with vulnerability and frailty.

Domain Frequently Reported Factors (N of Studies)

Biological/Clinical Chronological age (n = 26); Multimorbidity/co-morbidity (n = 15); Polypharmacy (n = 9); Uricemia (n = 2); Diabetes (n = 2); Ulcer risk (n = 2); Anticholinergic burden (n = 1).

Psychological Depressive symptoms (n = 15); Cognitive impairment/dementia (n = 15); Poor self-rated health (n = 5); Sleep disturbances (n = 2).

Socioeconomic/Relational Female gender (n = 14); Lower education (n = 7); Social isolation/low interaction (n = 7); Socioeconomic status/financial strain (n = 4); Living alone (n = 3); Marital status (unmarried/widowed) (n = 3); Race/ethnicity (n = 3); Infrequent visitation (n = 2).

Behavioral/Lifestyle Physical inactivity (n = 4); Smoking history (n = 3); Alcohol intake (n = 2); Low health literacy (n = 1);

Self-care Profile Functional dependence (ADL1 *-/IADL2 **-limitations) (n = 21); Objective mobility/strength decline (n = 10); Malnutrition/nutritional risk (n = 9); Body Mass Index abnormalities (n = 6); Fear of falling/fall history (n = 5); Swallowing difficulties (n = 3); Biomarkers (low albumin, anemia, HRR elevated CRP3 ***, plasma oxylipins); Oral health (n = 3) Pressure injury risk (n = 1); Chronic pain (n = 1).

1 * ADL = Activities of Daily Living, 2 ** IADL = Instrumental Activities of Daily Living/3 *** CRP = C-Reactive Protein. Values indicate the number of primary studies identifying each variable.

4.1. Biological Determinants

Chronological age was the most consistently evaluated demographic variable across the corpus (n = 26; Table A5). Prospective cohort investigations confirmed that advanced age significantly increases the risk of transitioning from robustness to prefrailty or frailty [56,58,66]. Rather than acting in isolation, age operates within a broader matrix of clinical decline, in which multimorbidity (n = 15) serves as a primary longitudinal predictor of functional decline and mortality [57,61] while showing high cross-sectional correlation with frailty severity in community and hospital samples [38,44,53].

Pharmacotherapeutic burden emerged as a prominent clinical determinant, though primary studies evaluated distinct operational exposures and epidemiological designs that must be clearly differentiated. General polypharmacy—predominantly operationalized as the concurrent use of five or more daily medications—demonstrated strong concurrent correlations with frailty severity across cross-sectional acute hospital and outpatient evaluations [34,44,47,50]. In prospective evaluations (Stream 2), baseline polypharmacy independently predicted adverse vulnerability trajectories: in the Longitudinal Aging Study Amsterdam (LASA), de Breij et al. [56] demonstrated that baseline polypharmacy (≥5 prescription drugs) conferred more than twice the odds of 3-year incident frailty (optimism-corrected aOR = 2.14, 95% CI: 1.53–4.19), retained in a final multivariable model adjusted for age, sex, number of chronic diseases, depressive symptoms, cognitive functioning, self-rated health, and pain (univariable OR = 4.76, 95% CI: 3.26–6.94). Similarly, continuous medication count independently hindered positive transitions from prefrailty back to robustness over long-term follow-up in the Toledo Study on Healthy Aging (TSHA) [64] (adjusted OR = 0.92 per additional drug, 95% CI: 0.86–0.99), while Lewis [55] identified high prescription counts (mean 8.24 drugs) and fall-risk-increasing drugs (≥2 FRIDs) as key markers of clinical instability in a community virtual ward cohort. However, interpreting numerical polypharmacy requires substantial caution regarding confounding by indication. Older adults are prescribed multiple medications to manage multimorbidity and advanced disease severity; although several included investigations adjusted for comorbidity indices (e.g., Charlson Comorbidity Index), residual confounding by underlying clinical severity and functional vulnerability cannot be excluded. Crucially, observational associations between high medication counts and frailty status cannot be construed as empirical evidence that deprescribing or reducing medication counts will reverse established frailty.

In contrast to simply counting medications, qualitative pharmacological properties provide a more specific prognostic perspective. In a 9-year follow-up cohort study of 1549 community-dwelling older adults, Tanaka et al. [69] assessed the burden of anticholinergic drugs using the validated Japanese Risk Scale. Older adults with a high baseline anticholinergic burden (score ≥ 3) exhibited a significantly elevated risk of incident physical frailty [adjusted HR = 2.45, 95% CI: 1.52–3.94] and sarcopenia [adjusted HR = 2.01, 95% CI: 1.20–3.35]. While this prospective finding confirms that anticholinergic burden operates as an independent observational marker of clinical decline, interventional randomized trials remain necessary to establish whether targeted pharmacological optimization can prevent or mitigate vulnerability trajectories.

4.2. Psychological Determinants

Depressive symptoms (n = 15) and cognitive decline (n = 15) were prominent psychological factors associated with vulnerability across the included corpus. In longitudinal cohort investigations, baseline depressive symptoms consistently emerged as independent prospective predictors of worsening frailty states, transitions from robustness to prefrailty, and adverse outcomes such as institutionalization and mortality [56,58,61,63,66,68]. Concurrently, cross-sectional surveys and quasi-experimental studies demonstrated high co-occurrence rates between depressive symptoms and physical frailty phenotypes [31,33,38,42,43,45,47,49,52,53]. Several investigations noted that depressive affect was linked to subsequent functional limitations, while worsening physical exhaustion reciprocally contributed to mood disturbances, suggesting a potential bidirectional relationship between affective state and clinical decline [33,38,49]. Cognitive impairment, ranging from subjective memory complaints to diagnosed dementia, was prospectively associated with adverse clinical trajectories in cohort studies, including accelerated functional dependency, residential admission, and mortality [54,55,56,58,60,61,66]. In cross-sectional evaluations, cognitive deficits showed marked concurrent associations with physical frailty markers and deficit accumulation [32,33,34,45,52]. Specifically, two studies explicitly operationalized the construct of ‘cognitive frailty’—defined as the simultaneous manifestation of physical frailty and cognitive impairment without overt dementia. In cross-sectional analysis, cognitive frailty was highly prevalent among older adults with lower educational attainment, malnutrition, and depressive symptoms [52]. Prospectively, longitudinal cohort evidence demonstrated that baseline factors such as advancing age, multi-morbidity, and lack of regular exercise significantly predict the clinical progression of cognitive frailty over time [68].

Poor self-rated health (n = 5) served as an integrative indicator of global vulnerability. In longitudinal cohorts, negative health perceptions independently predicted functional decline and frailty transitions [56,60], while cross-sectional investigations confirmed a linear relationship with physical frailty tiers across community and primary care samples [33,37,48]. Finally, sleep disturbances (n = 2), characterized by poor sleep quality or irregular patterns, were associated with higher frailty odds in community-based cohorts [48,70], reflecting disruptions in neuroendocrine and circadian homeostasis that commonly accompany clinical decline.

4.3. Socioeconomic/Relational Determinants

The synthesized literature identified consistent patterns across sociodemographic factors. Female gender was repeatedly identified as an associated factor with a higher prevalence of clinical frailty across cross-sectional investigations [34,35,38,41,47,50,52]. In contrast, longitudinal cohort studies observed that male older adults, despite demonstrating lower baseline frailty prevalence, experienced faster clinical deterioration and higher mortality rates [60,61] or demonstrated distinct transition trajectories between robust and frail states [56,65,68]. This divergence is consistent with the recognized male-female health-survival paradox in gerontological epidemiology. Lower educational attainment (n = 7) and socioeconomic hardship (n = 4) were recurring social determinants associated with vulnerable states. Dong et al. [70] conducted a prospective cohort study which found that lower levels of education and material deprivation were associated with an increased risk of frailty progression, transitions to functional disability and higher mortality. Similarly, population-based cross-sectional studies documented that limited formal education and neighborhood-level material deprivation were strongly associated with higher odds of both physical and cognitive frailty [32,41,48,50,52,53]. Furthermore, racial and ethnic minority backgrounds were identified in three studies as factors associated with cumulative psychosocial adversity and elevated frailty risk [41,48,70]. Relational determinants, including reduced social interaction (n = 7), living alone (n = 3), unmarried or widowed status (n = 3), and infrequent visitation (n = 2), were consistently associated with vulnerable phenotypes. In longitudinal investigations, diminished social participation and restricted community support networks independently predicted adverse frailty transitions and institutionalization [59,63,65]. Cross-sectional studies link “living without a partner”, “loneliness”, and “relational distance” to physical vulnerability and a reduced quality of life [33,34,42,43,48]. These associations, described as “social frailty,” may have a dimension that goes beyond the biological and may be important for overall vulnerability.

4.4. Behavioral/Lifestyle Determinants

Modifiable lifestyle factors display consistent associations with vulnerability and frailty trajectories across the reviewed literature. Sedentary behavior and insufficient physical activity were identified as significant risk factors for functional decline and frailty progression across prospective cohorts [58,59,67,68,69,70] while being associated with phenotypic frailty in cross-sectional and quasi-experimental evaluations [31,48,53]. Conversely, engagement in regular exercise and health-promoting behaviors was associated with preserved functional status, psychological resilience, and a higher probability of maintaining or regaining robustness in longitudinal and intervention studies [65,67,68]. Cigarette smoking was identified as an independent longitudinal predictor of adverse frailty transitions and accelerated clinical deterioration in prospective cohort designs [58,70]. In parallel, alcohol consumption patterns were evaluated across two investigations [48,64]: prospective cohort evidence from the TSHA study demonstrated that non-drinking was associated with frailty state transitions [64], while cross-sectional multivariable modeling in the CLHLS identified alcohol intake patterns within the broader lifestyle risk profile [48]. Furthermore, lower health literacy was identified by Yoshizawa et al. [63] in a prospective cohort as an independent predictor of frailty progression over 4 years. In that study, limited health literacy compromised older adults’ capacity to comprehend health information and navigate health resources, highlighting a key target for preventive educational and nursing interventions in primary care.

4.5. Self-Care Profile and Functional Determinants

Within the self-care profile, functional limitations in activities of daily living (ADLs and IADLs) represented the most prevalent functional indicators across the included literature. In prospective cohort studies, baseline limitations in instrumental ADLs (e.g., medication management, meal preparation, financial administration) and basic ADLs were identified as independent longitudinal predictors of frailty progression, institutionalization, and all-cause mortality [55,59,61,65,67,69,70]. Cross-sectional surveys and clinical evaluations further confirmed strong concurrent associations between functional dependence and frailty across community and acute-care populations. Notably, Costenoble et al. [65] demonstrated that restrictions in advanced ADLs (a-ADLs) and instrumental tasks precede basic ADL dependence, functioning as early indicators of transitions from robustness to prefrailty and frailty. Objective measures of physical performance and neuromuscular capacity—including decreased gait speed, longer walking time, and reduced or asymmetric handgrip strength—served as sensitive clinical markers of sarcopenia and biological decline. In prospective evaluations, impaired walking time and reduced muscle strength prospectively predicted frailty worsening and adverse health transitions [58,69]. In parallel, cross-sectional and sensor-based biomechanical studies highlighted that handgrip strength weakness and gait asymmetry reliably distinguish frail from non-frail older adults [35,39,46,47,50,51,53]. Nutritional compromise (n = 9), identified via validated screening instruments (e.g., MNA, MUST), was consistently associated with frailty onset and progression [31,32,34,36,47,49,50,52,62]. In residential aged care, Thompson et al. [62] identified malnutrition and risk of malnutrition as prospective predictors of 12-month frailty worsening, while cross-sectional studies documented significant associations between poor nutritional status and physical exhaustion [34,36,49,52]. This vulnerability was compounded by age-related swallowing difficulties and presbyphagia (n = 3), which were correlated with depressive symptoms, inadequate caloric intake, and exacerbated physical frailty [36,49,63]. Nutritional and functional decline was frequently mirrored by objective laboratory biomarkers: low serum albumin levels were prospectively predictive of mortality and frailty progression [57] and cross-sectionally linked to frailty severity [39]. Furthermore, hematological markers—including lower hemoglobin concentrations, elevated C-reactive protein (CRP), and a reduced hemoglobin-to-red blood cell distribution width ratio (HRR)—correlated with physical exhaustion, chronic systemic inflammation, and muscle weakness [32,40]. Finally, psychophysical concerns heavily intersected with functional capacity. A documented history of falls and fear of falling (n = 5) were correlated with voluntary activity restriction and deconditioning, contributing directly to functional dependency [33,52,53,55,63]. Chronic pain (n = 1), identified as an independent predictor of frailty in the LASA cohort [56], restricted mobility and compromised self-care routines. Pressure injury risk was similarly identified as a clinical marker reflecting severe immobility, malnutrition, and advanced clinical dependence [55].

4.6. Conceptual Heterogeneity and Methodological Synthesis

A critical finding emerging from this systematic review is the marked conceptual and operational heterogeneity across primary studies. Although our review conceptualizes vulnerability as an overarching biopsychosocial state, primary investigations predominantly operationalized risk through specific frailty instruments:

The Physical Phenotype: Studies utilizing Fried’s criteria focused exclusively on biological and physiological dimensions (exhaustion, unintentional weight loss, low grip strength, slow walking speed, low physical activity), assessing vulnerability through the lens of physiological exhaustion and sarcopenia [31,32,37,39,41,46,53,58,60,65,66].

Deficit Accumulation (Frailty Index): Investigations adopting the Frailty Index (FI) conceptualized vulnerability as a mathematical ratio of accumulated clinical, functional, and psychological deficits across multiple physiological systems [45,48,59,62,70].

Rapid Clinical and Inpatient Scales: Studies set in acute-care, emergency, or hospital wards prioritized rapid screening instruments, such as the Clinical Frailty Scale (CFS) [38,44,47,50,55] and the Edmonton Frail Scale (EFS) [34,49,57], which blend functional mobility with clinical judgement.

Multidimensional and Domain-Specific Tools: A smaller subset of studies deployed instruments that incorporate social and psychological indicators, such as the Tilburg Frailty Indicator (TFI) [42,67] and the Kihon Checklist (KCL) [63], or isolated sub-constructs such as “cognitive frailty” [52,68] and “social frailty” [42,43].

This operational divergence illustrates that while primary literature frequently employs “frailty” as an operational surrogate for vulnerability, frailty scales primarily measure biological reserve and physical performance. True multidimensional vulnerability necessitates integrating social support, cognitive capacity, environmental conditions, and self-care profiles alongside traditional physical phenotypes.

Finally, these synthesized determinants must be interpreted in light of the underlying methodological designs. As detailed in the results, more than half of the evidence base relies on analytical cross-sectional studies; consequently, these findings reflect concurrent associations and shared vulnerability profiles rather than causal pathways or temporal precedence. True prospective prognostic capability is restricted to the candidate determinants evaluated and validated within the 17 longitudinal cohort investigations.

A critical methodological caveat identified in this synthesis concerns predictor–outcome construct overlap (incorporation bias). Several frequently reported physical performance factors—most notably gait speed, handgrip strength, and physical exhaustion—serve simultaneously as candidate predictor variables and as core diagnostic criteria in phenotypic frailty instruments (such as the Fried phenotype). Consequently, statistical associations between these neuromuscular parameters and frailty status in cross-sectional and unadjusted models partially reflect mathematical and conceptual circularity rather than independent etiological prediction. In contrast, extrinsic multidimensional determinants—including social detachment, educational deprivation, depressive mood, and pharmacotherapeutic burden—operate independently of physical diagnostic definitions. Rather than asserting superior prognostic performance, these multidimensional determinants are less susceptible to predictor–outcome incorporation bias and provide complementary prognostic information alongside established physical phenotypes to support comprehensive clinical risk evaluation.

4.7. Limitations

Although this systematic review provides a comprehensive evidence-based synthesis of multidimensional reported risk factors in older adults, several methodological and contextual limitations must be acknowledged.

First, from a design perspective, more than half of the included evidence base (n = 22) relies on analytical cross-sectional designs, precluding causal inference and the determination of temporal precedence between candidate determinants and vulnerability states. While 17 longitudinal cohort studies were analyzed, their follow-up horizons varied substantially—ranging from 12 months in specific residential cohorts [62,65] to 9 years in population registries [69]—which may differentially capture acute transitional fluctuations versus insidious, long-term biological decline.

Second, the cumulative sample of 77,229 participant observations reflects aggregate study reports rather than entirely unique individuals. Two included investigations performed distinct, complementary analyses on shared national databases (specifically, the CLHLS in China [48,70] and the TSHA in Spain [58,64]. While their analytical designs and outcome frameworks differed, this shared provenance means that population-level evidence is partly concentrated in specific cohorts. Furthermore, many primary studies excluded older adults with severe dementia, total functional dependency, or acute critical illness, potentially underestimating vulnerability prevalence and risk burdens in the most vulnerable strata.

Third, marked conceptual and operational heterogeneity across frailty instruments (ranging from physical phenotypes and deficit accumulation indices to rapid emergency screening tools) precluded quantitative meta-analysis. Consequently, our synthesis matrix (Table 3 and Table A5) reflects the reporting frequency of candidate variables across published studies rather than a standardized hierarchy of statistical effect sizes or independent prognostic weights. Relatedly, while our review conceptualized vulnerability as an overarching biopsychosocial state, primary research predominantly operationalized risk through physical frailty phenotypes. Although physical frailty provides an established, clinically translatable proxy for adverse geriatric outcomes [8,23,25], it captures only a portion of the broader multidimensional vulnerability construct.

Fourth, information and reporting biases must be considered. The reliance on self-reported questionnaires for behavioral, lifestyle, nutritional, and relational variables across multiple investigations introduces potential recall and social desirability biases, whereas objective laboratory biomarkers were evaluated in only a small subset of hospital-based studies. Additionally, restricting our systematic search to peer-reviewed articles published between 2020 and 2025 in four languages (English, French, Portuguese, and Spanish) may have omitted relevant earlier benchmark investigations or literature published in other linguistic regions.

Finally, while we acknowledge the critical issue of predictor–outcome overlap (incorporation bias)—particularly between physical performance predictors and physical frailty diagnostic criteria—we were unable to assess this overlap separately for each specific multidimensional instrument and measurement time point. Because we could not statistically isolate and compare the prognostic robustness of physical versus nonphysical determinants after accounting for overlapping variables, this remains an unresolved limitation of the current synthesized evidence.

4.8. Future Research

Future research should prioritize prospective studies integrating objective digital phenotyping (such as wearable sensors for gait kinematics and mobility) alongside inflammatory and nutritional biomarkers (e.g., CRP, albumin) to minimize self-report bias and clarify pathophysiological trajectories. Furthermore, investigators should utilize the multidimensional candidate variables synthesized in this review to develop and externally validate machine learning algorithms, rigorously assessing model calibration, discrimination, and clinical utility while avoiding predictor–outcome incorporation bias. Finally, randomized interventional trials are needed to establish whether targeted deprescribing and the systematic reduction in anticholinergic burden can actively mitigate vulnerability trajectories across clinical and community care settings.

5. Conclusions

This systematic review synthesized the multifaceted determinants associated with vulnerability and frailty in adults aged 65 and older across community, residential, and acute-care settings. A central insight emerging from this evidence base is that primary research rarely operationalizes vulnerability as a uniform, single outcome. Instead, investigations evaluate risk through distinct frailty paradigms—ranging from physical phenotype models and cumulative deficit indices to multidimensional tools and domain-specific constructs, such as cognitive or social frailty.

Advanced age was the most consistently documented non-modifiable factor. Across biological and clinical dimensions, multimorbidity and numerical polypharmacy correlated with greater frailty severity and prospectively predicted adverse trajectories, including institutionalization and mortality, though these associations remain subject to confounding by indication. Furthermore, longitudinal cohort evidence highlighted qualitative pharmacological risk, demonstrating that high baseline anticholinergic burden independently predicts incident physical frailty over time.

Within the psychological domain, depressive symptoms and cognitive decline were repeatedly linked to vulnerability, independently predicting accelerated functional loss and mortality. When cognitive impairment co-occurs with physical frailty—defined operationally as cognitive frailty—it is prospectively associated with an elevated risk of progressive clinical deterioration. Poorer self-rated health and persistent sleep disturbances further delineate a psychophysical profile that parallels vulnerable states.

Socioeconomic and relational determinants also proved pivotal. Lower educational attainment, financial hardship, and racial or ethnic disparities were associated with heightened vulnerability, reflecting cumulative life-course inequities and limited health literacy. Concurrently, relational detachment—operationalized through living alone, widowhood, and restricted social interaction—consistently correlated with adverse physical outcomes and compromised quality of life, underscoring social frailty as an indispensable component of comprehensive geriatric assessment.

Regarding functional and self-care profiles, reduced gait speed, handgrip weakness, step asymmetry, and nutritional deficits (malnutrition, low serum albumin) were among the most prominent indicators. Rather than operating as isolated prospective causes, objective physical performance parameters largely function as sensitive indicators of depleted physiological reserve and, in physical frailty phenotypes, partially represent operational diagnostic criteria (incorporation bias). Notably, longitudinal findings indicate that functional decline typically manifests in advanced and instrumental activities of daily living before progressing to basic self-care dependence.

Critically, this systematic review does not deliver an empirically validated prediction algorithm or an established clinical risk-stratification instrument. Instead, it provides an evidence-based synthesis and categorization of candidate variables across biological, functional, psychological, behavioral, and relational domains. Future research must prioritize large, prospective longitudinal cohorts and advanced machine learning techniques to externally validate these candidate features, evaluate their statistical independence, and determine their calibration and clinical net benefit. Such validation will be essential for translating these observational determinants into precise, person-centered risk screening in clinical practice.

Acknowledgments

During the preparation of this manuscript, the authors used Gemini (Google) (https://gemini.google.com/) (accessed on 15 September 2025) and NotebookLM (https://notebooklm.google.com/) (accessed on 20 September 2025) for text editing, narrative summarization assistance, and English language refinement. Following the use of these tools, the authors reviewed, manually checked, and edited all content against original source publications and assume full responsibility for the accuracy and integrity of the published work. No artificial intelligence tools were employed for screening decisions, critical quality appraisal, or numerical data extraction.

Abbreviations

The following abbreviations are used in this manuscript:

a-ADL-DI Advanced Activities of Daily Living–Disability Index

ADL Activities of Daily Living

ALB Serum Albumin

BMI Body Mass Index

CCI Charlson comorbidity index

CD-RISC Connor–Davidson Resilience Scale

CF Cognitive Frailty

CFS Clinical Frailty Scale

CHF Congestive Heart Failure

CKD Chronic Kidney Disease

CLHLS Chinese Longitudinal Healthy Longevity Survey

CRP C-Reactive Protein

EFS Edmonton Frail Scale

ELSA English Longitudinal Study of Ageing

FI Frailty Index

FREEDOM French Acronym For Fragilité Et Evaluation à Domicile

HCBS Home and Community-Based Services

HL Health Literacy

HRR Haemoglobin/Rdw Ratio

IADL Instrumental Activities of Daily Living scale

JBI Joanna Briggs Institute

K-ADL Korean Activities of Daily Living scale

KCL Kihon Checklist

KFACS Korean Frailty and Aging Cohort Study

LASA Longitudinal Aging Study Amsterdam

LPOPS Living Profiles of Older People Survey

LSTM Long Short-Term Memory

LWF Living With Frailty

MUSIC McMaster University Sentinel and Information Collaboration

MUST Malnutrition Universal Screening Tool

PASE Physical Activity Scale for the Elderly

PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses

FP Fried phenotype

SARC-F Strength, Assistance with walking, Rising from chair, Climbing stairs, Falls

SARS-CoV-2 Severe Acute Respiratory Syndrome Coronavirus 2

SES Socioeconomic Status

TFI Tilburg frailty indicator

TSHA Toledo Study on Healthy Ageing

TUG/TUGT Timed Up and Go

Tw-FROP- Falls Risk for Older People in the Community—Taiwan version

TxB2/7-HDoHE Thromboxane B2/7-Hydroxydocosahexaenoic acid

WHO-5 WHO-5 Well-being Index

Appendix A

Table A1Full search syntax and database query parameters.

Database/Interface Full Search Query/Advanced Syntax Full Search Query/Advanced Syntax Limits & Filters Applied Initial Yield (n)

PubMed/MEDLINE (“Vulnerable Populations”[Mesh] OR “Frail Elderly” [Mesh] OR “functionally impaired elderly”) AND (“Geriatric Assessment”[Mesh] OR “International Classification of Functioning, Disability and Health” [Mesh] OR “Health Status Indicators” [Mesh] OR “Quality of Life” [Mesh]) AND (prognosis OR predictor) NOT (review) 2 June 2025 Publication date: Last 5 years (2020–2025); Languages: English, French, Portuguese, Spanish. 466

Scopus (Elsevier) TITLE-ABS-KEY ((“Vulnerable Populations” OR “Frail Elderly” OR “functionally impaired elderly”) AND (“Geriatric Assessment” OR “International Classification of Functioning, Disability and Health” OR “Health Status Indicators” OR “Quality of Life”) AND (prognosis OR predictor)) AND PUBYEAR > 2019 AND PUBYEAR < 2026 AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (LANGUAGE, “Portuguese”) OR LIMIT-TO (LANGUAGE, “Spanish”) OR LIMIT-TO (LANGUAGE, “English”) OR LIMIT-TO (LANGUAGE, “French”)) 3 June 2025 Publication years: 2020–2025; Document type: Article (“ar”); Languages: English, French, Portuguese, Spanish. 599

Web of Science (Clarivate) (“Vulnerable Populations” OR “Frail Elderly” OR “functionally impaired elderly”) AND (“Geriatric Assessment” OR “International Classification of Functioning, Disability and Health” OR “Health Status Indicators” OR “Quality of Life”) AND (prognosis OR predictor)Query session: (https://www.webofscience.com/wos/woscc/summary/7df9cdc3-0dc7-4a76-86f4-1ff4400ac902-0165608ee8/relevance/1) (accessed on 4 June 2025) 4 June 2025 Timespan: 1 January 2020 to 29 May 2025; Document types: Articles; Languages: English, French, Portuguese, Spanish. 39

CINAHL (EBSCOhost) (“Vulnerable Populations” OR “Frail Elderly” OR “functionally impaired elderly”) AND (“Geriatric Assessment” OR “International Classification of Functioning, Disability and Health” OR “Health Status Indicators” OR “Quality of Life”) AND (prognosis OR predictor) 4 June 2025 Published Date: 1 January 2020–29 May 2025; Languages: English, French, Portuguese, Spanish. 77

Subtotal (Databases) Consolidated database search yield

1181

Google Scholar allintitle: ((Vulnerable OR FRAIL OR risk) AND aged 65) 5 June 2025 Screened for non-indexed grey literature; first 100 consecutive records sorted by default relevance. 100

Total Identified Cumulative records entered into PRISMA 2020 Flow Diagram

1281

Note: The complete unedited strategy document and search logs are permanently accessible in the prospective PROSPERO record (CRD420251039796) at: (https://www.crd.york.ac.uk/PROSPEROFILES/d54273f12a9efe9533ad20638d86ba83.pdf) (accessed on 25 August 2026).

Appendix B

Table A2Results of the methodological quality assessment for the JBI critical appraisal tool of analytical quasi-experimental studies.

JBI Critical Appraisal Tool of Analytical Quasi-Experimental Studies (Non-Randomized Experimental Studies)

Study Rev 1 2 3 4 5 6 7 8 9 Score Level

Huang et al. (2023) [31] AR Y Y Y N N N Y Y Y 6/9(67%) Moderate

HP Y Y Y N N N Y Y Y 6/9(67%) Moderate

Y—Yes; N—No.

Table A3Results of the methodological quality assessment for the JBI critical appraisal tool of analytical cross-sectional studies.

JBI Critical Appraisal Tool of Analytical Cross-Sectional Studies

Study Rev 1 2 3 4 5 6 7 8 Score Level

Steinmeyer et al. (2020) [32] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Venturini et al. (2021) [33] AR Y Y Y Y U N Y Y 6/8 (75%) Moderate

HP Y Y Y Y U N Y Y 6/8 (75%) Moderate

Sharma et al. (2021) [34] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Jung et al. (2021) [35] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Sella-Weiss et al. (2021) [36] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Tavares et al. (2022) [37] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Mangin et al. (2023) [38] AR Y Y Y Y N N Y Y 6/8 (75%) Moderate

HP Y Y Y Y N N Y Y 6/8 (75%) Moderate

Yan et al. (2023) [39] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Zhu et al. (2023) [40] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Shakya et al. (2023) [41] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Cömertoğlu et al. (2023) [42] AR U Y Y Y Y Y Y Y 7/8 (88%) High

HP U Y Y Y Y Y Y Y 7/8 (88%) High

Henry et al. (2023) [43] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Alamri & Mealif (2024) [44] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Johnson et al. (2024) [45] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Sepúlveda et al. (2024) [46] AP Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Güner et al. (2024) [47] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Tang et al. (2024) [48] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Savci et al. (2025) [49] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Ceylan et al. (2025) [50] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Ortiz-Navarro et al. (2025) [51] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Park et al. (2025) [52] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Tzeng et al. (2025) [53] AR Y Y Y Y Y Y Y Y 8/8 (100%) High

HP Y Y Y Y Y Y Y Y 8/8 (100%) High

Y—Yes; N—No; U—Unclear.

Table A4Results of the methodological quality assessment for the JBI critical appraisal tool of analytical cohort studies.

JBI Critical Appraisal Tool of Analytical Cohort Studies

Study Rev 1 2 3 4 5 6 7 8 9 10 11 Score Level

Hajek et al. (2020) [54] AR Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

HP Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

Lewis (2020) [55] AR Y Y Y N * N * N Y Y Y N Y 7/11 (64%) Moderate

HP Y Y Y N * N * N Y Y Y N Y 7/11 (64%) Moderate

Breij et al. (2020) [56] AR Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

HP Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

Varan et al. (2021) [57] AR Y Y Y Y Y Y Y Y Y Y U 10/11 (91%) High

HP Y Y Y Y Y Y Y Y Y Y U 10/11 (91%) High

Rodríguez-Laso et al. (2021) [58] AR Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

HP Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

Tsay et al. (2021) [59] AR Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

Mielke et al. (2022) [60] AR Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

Lee et al. (2022) [61] AR Y Y Y Y Y Y Y Y Y Y Y 11/11 (100%) High

HP Y Y Y Y Y Y Y Y Y Y Y 11/11 (100%) High

Thompson et al. (2022) [62] AR Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

Yoshizawa (2022) [63] AR Y Y Y Y Y Y Y Y N N Y 9/11 (82%) High

HP Y Y Y Y Y Y Y Y N N Y 9/11 (82%) High

Rodríguez-Laso et al. (2023) [64] AR Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

Costenoble et al. (2023) [65] AR Y Y Y Y Y Y Y N Y N Y 9/11 (82%) High

HP Y Y Y Y Y Y Y N Y N Y 9/11 (82%) High

Tchalla et al. (2023) [66] AR Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y Y N Y 9/11 (82%) High

Van der Ploeg et al. (2024) [67] AR Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

HP Y Y Y Y Y N Y Y N N Y 8/11 (73%) Moderate

Suprawesta et al. (2024) [68] AR Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

HP Y Y Y Y Y N Y Y N Y Y 9/11 (82%) High

Tanaka et al. (2025) [69] AR Y Y Y Y Y Y Y Y N Y Y 10/11 (91%) High

HP Y Y Y Y Y Y Y Y N Y Y 10/11 (91%) High

Dong et al. (2025) [70] AR Y Y Y Y Y Y Y Y N Y Y 10/11 (91%) High

HP Y Y Y Y Y Y Y Y N Y Y 10/11 (91%) High

Y—Yes; N—No; U—Unclear. * Note on Consensus Adjudication: For Lewis (2020) [55], initial reviewer discordance on Items 4 and 5 (measurement of exposure and identification of confounding) was arbitrated with a third reviewer (A.C.); because exposure protocols in the virtual ward setting relied on administrative health provider counts and confounding adjustments were restricted to baseline univariable comparisons for certain sub-outcomes, Items 4 and 5 were adjudicated as ‘No’, confirming a final consensus score of 7/11 (64%, Moderate quality).

Appendix C

Table A5Master Study Identification and Cohort Mapping Table (N = 40).

ID Author & Year & Reference Full Article Title & Journal Citation Persistent Identifier (DOI/PMID) Underlying Cohort/Data Source Study Design Country & Setting Sample Size (n)

1 Hajek et al. (2020) [54] Which Factors Contribute to Frailty among the Oldest Old? Results of the AgeCoDe and AgeQualiDe Study. Gerontology, 66(5), 460–466. DOI: 10.1159/000508723 AgeCoDe/AgeQualiDe Cohort Study Germany (Community-dwelling) 486

2 Lewis (2020) [55] Risk Prediction for Adverse Outcomes for Frail Older Persons with Complex Healthcare and Social Care Needs Admitted to a Community Virtual Ward Model DOI: 10.2147/CIA.S236895 Community Virtual Ward (CVW) Cohort Cohort Study Ireland (Community-dwelling) 88

3 Steinmeyer et al. (2020) [32] Hemoglobin concentration; a pathway to frailty. BMC Geriatrics, 20, 203. DOI: 10.1186/s12877-020-01597-6 Toulouse University Hospital Geriatric Day Hospital Cohort Cross-sectional Study France (Hospital) 1829

4 De Breij et al. (2020) [56] Predictors of Frailty and Vitality in Older Adults Aged 75 years and Over: Results from the Longitudinal Aging Study Amsterdam. Gerontology, 67(1), 69–77. DOI: 10.1159/000512049 Longitudinal Aging Study Amsterdam (LASA) Cohort Study Netherlands (Community-dwelling) 569

5 Venturini et al. (2021) [33] A multidimensional approach to frailty in Brazilian older adults: the FIBRA-BR study. BMC Geriatrics, 21, 143. DOI: 10.1186/s12877-021-02193-y FIBRA-BR Study (Frailty in Brazilian Older People) Cross-sectional Study Brazil (Community-dwelling) 3569

6 Varan et al. (2021) [57] Comparative Evaluation of Predictive Ability of Comprehensive Geriatric Assessment Components Including Frailty on Long-Term Mortality DOI: 10.1080/0361073X.2021.1876395 Gazi University Geriatric Outpatient Clinic Cohort Cohort Study Turkey (Community-dwelling) 967

7 Sharma et al. (2021) [34] Nutritional risk assessed by the Malnutrition Universal Screening Tool as a predictor of frailty in acutely hospitalised older patients: An observational study. Asia Pacific Journal of Clinical Nutrition, 30(2), 187–195. DOI: 10.6133/apjcn.202106_30(2).0001 Flinders Medical Centre Acute Inpatient Cohort Cross-sectional Study Australia (Hospital) 243

8 Jung et al. (2021) [35] Frailty Assessment Using Temporal Gait Characteristics and a Long Short-Term Memory Network. IEEE Journal of Biomedical and Health Informatics DOI: 10.1109/JBHI.2021.3067931 Kyung Hee University Medical Center Cohort Cross-sectional Study South Korea (Community-dwelling) 74

9 Sella-Weiss et al. (2021) [36] Association between swallowing function, malnutrition and frailty in community dwelling older people. Clinical Nutrition ESPEN, 45, 476–485. DOI: 10.1016/j.clnesp.2021.06.028 Haifa & Ono Academic College Community Cohort Cross-sectional Study Israel (Community-dwelling) 180

10 Rodríguez-Laso et al. (2021) [58] Transitions Between Frailty States and Its Predictors in a Cohort of Community-Dwelling Spaniards. JAMDA, 23(4), 524.e1–524.e11. DOI: 10.1016/j.jamda.2021.07.008 Toledo Study for Healthy Aging (TSHA) Cohort Study Spain (Community-dwelling) 1748

11 Tsay et al. (2021) [59] Predictors of Frailty Change in Home and Community-Based Services Populations DOI: 10.1016/j.jamda.2021.07.032 Connecticut Medicaid HCBS Database (interRAI-HC) Cohort Study USA (Community-dwelling) 16,309

12 Tavares et al. (2022) [37] Predictors of frailty in older people users of Primary Health Care. Revista Brasileira de Enfermagem DOI: 10.1590/0034-7167-2020-1292 Vila Nova de Gaia Family Health Unit (FHU) Cohort Cross-sectional Study Portugal (Community-dwelling) 136

13 Mielke et al. (2022) [60] Gender differences in frailty transition and its prediction in community-dwelling old adults. Scientific Reports DOI: 10.1038/s41598-022-11358-7 Berlin Initiative Study (BIS) Cohort Study Germany (Community-dwelling) 1158

14 Lee et al. (2022) [61] The influence of multiple frailty profiles on institutionalization and all-cause mortality in community-living older adults. Journal of Cachexia, Sarcopenia and Muscle, 13(5), 2322–2330. DOI: 10.1002/jcsm.13033 Living Profiles of Older People Survey (LPOPS) Cohort Study South Korea (Community-dwelling) 9171

15 Thompson et al. (2022) [62] Frailty change based on minimally important difference in nursing home residents: FIRST cohort study findings. Age and Ageing, 51(5), afac246. DOI: 10.1093/ageing/afac246 FIRST Cohort Study (Frailty in Residential Sector Teams) Cohort Study Australia (Long-term care/nursing home) 548

16 Yoshizawa (2022) [63] Impact of Health Literacy on the Progression of Frailty after 4 Years among Community-Dwelling Older Adults. IJERPH, 19(1), 394. DOI: 10.3390/ijerph19010394 Kashiwa Longitudinal Study Cohort Study Japan (Community-dwelling) 621

17 Rodríguez-Laso et al. (2023) [64] Predictors of Maintained Transitions Between Robustness and Prefrailty in Community-Dwelling Older Spaniards. JAMDA, 24(1), 57–64. DOI: 10.1016/j.jamda.2022.10.010 Toledo Study for Healthy Aging (TSHA) Cohort Study Spain (Community-dwelling) 949

18 Mangin et al. (2023) [38] Association between frailty, chronic conditions and socioeconomic status in community-dwelling older adults attending primary care: a cross-sectional study using practice-based research network data. BMJ Open DOI: 10.1136/bmjopen-2022-066269 MUSIC PBRN (McMaster University Practice-Based Network) Cross-sectional Study Canada (Community-dwelling) 2043

19 Yan et al. (2023) [39] Analysis of systolic and diastolic blood pressure and frailty in community-dwelling older adults. Medicine DOI: 10.1097/MD.0000000000032874 Nantong University Affiliated Hospital Geriatric Cohort Cross-sectional Study China (Hospital) 150

20 Huang et al. (2023) [31] Physical activity plays a crucial role in multidomain intervention for frailty prevention. Aging Clinical and Experimental Research, 35(6), 1283–1292. DOI: 10.1007/s40520-023-02412-z Kaohsiung Community Care Centers Intervention Cohort Quasi-experimental Study Taiwan (Kaohsiung, Community Centers) 106

21 Zhu et al. (2023) [40] Lower haemoglobin-to-red blood cell distribution width ratio is independently associated with frailty in community-dwelling older adults: a cross-sectional study. BMJ Open. DOI: 10.1136/bmjopen-2022-069141 Wuhan Community Older Adult Health Study Cross-sectional Study China (Community-dwelling) 1296

22 Costenoble et al. (2023) [65] Transitions in robust and prefrail octogenarians after 1 year: the influence of activities of daily living, social participation, and psychological resilience on the frailty state. BMC Geriatrics, 23, 485. DOI: 10.1186/s12877-023-04178-5 BUTTERFLY Study (Vrije Universiteit Brussel) Cohort Study Belgium (Community-dwelling) 322

23 Shakya et al. (2023) [41] Does cumulative psychosocial stress explain frailty disparities in community-dwelling older adults? Archives of Gerontology and Geriatrics, 113, 105055. DOI: 10.1016/j.archger.2023.105055 Health and Retirement Study (HRS) Cross-sectional Study USA (Community-dwelling) 7679

24 Tchalla et al. (2023) [66] Patterns, predictors, and outcomes of frailty trajectories in community-dwelling older adults: Results from The FREEDOM Cohort Study. Archives of Gerontology and Geriatrics, 113, 105080. DOI: 10.1016/j.archger.2023.105101 FREEDOM Cohort Study Cohort Study France (Community-dwelling) 497

25 Cömertoğlu et al. (2023) [42] The Effect of Social Frailty on Mental Health and Quality of Life in Older People: A Cross-Sectional Study. Research Square. DOI: 10.21203/rs.3.rs-2852088/v1 Hacettepe University Geriatric Outpatient Cohort Cross-sectional Study Turkey (Community-dwelling) 136

26 Henry et al. (2023) [43] Social Frailty in Late Adulthood: Social Cognitive and Psychological Well-Being Correlates. The gerontological society of america. DOI: 10.1093/geronb/gbac157 Brisbane Community Older Adult Survey Cross-sectional Study Australia (Community-dwelling) 90

27 Van der Ploeg et al. (2024) [67] Disability transitions in Dutch community-dwelling older people aged 75 years or older. Archives of Gerontology and Geriatrics, 116, 105165. DOI: 10.1016/j.archger.2023.105165 Roosendaal Longitudinal Study Cohort Study Netherlands (Community-dwelling) 479

28 Suprawesta et al. (2024) [68] Factors affecting cognitive frailty improvement and progression in Taiwanese older adults. BMC Geriatrics, 24, 105. DOI: 10.1186/s12877-024-04700-3 Taipei Medical University Outpatient & Community Cohort Cohort Study Taiwan (Community-dwelling) 832

29 Alamri & Mealif (2024) [44] Exploring the correlates of frailty among hospitalized older adults: A cross-sectional study in a Saudi teaching hospital. Medicine, 103(25), e38603. DOI: 10.1097/MD.0000000000038603 King Abdulaziz University Hospital (KAUH) Inpatient Database Cross-sectional Study Saudi Arabia (Hospital) 147

30 Johnson et al. (2024) [45] Frailty or frailties: exploring frailty index subdimensions in the English Longitudinal Study of Ageing. JECH, 78(10), 609–115. DOI: 10.1136/jech-2023-221829 English Longitudinal Study of Ageing (ELSA, Wave 9) Cross-sectional Study UK (Community-dwelling) 4971

31 Sepúlveda et al. (2024) [46] Physiological changes associated with aging: Identification of novel biomarkers for frailty syndrome in women. Free Radical Biology and Medicine, 223, 160–171. DOI: 10.1016/j.freeradbiomed.2024.07.022 PIEI-ES Healthy Aging Program (Talca/San Clemente) Cross-sectional Study Chile (Community-dwelling) 71

32 Güner et al. (2024) [47] A hand-in-hand phenomenon in older adults: Increased risk of frailty in geriatric outpatients associated with handgrip strength asymmetry and weakness. Clinical Nutrition, 43(10), 2381–2387. DOI: 10.1016/j.clnu.2024.09.010 Hacettepe University Geriatric Medicine Outpatient Cohort Cross-sectional Study Turkey (Community-dwelling) 512

33 Tang et al. (2024) [48] Development of a Frailty Prediction Model Among Older Adults in China: A Cross-Sectional Analysis Using the Chinese Longitudinal Healthy Longevity Survey. Nursing Open, 11(4), e70070. DOI: 10.1002/nop2.70070 Chinese Longitudinal Healthy Longevity Survey (CLHLS 2018 Wave) Cross-sectional Study China (Community-dwelling) 9006

34 Savci et al. (2025) [49] Relationship between frailty, nutritional status, and physical performance in community-dwelling older adults. Archives of Gerontology and Geriatrics DOI: 10.1016/j.gerinurse.2025.02.001 Community Health Care Centers Cohort Cross-sectional Study Turkey (Long-term care/nursing home) 205

35 Tanaka et al. (2025) [69] Anticholinergic burden quantified using the Japanese risk scale as a predictor of frailty and sarcopenia among community-dwelling older adults: A 9-year Kashiwa cohort study. Geriatrics & Gerontology International, 25, 520–527. DOI: 10.1111/ggi.70012 Kashiwa Longitudinal Population-Based Cohort Study Cohort Study Japan (Community-dwelling) 1549

36 Ceylan et al. (2025) [50] Can handgrip strength alone detect individuals living with frailty according to the Clinical Frailty Scale? DOI: 10.1590/1806-9282.20240237 Hacettepe University Geriatric Outpatient Cohort Cross-sectional Study Turkey (Community-dwelling) 742

37 Ortiz-Navarro et al. (2025) [51] Identification of Target Body Composition Parameters by DXA, Bioelectrical Impedance, and Ultrasonography to Detect Older Adults With Frailty Using a Mobile App in Primary Care. JMIR Aging, 8, e67982. DOI: 10.2196/67982 Andalusian Public Health System Primary Care Cohort Cross-sectional Study Spain (Community-dwelling) 94

38 Park et al. (2025) [52] Predicting cognitive frailty in community-dwelling older adults: a machine learning approach based on multidomain risk factors. Scientific Reports. DOI:10.1038/s41598-025-00844-3 Korean Frailty and Aging Cohort Study (KFACS) Cross-sectional Study South Korea (Community-dwelling) 2404

39 Dong et al. (2025) [70] Development and validation of a prediction model of frailty risk in community-dwelling older adults: From a national longitudinal survey. Public Health, 240, 63–70. DOI: 10.1016/j.puhe.2024.12.055 Chinese Longitudinal Healthy Longevity Survey (CLHLS) Cohort Study China (Community-dwelling) 4878

40 Tzeng et al. (2025) [53] Examining frailty phenotypes of community-dwelling older adults in Taiwan using the falls risk for older people in the community—Taiwan version (Tw-FROP-Com). Aging, 17, 206231. DOI: 10.18632/aging.206231 Taiwan Community Frailty Assessment Survey Cross-sectional Study Taiwan (Community-dwelling) 375

— Cumulative Total

77,229

Table A6Effect Estimates and Adjustment Models across Epidemiological Evidence Streams.

Study (Author, Year) Setting/Cohort Follow-Up Exposure/Predictor Outcome/Endpoint Adjusted Effect Estimate 95% CI Adjusted Covariates

Stream 1: Concurrent Cross-Sectional Associations

Steinmeyer et al. (2020) [32] France (Outpatient Day Hospital) N/A Low Hemoglobin (<12 g/dL women, <13 men) Gerontopôle Frailty Screening Tool (GFST) OR = 0.86 per g/dL increase 0.79–0.94 Age, sex, eGFR, CRP, MMSE, MNA, living situation

Venturini et al. (2021) [33] Brazil (Community, 15 cities) N/A Poor SRH, incontinence, depressive symptoms, no visits Fried Frailty Phenotype (FIBRA-BR) Factor 1 Phys: p < 0.001; Factor 3 Psych: p < 0.001 N/A (Factor scores) Age, sex, sociodemographic factors

Sharma et al. (2021) [34] Australia (Acute Hospital Wards) N/A MUST score ≥ 1 (Medium/High Risk) Edmonton Frail Scale (EFS) OR = 3.82 (High vs. Low risk) 1.74–8.38 Age, sex, comorbidity, polypharmacy, mobility

Jung et al. (2021) [35] South Korea (Community) N/A Temporal gait variability (LSTM wearable sensors) FRAIL Scale Accuracy = 88.5%; Gait asymmetry p < 0.01 N/A (ML classifier) Age, sex, body height

Sella-Weiss et al. (2021) [36] Israel (Community) N/A H-EAT-10 score (Dysphagia risk) FRAIL Scale & Grip Strength β = 0.057 (Linear regression) 0.004–0.111 Age, MNA-SF score, Geriatric Depression Scale (GDS-SF)

Tavares et al. (2022) [37] Portugal (Primary Care) N/A Age (years) & Physical capacity score Fried Frailty Phenotype OR = 1.11 per year; OR = 0.67 per capacity unit 1.03–1.20; 0.51–0.89 Age, sex, self-rated health, living conditions

Mangin et al. (2023) [38] Canada (Primary Care PBRN) N/A ≥5 chronic conditions & Neighborhood income decile Clinical Frailty Scale (CFS) χ2 = 137.92 (Multimorbidity); χ2 = 61.42 (SES) N/A (Bivariate χ2) Retained in unadjusted/bivariate stratification

Yan et al. (2023) [39] China (Hospital Inpatient) N/A 24h SBP/DBP variability & Low Albumin (<35 g/L) Fried Frailty Phenotype Albumin: p = 0.002; SBP variability: p = 0.014 N/A (ANOVA/Spearman) Age, sex, BMI, fasting blood glucose

Zhu et al. (2023) [40] China (Community) N/A Lower HRR (<9.97 cutoff) Fried Frailty Phenotype AUC = 0.802; OR = 3.42 (Low vs. Normal HRR) 2.11–5.54 Age, sex, BMI, SBP, DBP, fasting glucose, lipids

Shakya et al. (2023) [41] USA (Nationwide HRS) N/A Cumulative stress index (0–6) & Minority status Fried Frailty Phenotype aOR = 1.50 per stress unit 1.32–1.71 Age, sex, income, education, chronic conditions

Cömertoğlu et al. (2023) [42] Turkey (Geriatric Outpatient) N/A HALFT Scale (Social frailty score ≥ 2) Clinical Frailty Scale & FRAIL r = 0.428 (CFS correlation) N/A (Correlation) GDS, MMSE, ADL, IADL

Henry et al. (2023) [43] Australia (Community) N/A Inappropriate social behavior & Demoralization Physical & Cognitive Frailty β = 0.24 (Inappropriate behavior) 0.04–0.44 Age, sex, depression, physical frailty, cognitive frailty

Alamri & Mealif (2024) [44] Saudi Arabia (Hospital Inpatient) N/A Comorbidity index & Polypharmacy (≥5 drugs) Clinical Frailty Scale (CFS) OR = 1.49 per comorbidity unit; Polypharmacy p = 0.003 1.14–1.99 Height, comorbidity index, heart failure

Johnson et al. (2024) [45] UK (Nationwide ELSA Wave 9) N/A Mobility, Daily Activities, Mental Health, Orientation Frailty Index (58 items) Factor 1 (Mobility): r = 0.929; Factor 3 (Mental): r = 0.512 N/A (Factor analysis) Age, sex, CASP-19 quality of life score

Sepúlveda et al. (2024) [46] Chile (Community) N/A TxB2 & 7-HDoHE plasma levels Edmonton Frail Scale (EFS) p = 0.012 (Oxylipin divergence) N/A (Biomarker comparison) Age, sex, cardiovascular risk factors

Güner et al. (2024) [47] Turkey (Geriatric Outpatient) N/A Asymmetric HGS (>10% diff) & Weakness Clinical Frailty Scale (CFS) aOR = 2.85 (Asymmetry + Weakness) 1.62–5.01 Age, sex, BMI, MNA-SF, MMSE, GDS, polypharmacy

Tang et al. (2024) [48] China (Nationwide CLHLS 2018) N/A Rural residence, education, sleep, exercise Frailty Index (38 items, cutoff ≥ 0.25) AUC = 0.831; Age 80–94 OR = 4.58; Exercise OR = 0.76 3.88–5.42; 0.63–0.92 Age, sex, residence, living status, education, BMI, sleep, smoking, alcohol

Savci et al. (2025) [49] Turkey (Nursing Home Residents) N/A EAT-10 score ≥ 3 & GDS score ≥ 5 Edmonton Frail Scale (EFS) r = 0.384 (Dysphagia vs. Frailty); r = 0.412 (GDS) N/A (Correlation) Age, sex, comorbidity count, medication count

Ceylan et al. (2025) [50] Turkey (Geriatric Outpatient) N/A HGS < 20.5 kg (women), <31.5 kg (men) Clinical Frailty Scale (CFS ≥ 4) aOR = 2.11 (Female sex); aOR = 0.88 per kg HGS 1.34–3.32; 0.84–0.92 Age, sex, HGS, education, multimorbidity, polypharmacy

Ortiz-Navarro et al. (2025) [51] Spain (Primary Care) N/A Relative muscle power (RPOW) & Rectus femoris US Fried Phenotype & SPPB AUC = 0.892 (RPOW); US thickness p = 0.004 0.812–0.972 Age, sex, BMI, MNA, GPCOG, 5STS, 4 m walk speed

Park et al. (2025) [52] South Korea (Nationwide KFACS) N/A Sensor-based STS velocity & Cognitive impairment Cognitive Frailty (PF + IC) aOR = 4.71 (Slowness); aOR = 5.09 (Weakness) 3.74–5.93; 3.98–6.51 Age, sex, education, comorbidities, prescription count, GDS

Tzeng et al. (2025) [53] Taiwan (Community) N/A Fall history & 5 Fried Phenotype items Fried Phenotype & Fall Risk Score aOR = 15.51 (Frail vs. Non-frail on Fall Risk) 4.53–53.15 Age, education, BMI, comorbidities, physical activity, gait speed

Stream 2: Baseline Determinants Predicting Frailty Onset

Hajek et al. (2020) [54] Germany (AgeCoDe/AgeQualiDe) 20 months Age ≥ 85 years; Oldest-old primary care Clinical Frailty Scale (CFS) Fixed Effects β = 0.14 per age year 0.05–0.23 Panel fixed-effects controlling all time-invariant unobserved heterogeneity

Lewis (2020) [55] Ireland (Community Virtual Ward) 90 days >7 HCPs in Red CVW; <2 care episodes Rockwood Clinical Frailty Scale (CFS) OR = 14.77 (>7 HCPs); OR = 44.89 (<2 care episodes) 2.69–81.2; 12.4–162.1 Age, sex, comorbidity index, baseline CFS, home care hours

Breij et al. (2020) [56] Netherlands (LASA Cohort) 3 years Polypharmacy (≥5 drugs), MMSE, GDS LASA Frailty Index (FI ≥ 0.25) aOR = 2.14 (Polypharmacy); OR = 1.33 (Pain) 1.53–4.19; 1.18–1.71 Age, sex, depressive symptoms, chronic diseases, self-rated health

Varan et al. (2021) [57] Turkey (Gazi Geriatric Outpatient) 47 months (median) Albumin < 3.5 g/dL; 15ft walking time > 8.9 s Edmonton Frail Scale (EFS) & Fried Phenotype HR = 0.68 per g/dL Albumin; HR = 1.12 per sec walking 0.52–0.89; 1.05–1.21 Age, sex, comorbidity count, ADL, IADL, MMSE

Rodríguez-Laso et al. (2021) [58] Spain (Toledo Study TSHA Wave 1–2) 5.2 years (median) Fried phenotype items & lifestyle factors Fried Frailty Phenotype (Transitions) OR = 1.10 per age year; OR = 1.12 per GDS point 1.06–1.15; 1.01–1.24 Age, sex, education, BMI, alcohol, PASE, IADL, gait speed, grip strength

Tsay et al. (2021) [59] USA (Connecticut Medicaid HCBS) 8–16 months In-home care vs. Adult day services received interRAI-HC Frailty Index OR = 1.48 (ADL decline); OR = 0.68 (In-home care improvement) 1.03–2.12; 0.47–0.97 Age, sex, race/ethnicity, living alone, baseline frailty tier

Mielke et al. (2022) [60] Germany (Berlin Initiative Study BIS) 2.1 years Cognitive impairment, CHF, Diabetes, ACR Fried Frailty Phenotype Women Worsening OR = 3.70 (Cognitive); Men Worsening OR = 2.43 1.73–7.92; 1.20–4.93 Age, self-rated health, CHF, ACR, GFR, stroke, diabetes, living alone, smoking

Lee et al. (2022) [61] South Korea (LPOPS Nationwide) 3 years Single vs. 2-, 3-, or 4-domain frailty profiles Institutionalization & All-Cause Mortality Mixed Frailty (4 domains) HR = 3.89 (Mortality) 2.57–5.85 Age, sex, smoking, alcohol drinking, comorbidities

Thompson et al. (2022) [62] Australia (FIRST Cohort Residential) 12 months MNA At Risk/Malnourished vs. Normal 47-item Frailty Index (MID change) OR = 1.88 (At Risk); OR = 2.02 (Malnourished) 1.28–2.75; 1.17–3.48 Age, sex, length of stay, MI, stroke, COPD, dementia, diabetes, polypharmacy

Yoshizawa (2022) [63] Japan (Kashiwa Study) 4 years Limited vs. High Health Literacy Kihon Checklist (KCL) aOR = 1.85 (Limited Health Literacy) 1.18–2.91 Age, sex, education, living alone, baseline KCL score, comorbidities

Rodríguez-Laso et al. (2023) [64] Spain (Toledo Study TSHA Wave 1–3) 8 years (6.5–9.5 years) Number of medications (continuous) & Low grip strength Fried Phenotype (Persistent Transitions) OR = 0.92 per additional drug; OR = 0.80 per age year 0.86–0.99; 0.72–0.89 Age, sex, alcohol consumption, chair stands time, grip strength, uricemia

Costenoble et al. (2023) [65] Belgium (BUTTERFLY Cohort) 1 year Brussels Integrated ADL Inventory (a-ADL-DI) Adapted Fried Frailty Phenotype a-ADL Worsening OR = 1.048; Women Improvement OR = 3.682 1.01–1.09; 1.38–10.14 Age, sex, b-ADL, i-ADL, board membership change, resilience score

Tchalla et al. (2023) [66] France (FREEDOM Cohort) 6 years (75 months) GDS > 9 & MMS < 24 Fried Phenotype (Group-based Trajectories) OR = 10.51 (Depressive symptoms); OR = 4.56 (MMS < 24) 4.61–23.95; 1.97–10.58 Age, sex, hypertension, MMS, GDS, personalized care plan

Van der Ploeg et al. (2024) [67] Netherlands (Roosendaal Cohort) 9 years ≥2 chronic diseases & Unhealthy lifestyle (TFI) Groningen Activity Restriction Scale (GARS) GEE β = 5.3 (Multimorbidity); β = 6.6 (Unhealthy lifestyle) p = 0.002; p = 0.029 Age, sex, marital status, education, monthly income, life events

Suprawesta et al. (2024) [68] Taiwan (Taipei Community) 2 years Tinetti balance score, gait velocity, exercise habit Cognitive Frailty Trajectories (GBTM) OR = 0.58 (Regular exercise); OR = 2.90 (≥4 comorbidities) 0.36–0.94; 1.65–5.10 Age, sex, Tinetti balance, gait velocity, comorbidities, GDS score

Tanaka et al. (2025) [69] Japan (Kashiwa 9-year Cohort) 9 years (6.0 years median) Japan ACB score ≥ 3 vs. 0 CHS Index (Incident Frailty) & AWGS (Sarcopenia) Adjusted HR = 2.45 (Frailty); Adjusted HR = 2.01 (Sarcopenia) 1.52–3.94; 1.20–3.35 Age, sex, BMI, education, income, living alone, cognitive, depression, food diversity, exercise, alcohol, comorbidities, polypharmacy, PIMs

Dong et al. (2025) [70] China (Nationwide CLHLS 3 Cohorts) 9–10 years (65 mo median) Cox stepwise selected 9-variable panel Frailty Index (46 health deficits, cutoff ≥0.25) AUC = 0.80 at 96 mo; High Risk PI p < 0.0001 0.78–0.82 Age, logBMI, MMSE, sex, ethnicity, education, occupation, smoking, natural teeth

Stream 3: Baseline Frailty Predicting Downstream Outcomes

Lewis (2020) [55] Ireland (CVW Cohort) 90 days Rockwood CFS (Red/Amber/Green Wards) Institutionalization/Unplanned Hospital Admission OR = 14.77 (Red Ward > 7 HCPs); OR = 9.43 (Unstable state) 2.69–81.2; 1.82–48.9 Age, comorbidity index, baseline CFS, care hours, FRIDs

Varan et al. (2021) [57] Turkey (Gazi Outpatient) 47 months (median) Edmonton Frail Scale (EFS ≥ 7) & Fried Phenotype All-Cause Mortality HR = 2.84 (EFS Frail vs. Robust); HR = 2.12 (Fried Frail) 1.65–4.88; 1.34–3.35 Age, sex, comorbidity count, serum albumin, IADL, walking time

Lee et al. (2022) [61] South Korea (LPOPS Cohort) 3 years Multidimensional Frailty Profiles (Physical, Cog, Psych, Soc) Nursing Home Institutionalization & All-Cause Mortality SHR = 3.42 (Institutionalization); HR = 3.89 (Mortality for Mixed Frailty) 1.79–6.52; 2.57–5.85 Age, sex, smoking, alcohol drinking, comorbidities

Thompson et al. (2022) [62] Australia (FIRST Cohort Nursing Home) 1 year 47-item Frailty Index (Continuous deciles) 12-month All-Cause Mortality HR = 2.15 per FI decile increment 1.42–3.25 Age, sex, length of stay, comorbidities, polypharmacy, MNA status

Mielke et al. (2022) [60] Germany (BIS Cohort) 2.1 years Fried Frailty Phenotype (Frail vs. Robust) 2.1-year All-Cause Mortality 33% mortality in Frail Men vs. 18% in Frail Women p < 0.001 (Log-rank) Age, sex, eGFR, Charlson Comorbidity Index, hospitalization history

Tchalla et al. (2023) [66] France (FREEDOM Cohort) 6 years Frailty Trajectory D (Severe Worsening) Incident Falls & Unplanned Hospitalization Falls: 43.8% in Traj D vs. 18.5% in Traj B; Hosp: 29.0% p = 0.012 (Chi-squared across trajectories) Age, sex, hypertension, MMS score, GDS score

Tanaka et al. (2025) [69] Japan (Kashiwa Cohort) 9 years Baseline Anticholinergic Burden (ACB ≥ 3) & Physical Frailty Incident Sarcopenia & Long-Term Care Certification Adjusted HR = 2.01 (Sarcopenia); HR = 2.45 (Frailty) 1.20–3.35; 1.52–3.94 Age, sex, BMI, education, income, living alone, cognitive, depression, exercise, comorbidities, polypharmacy, PIMs

Abbreviations: 5STS, Five-Times Sit-to-Stand test; a-ADL-DI, Advanced Activities of Daily Living Disability Index; ACB, Anticholinergic Cognitive Burden; ACR, Albumin-to-Creatinine Ratio; ADL, Activities of Daily Living; AgeCoDe, German Study on Ageing, Cognition and Dementia in Primary Care Patients; AgeQualiDe, follow-up study on needs, health service use, costs and health-related quality of life in the oldest old; aOR, adjusted Odds Ratio; AUC, Area Under the Curve; AWGS, Asian Working Group for Sarcopenia; b-ADL, Basic Activities of Daily Living; BIS, Berlin Initiative Study; BMI, Body Mass Index; BUTTERFLY, Brussels Study on the Early Predictors of Frailty; CASP-19, Control, Autonomy, Self-realization and Pleasure (19-item scale); CFS, Clinical Frailty Scale; CHF, Congestive Heart Failure; CHS, Cardiovascular Health Study; CI, Confidence Interval; CLHLS, Chinese Longitudinal Healthy Longevity Survey; COPD, Chronic Obstructive Pulmonary Disease; CRP, C-Reactive Protein; CSHA-FI, Canadian Study of Health and Aging Frailty Index; CVW, Community Virtual Ward; DBP, Diastolic Blood Pressure; EAT-10/H-EAT-10, (Hebrew) Eating Assessment Tool-10; eGFR, estimated Glomerular Filtration Rate; EFS, Edmonton Frail Scale; ELSA, English Longitudinal Study of Ageing; FI, Frailty Index; FIBRA-BR, Rede FIBRA (Frailty in Brazilian Elderly Network Study); FIRST, Frailty In Residential Sector over Time; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Loss of weight; FREEDOM, FRagilité Et Evaluation à DOMicile (Frailty and Evaluation at Home); FRIDs, Fall-Risk-Increasing Drugs; GARS, Groningen Activity Restriction Scale; GBTM, Group-Based Trajectory Modeling; GDS(-SF), Geriatric Depression Scale (-Short Form); GEE, Generalized Estimating Equations; GFR, Glomerular Filtration Rate; GFST, Gerontopôle Frailty Screening Tool; GPCOG, General Practitioner Assessment of Cognition; HALFT, social frailty scale (Help, Participation, Loneliness, Financial difficulty, Talk); HCBS, Home and Community-Based Services; HCPs, Health Care Professionals; HGS, Handgrip Strength; HR, Hazard Ratio; HRR, Hemoglobin-to-Red cell distribution width Ratio; HRS, Health and Retirement Study; i-ADL/IADL, Instrumental Activities of Daily Living; IC, Impaired Cognition; interRAI-HC, interRAI Home Care; KCL, Kihon Checklist; KFACS, Korean Frailty and Aging Cohort Study; LASA, Longitudinal Aging Study Amsterdam; LPOPS, Living Profiles of Older People Survey; LSTM, Long Short-Term Memory; MI, Myocardial Infarction; MID, Minimal/Minimally Important Difference; MMS/MMSE, Mini-Mental State (Examination); MNA(-SF), Mini Nutritional Assessment (-Short Form); MUST, Malnutrition Universal Screening Tool; N/A, Not Applicable; OR, Odds Ratio; PASE, Physical Activity Scale for the Elderly; PIMs, Potentially Inappropriate Medications; RPOW, Relative muscle Power; SBP, Systolic Blood Pressure; SES, Socioeconomic Status; SHR, Subdistribution Hazard Ratio; SPPB, Short Physical Performance Battery; SRH, Self-Rated Health; STS, Sit-To-Stand test; TFI, Tilburg Frailty Indicator; Traj, Trajectory; TSHA, Toledo Study for Healthy Aging; TxB2, Thromboxane B2; 7-HDoHE, 7-Hydroxydocosahexaenoic acid; US, Ultrasound.

Table A7Distribution matrix of variables and health determinants identified in the studies included in the systematic review (n = 40).

Publication Age Functional Status/Physical Capacity/Exercise Co-mobility Depressive Symptoms/Sadness Gender Dementia/Memory Problems/Cognitive Status Malnutrition or Risk Malnut. Longer Walking Time/Asymmetry in Steps/Handstring Strength Polypharmacy/Medication Education Social Interaction BMI Fear of Falling/Fall History Poor Self-Rated Health Marital Status Smoke Health Behaviors/Resilience Socioeconomic Status Living Alone Alcohol Swallowing Hg or HRR (<9.97) or CRP Race/Ethnic. Oral Health/Teeth Status Sleep Uricemia Low Albumin Not visiting/Receiving Diabetes Pain Ulcer risk Anticholinergic Burden (Score ≥ 3) Hospitalization Long-Term Care Biomarkers: txb2 and 7-hdohe Low Health Literacy High Blood Pressure (Or Bad Control)

Hajek et al. (2020) [54]
✔

✔

✔

Lewis (2020) [55]

✔

✔

✔

✔

Steinmeyer et al. (2020) [32]

✔

✔

✔

✔

✔

Breij et al. (2020) [56]
✔

✔

✔

✔

✔

✔

✔

Venturini et al. (2021) [33]

✔

✔

✔

✔

✔

✔

✔

Varan et al. (2021) [57]

✔

✔

✔

✔

✔

Sharma et al. (2021) [34]
✔

✔

✔

✔

✔

✔

Jung et al. (2021) [35]
✔

✔

✔

Sella-Weiss et al. (2021) [36]
✔

✔

✔

✔

✔

Rodríguez-Laso et al. 2022 [58]
✔

✔

✔

✔

✔

✔

✔

✔

Tsay et al. (2021) [59]
✔

✔

✔

✔

✔

Tavares et al. (2022) [37]
✔

✔

✔

Mielke et al. (2022) [60]
✔

✔

✔

✔

✔

✔

✔

✔

Lee et al. (2022) [61]
✔

✔

✔

✔

✔

✔

✔

Thompson et al. (2022) [62]
✔

✔

✔

Yoshizawa (2022) [63]
✔

✔

✔

✔

✔

✔

✔

✔

Rodríguez-Laso et al. (2023) [64]
✔

✔

✔

✔

✔

✔

Mangin et al. (2023) [38]
✔

✔

✔

✔

Yan et al. (2023) [39]
✔

✔

✔

✔

✔

Huang et al. (2023) [31]
✔

✔

✔

✔

Zhu et al. (2023) [40]

✔

Costenoble et al. (2023) [65]

✔

✔

✔

Shakya et al. (2023) [41]

✔

✔

✔

Tchalla et al. (2023) [66]
✔

✔

✔

✔

✔

Cömertoğlu et al. (2023) [42]

✔

✔

✔

✔

✔

✔

Henry et al. (2023) [43]

✔

✔

Van der Ploeg et al. (2024) [67]

✔

✔

✔

Suprawesta et al. (2024) [68]
✔

✔

✔

✔

✔

Alamri & Mealif (2024) [44]

✔

✔

Johnson et al. (2024) [45]

✔

✔

✔

✔

Sepúlveda et al. (2024) [46]

✔

Güner et al. (2024) [47]
✔

✔

✔

✔

✔

✔

✔

✔

Tang et al. (2024) [48]
✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

Savci et al. (2025) [49]
✔

✔

✔

✔

✔

Tanaka et al. (2025) [69]
✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

Ceylan et al. (2025) [50]
✔

✔

✔

✔

✔

✔

Ortiz-Navarro et al. (2025) [51]

✔

✔

Park et al. (2025) [52]
✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

Dong et al. (2025) [70]
✔

✔

✔

✔

✔

✔

✔

✔

✔

Tzeng et al. (2025) [53]
✔

✔

✔

✔

✔

✔

✔

✔

Total 26 21 15 15 14 15 9 10 9 7 7 6 5 5 3 3 4 4 3 2 3 3 3 3 2 2 2 2 2 1 1 1 1 1 1 1 1

The checkmarks (✔) indicate that the variable was identified as a statistically significant factor or retained correlate/predictor in the corresponding study’s primary analytical model, with the final sum indicating the reporting frequency of each determinant across the selected literature (N = 40).

Table A8Study-specific definitions, operational thresholds, and adjusted findings for medication and polypharmacy reports (n = 9).

Author & Year Study Design & Setting Medication Exposure Name Specific Exposure Threshold/Definition Outcome Measure Follow-Up Duration Adjusted Effect Estimate (OR/HR/β) 95% Confidence Interval p-Value Adjusted Covariates in Model Reconciled Design Role

Sharma et al. (2021) [34] Cross-sectional, Acute Hospital Wards Polypharmacy Concurrent use of ≥5 daily medications Edmonton Frail Scale (EFS) Cross-sectional (Admission) Mean = 8.9 drugs (Frail) vs. 7.2 (Non-frail); p = 0.002 N/A (Mean diff) 0.002 Age, sex, MUST score, CCI, mobility, cognitive impairment Concurrent Correlate (Stream 1)

Alamri & Mealif (2024) [44] Cross-sectional, Hospital Inpatient Wards Polypharmacy Concurrent use of ≥5 daily medications Clinical Frailty Scale (CFS) Cross-sectional Prevalence 47.6% (Frail) vs. 21.4% (Non-frail); p = 0.003 N/A (Chi-squared) 0.003 Height, comorbidity index, heart failure Concurrent Correlate (Stream 1)

Ceylan et al. (2025) [50] Cross-sectional, Geriatric Outpatient Clinic Polypharmacy Concurrent use of ≥5 daily medications Clinical Frailty Scale (CFS ≥ 4) Cross-sectional 71.3% (LWF) vs. 55.3% (Robust); p < 0.001 N/A (Univariate retained) <0.001 Age, sex, HGS, education, multimorbidity Concurrent Correlate (Stream 1)

Güner et al. (2024) [47] Cross-sectional, Geriatric Outpatient Clinic Polypharmacy Concurrent use of ≥5 daily medications Clinical Frailty Scale (CFS) Cross-sectional 56.9% (Frail) vs. 45.7% (Non-frail); p = 0.015 N/A (Univariate retained) 0.015 Age, sex, BMI, MNA-SF, MMSE, GDS, HGS asymmetry Concurrent Correlate (Stream 1)

De Breij et al. (2021) [56] Longitudinal Prospective Cohort (LASA) Polypharmacy Concurrent use of ≥5 prescription drugs LASA Frailty Index (FI ≥ 0.25) 3 years (2008–09 to 2011–12) Adjusted aOR = 2.14 (39.6% baseline prevalence) 1.53–4.19 <0.001 Age, sex, depressive symptoms, chronic diseases, self-rated health, pain Longitudinal Predictor (Stream 2)—Reconciled as Cohort

Rodríguez-Laso et al. (2023) [64] Longitudinal Prospective Cohort (TSHA 3-Wave) Continuous Medication Count Total number of daily prescribed drugs (continuous) Persistent Prefrailty to Robustness Transition 8 years (6.5–9.5 years) Adjusted OR = 0.92 per additional drug 0.86–0.99 0.008 Age, grip strength, chair stands, drinking alcohol, uricemia Longitudinal Predictor (Stream 2)

Tanaka et al. (2025) [69] Longitudinal Prospective Cohort (Kashiwa) Anticholinergic Burden & Polypharmacy Japanese Anticholinergic Risk Scale (Score ≥ 3) & Polypharmacy (≥5 drugs) Incident Physical Frailty & Sarcopenia 9 years (6.0 years median) Adjusted HR = 2.45 (ACB ≥ 3 for Frailty); Adjusted HR = 2.01 (Sarcopenia) 1.52–3.94; 1.20–3.35 <0.001 Age, sex, BMI, education, income, living alone, cognitive, depression, food diversity, exercise, alcohol, comorbidities, polypharmacy, PIMs Longitudinal Predictor (Stream 2)

Mielke et al. (2022) [60] Longitudinal Prospective Cohort (BIS) Polypharmacy Concurrent use of ≥5 daily medications Fried Frailty Phenotype (Worsening/Improvement) 2.1 years 73% baseline prevalence; retained in candidate predictor pool N/A (Bootstrap selection) N/S in final model Age, self-rated health, cognitive impairment, CHF, stroke, diabetes, ACR, GFR, living alone, smoking Longitudinal Predictor (Stream 2)

Lewis (2020) [55] Prospective Correlational Cohort (CVW) Medication Count & FRIDs Mean medication count (8.24 SD 3.66) & Fall Risk Increasing Drugs (≥2 FRIDs) Virtual Ward Risk Stratification/Unstable State 90 days 42.0% taking ≥ 2 FRIDs; Beta-blockers 32.9%, ACEi 26.1% N/A (Baseline profile) <0.05 Age, sex, Charlson Comorbidity Index, baseline CFS, home care hours Longitudinal Cohort (Stream 2)

Note: Stream 1 denotes concurrent cross-sectional correlates; Stream 2 denotes baseline determinants prospectively predicting frailty onset, state progression, or transitions over longitudinal follow-up. Abbreviations: ACB, Anticholinergic Cognitive Burden; ACEi, Angiotensin Converting Enzyme inhibitors; ACR, Albumin-to-Creatinine Ratio; BIS, Berlin Initiative Study; BMI, Body Mass Index; CCI, Charlson Comorbidity Index; CFS, Clinical Frailty Scale; CHF, Congestive Heart Failure; CI, Confidence Interval; CVW, Community Virtual Ward; EFS, Edmonton Frail Scale; FI, Frailty Index; FRIDs, Fall-Risk-Increasing Drugs; GDS, Geriatric Depression Scale; GFR, Glomerular Filtration Rate; HGS, Handgrip Strength; HR, Hazard Ratio; LASA, Longitudinal Aging Study Amsterdam; LWF, Living With Frailty; MMSE, Mini-Mental State Examination; MNA-SF, Mini Nutritional Assessment—Short Form; MUST, Malnutrition Universal Screening Tool; N/A, Not Applicable; N/S, Not Significant; OR, Odds Ratio; PIMs, Potentially Inappropriate Medications; SD, Standard Deviation; TSHA, Toledo Study for Healthy Aging.

Appendix D

Table A9Bibliographic details, identifiers, and specific reasons for non-retrieval of full-text reports (n = 3).

ID Bibliographic Details (DOI) Base Identifiers Specific Reasons

UR-01 Vafaei, A., Wu, Y.Y., Curcio, C.L., Gomes, C.S., Auais, M., & Gomez, F. (2023). A Regression Tree Analysis to Identify Factors Predicting Frailty: The International Mobility in Aging Study. Gerontology, 69(2), 130–139. DOI: 10.1159/000526737. Scopus/Web of Science Community-dwelling older adults (≥65 years) Full text unavailable through institutional library subscriptions or interlibrary loan networks; no response from the corresponding author following electronic inquiry.

UR-02 Kulal, N., Nandakumar, G., & Vaishali, K. (2023). Prediction of Risk of Frailty Among Institutionalized Older Adults in India. Current Aging Science, 16(1), 33–39. DOI: 10.2174/1874609815666220415130737. PubMed/Scopus Institutionalized older adults in residential care facilities Document access restricted by publisher paywall with no institutional license available; repository link inactive; reprint request to authors unanswered.

UR-03 Ponvel, P., Murukesu, R.R., Shahar, S., Rivan, N.F.M., Subramaniam, P., & Singh, D.K.A. (2024). Transition of Physical, Psychological, and Cognitive Frailty and Its Associated Determinants in Malaysian Older Adults: A 5-Year Follow-up Study. Rejuvenation Research, pp. 198–206. DOI: 10.1089/rej.2024.0047 [not independently verified]. CINAHL (EBSCOhost)
Full text unavailable through institutional library subscriptions or interlibrary loan networks; no response from the corresponding author following electronic inquiry.

Note: These reports were sought for retrieval during the eligibility screening phase (Figure 1, PRISMA 2020 flow diagram) but could not be obtained in full text prior to the eligibility assessment deadline. Author names and bibliographic details were cross-checked against the original publisher records; the DOI for UR-03 could not be independently confirmed at the time of writing and should be verified against the publisher’s record before submission.

Author Contributions

Conceptualization, H.P., A.R., C.F. and A.C.; methodology, H.P., A.R., C.F. and A.C.; software, H.P. and A.R.; validation, H.P., A.R., C.F., E.A. and A.C.; formal analysis, H.P., A.R., C.F., E.A. and A.C.; investigation, H.P., A.R., C.F., E.A. and A.C.; resources, H.P., A.R., C.F., E.A. and A.C.; data curation, H.P., A.R. and A.C.; writing—original draft preparation, H.P. and A.R.; writing—review and editing, H.P., A.R. and A.C.; visualization, H.P., A.R., C.F., E.A. and A.C.; supervision, A.R., C.F., E.A. and A.C.; project administration, H.P., A.R., C.F., E.A. and A.C. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable (this study is a systematic review of published literature and did not involve direct human participants or animals).

Informed Consent Statement

Not applicable.

Data Availability Statement

All datasets generated and analyzed during this systematic review are included within this article, its main tables (Table 1, Table 2 and Table 3), and its appendixes ([Table A2, Table A3, Table A4, Table A5, Table A6, Table A7, Table A8 and Table A9]). The search queries and initial screening yields are publicly accessible via the registered PROSPERO record (CRD420251039796). Additional screening records and extraction sheets are available from the corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research received no external funding.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

References

1.United Nations . World Population Prospects 2024: Summary of Results. United Nations; New York, NY, USA: 2024.

2.Ferrari A.J., Santomauro D.F., Aali A., Abate Y.H., Abbafati C., Abbastabar H., Abd ElHafeez S., Abdelmasseh M., Abd-Elsalam S., Abdollahi A., et al. Global Incidence, Prevalence, Years Lived with Disability (YLDs), Disability-Adjusted Life-Years (DALYs), and Healthy Life Expectancy (HALE) for 371 Diseases and Injuries in 204 Countries and Territories and 811 Subnational Locations, 1990–2021: A Systematic Analysis for the Global Burden of Disease Study 2021. Lancet. 2024;403:2133–2161. doi: 10.1016/S0140-6736(24)00757-8.

3.Li J., Li J., Xu X., Shen W., Sun Y., Fu Y., Tan X., Wang N., Qi L., Wang B., et al. Social Determinants of Health, Accelerated Biological Aging, and Long-Term Health Outcomes. Nat. Commun. 2025;17:900. doi: 10.1038/s41467-025-67622-7.

4.Brocklehurst H., Laurenson M. A Concept Analysis Examining the Vulnerability of Older People. Br. J. Nurs. 2008;17:1354–1357. doi: 10.12968/bjon.2008.17.21.31738.

5.Sanchini V., Sala R., Gastmans C. The Concept of Vulnerability in Aged Care: A Systematic Review of Argument-Based Ethics Literature. BMC Med. Ethics. 2022;23:84. doi: 10.1186/s12910-022-00819-3.

6.Fernandes Bolina A., Rodrigues R.A.P., Tavares D.M.D.S., Haas V.J. Factors Associated with the Social, Individual and Programmatic Vulnerability of Older Adults Living at Home. Rev. Esc. Enferm. USP. 2019;53:e03429. doi: 10.1590/s1980-220x2017050103429.

7.Huang Z.-T., Lai E.T.C., Luo Y., Woo J. Social Determinants of Intrinsic Capacity: A Systematic Review of Observational Studies. Ageing Res. Rev. 2024;95:102239. doi: 10.1016/j.arr.2024.102239.

8.Fried L.P., Tangen C.M., Walston J., Newman A.B., Hirsch C., Gottdiener J., Seeman T., Tracy R., Kop W.J., Burke G., et al. Frailty in Older Adults: Evidence for a Phenotype. J. Gerontol. Ser. A Biol. Sci. Med. Sci. 2001;56:M146–M157. doi: 10.1093/gerona/56.3.M146.

9.Kojima S., Murotani K., Zhou B., Kothari K.U., Fukushima M., Nagai Y. Assessing Long-Term Care Risk in Older Individuals with Possible Cognitive Decline: A Large Population-Based Study Using the Kihon Checklist. Geriatr. Gerontol. Int. 2019;19:598–603. doi: 10.1111/ggi.13677.

10.Lin K.-P., Li H.-Y., Chen J.-H., Lu F.-P., Wen C.-J., Chou Y.-C., Wu M.-C., Chan D.-C., Chen Y.-M. Prediction of Adverse Health Outcomes Using an Electronic Frailty Index among Nonfrail and Prefrail Community Elders. BMC Geriatr. 2023;23:474. doi: 10.1186/s12877-023-04160-1.

11.Christensen M.G., Jacobsen K.K., Nilsson C., Jepsen R., Thygesen L., Suetta C., Holm E.A. Prevalence and Population Characteristics Associated with Frailty in a Rural Low Socioeconomic Area in Denmark: The Lolland-Falster Health Study. BMJ Open. 2024;14:e073754. doi: 10.1136/bmjopen-2023-073754.

12.OECD . Is Care Affordable for Older People? OECD Publishing; Paris, France: 2024. (OECD Health Policy Studies).

13.Micheli K., Ratsika N., Vozikaki M., Chlouverakis G., Philalithis A. Family Ties and Functional Limitation in the Elderly: Results from the Survey of Health Ageing and Retirement in Europe (SHARE) Arch. Gerontol. Geriatr. 2018;78:23–29. doi: 10.1016/j.archger.2018.05.023.

14.Deng Y., Zhang K., Zhu J., Hu X., Liao R. Healthy Aging, Early Screening, and Interventions for Frailty in the Elderly. BioSci. Trends. 2023;17:252–261. doi: 10.5582/bst.2023.01204.

15.Nari F., Park E.-C., Nam C.-M., Jang S.-I. Impact of Frailty on Mortality and Healthcare Costs and Utilization among Older Adults in South Korea. Sci. Rep. 2023;13:21203. doi: 10.1038/s41598-023-48403-y.

16.Kokorelias K.M., Cronin S.M., Munce S.E.P., Eftekhar P., McGilton K.S., Vellani S., Colella T.J.F., Kontos P., Grigorovich A., Furlan A., et al. Conceptualization of Frailty in Rehabilitation Interventions with Adults: A Scoping Review. Disabil. Rehabil. 2023;45:117–153. doi: 10.1080/09638288.2021.2012844.

17.World Health Organization . WHO Guideline on Self-Care Interventions for Health and Well-Being, 2022 Revision. World Health Organization; Geneva, Switzerland: 2022.

18.Whelan E.G. Analysis and Application of Dorothea Orem’s Self-Care Practice Model. J. Nurs. Educ. 1984;23:342–345. doi: 10.3928/0148-4834-19841001-07.

19.Cavanagh S.J. Orem’s Model in Action. Macmillan Education; London, UK: 1991.

20.Attaallah S.A., Peters R.M., Benkert R., Yarandi H., Oliver-McNeil S., Hopp F. Developing a Middle-Range Theory of Heart Failure Self-Care. Nurs. Sci. Q. 2021;34:168–177. doi: 10.1177/0894318420987164.

21.Ramos A., Fonseca C., Pinho L., Lopes M., Brites R., Henriques A. Assessment of Functioning in Older Adults Hospitalized in Long-Term Care in Portugal: Analysis of a Big Data. Front. Med. 2022;9:780364. doi: 10.3389/fmed.2022.780364.

22.Charalambous A., editor. Informal Caregivers: From Hidden Heroes to Integral Part of Care. Springer International Publishing; Cham, Switzerland: 2023.

23.Levasseur M., Lussier-Therrien M., Biron M.L., Dubois M., Boissy P., Naud D., Dubuc N., Coallier J., Calvé J., Audet M. Scoping Study of Definitions and Instruments Measuring Vulnerability in Older Adults. J. Am. Geriatr. Soc. 2022;70:269–280. doi: 10.1111/jgs.17451. Erratum in J. Am. Geriatr. Soc. 2023, 71, E14.

24.Leghissa M., Carrera Á., Iglesias C.A. Machine Learning Approaches for Frailty Detection, Prediction and Classification in Elderly People: A Systematic Review. Int. J. Med. Inform. 2023;178:105172. doi: 10.1016/j.ijmedinf.2023.105172.

25.Dent E., Morley J.E., Cruz-Jentoft A.J., Woodhouse L., Rodríguez-Mañas L., Fried L.P., Woo J., Aprahamian I., Sanford A., Lundy J., et al. Physical Frailty: ICFSR International Clinical Practice Guidelines for Identification and Management. J. Nutr. Health Aging. 2019;23:771–787. doi: 10.1007/s12603-019-1273-z.

26.Pino H., Ramos A., Coelho A. Predicted Vulnerability in People Aged 65 and Over: Systematic Review Protocol. In: Moreira J., Bico I., Moguel E., Alves E., Fonseca C., Ferreira R., editors. Gerontechnology VII. Springer Nature; Cham, Switzerland: 2026. pp. 90–97. (Lecture Notes in Bioengineering).

27.De Cassai A., Dost B., Tulgar S., Boscolo A. Methodological Standards for Conducting High-Quality Systematic Reviews. Biology. 2025;14:973. doi: 10.3390/biology14080973.

28.Morgan R.L., Whaley P., Thayer K.A., Schünemann H.J. Identifying the PECO: A Framework for Formulating Good Questions to Explore the Association of Environmental and Other Exposures with Health Outcomes. Environ. Int. 2018;121:1027–1031. doi: 10.1016/j.envint.2018.07.015.

29.Haddaway N.R., Collins A.M., Coughlin D., Kirk S. The Role of Google Scholar in Evidence Reviews and Its Applicability to Grey Literature Searching. PLoS ONE. 2015;10:e0138237. doi: 10.1371/journal.pone.0138237.

30.Jia R.M., Stern C. The Inclusion or Exclusion of Studies Based on Critical Appraisal Results in JBI Qualitative Systematic Reviews: An Analysis of Practices. Res. Synth. Methods. 2026;17:277–292. doi: 10.1017/rsm.2025.10042.

31.Huang T.-Y., Chou M.-Y., Liang C.-K., Lin Y.-T., Chen R.-Y., Wu P.-F. Physical Activity Plays a Crucial Role in Multidomain Intervention for Frailty Prevention. Aging Clin. Exp. Res. 2023;35:1283–1292. doi: 10.1007/s40520-023-02412-z.

32.Steinmeyer Z., Delpierre C., Soriano G., Steinmeyer A., Ysebaert L., Balardy L., Sourdet S. Hemoglobin Concentration; a Pathway to Frailty. BMC Geriatr. 2020;20:202. doi: 10.1186/s12877-020-01597-6.

33.Venturini C., Sampaio R.F., De Souza Moreira B., Ferriolli E., Neri A.L., Lourenço R.A., Lustosa L.P. A Multidimensional Approach to Frailty Compared with Physical Phenotype in Older Brazilian Adults: Data from the FIBRA-BR Study. BMC Geriatr. 2021;21:246. doi: 10.1186/s12877-021-02193-y.

34.Sharma Y., Avina P., Ross E., Horwood C., Hakendorf P., Thompson C. Nutritional Risk Assessed by the Malnutrition Universal Screening Tool as a Predictor of Frailty in Acutely Hospitalised Older Patients: An Observational Study. Asia Pac. J. Clin. Nutr. 2021;30:185–191. doi: 10.6133/apjcn.202106_30(2).0001.

35.Jung D., Kim J., Kim M., Won C.W., Mun K.-R. Frailty Assessment Using Temporal Gait Characteristics and a Long Short-Term Memory Network. IEEE J. Biomed. Health Inform. 2021;25:3649–3658. doi: 10.1109/JBHI.2021.3067931.

36.Sella-Weiss O. Association between Swallowing Function, Malnutrition and Frailty in Community Dwelling Older People. Clin. Nutr. ESPEN. 2021;45:476–485. doi: 10.1016/j.clnesp.2021.06.028.

37.Tavares J.P.D.A., Sá-Couto P.M.F.D., Pedreira L.C. Predictors of Frailty in Older People Users of Primary Health Care. Rev. Bras. Enferm. 2022;75:e20201292. doi: 10.1590/0034-7167-2020-1292.

38.Mangin D., Lawson J., Risdon C., Siu H.Y.-H., Packer T., Wong S.T., Howard M. Association between Frailty, Chronic Conditions and Socioeconomic Status in Community-Dwelling Older Adults Attending Primary Care: A Cross-Sectional Study Using Practice-Based Research Network Data. BMJ Open. 2023;13:e066269. doi: 10.1136/bmjopen-2022-066269.

39.Yan G., Wan Y., Ren J.-H., Zhao Y., Wang Y., Shen J.-H. Analysis of Systolic and Diastolic Blood Pressure Variability in Frail, Pre-Frail, and Non-Frail Elderly Patients: The Relationship between Frailty Syndrome and Blood Pressure Variability in the Elderly. Medicine. 2023;102:e32874. doi: 10.1097/MD.0000000000032874.

40.Zhu M., Wei C., Yang X., Huang Y., Xu Y., Xiong Z. Lower Haemoglobin-to-Red Blood Cell Distribution Width Ratio Is Independently Associated with Frailty in Community-Dwelling Older Adults: A Cross-Sectional Study. BMJ Open. 2023;13:e069141. doi: 10.1136/bmjopen-2022-069141.

41.Shakya S., Silva S.G., McConnell E.S., McLaughlin S.J., Cary M.P. Does Cumulative Psychosocial Stress Explain Frailty Disparities in Community-Dwelling Older Adults? Arch. Gerontol. Geriatr. 2023;113:105055. doi: 10.1016/j.archger.2023.105055.

42.Cömertoğlu E.O., Ozturk Y., Hafizoglu M., Kahyaoglu Z., Cavusoglu C., Balci C., Dogu B.B., Halil M., Cankurtaran M. The Effect of Social Frailty on Mental Health and Quality of Life in Older People: A Cross-Sectional Study. Res. Sq. 2023 doi: 10.21203/rs.3.rs-2852088/v1.

43.Henry J.D., Coundouris S.P., Mead J., Thompson B., Hubbard R.E., Grainger S.A. Social Frailty in Late Adulthood: Social Cognitive and Psychological Well-Being Correlates. J. Gerontol. Ser. B. 2023;78:87–96. doi: 10.1093/geronb/gbac157.

44.Alamri S.H., Mealif H.M. Exploring the Correlates of Frailty among Hospitalized Older Adults: A Cross-Sectional Study in a Saudi Teaching Hospital. Medicine. 2024;103:e38603. doi: 10.1097/MD.0000000000038603.

45.Johnson L., Guthrie B., Kelly P.A.T., Anand A., Marshall A., Seth S. Frailty or Frailties: Exploring Frailty Index Subdimensions in the English Longitudinal Study of Ageing. J. Epidemiol. Community Health. 2024;78:609–615. doi: 10.1136/jech-2023-221829.

46.Sepúlveda M., Palomo I., Montecino-Garrido H., Wehinger S., Rodriguez-Mañas L., Trostchansky A., Fuentes E. Physiological Changes Associated with Aging: Identification of Novel Biomarkers for Frailty Syndrome in Women. Free Radic. Biol. Med. 2024;223:160–171. doi: 10.1016/j.freeradbiomed.2024.07.022.

47.Güner M., Ceylan S., Okyar Baş A., Koca M., Öztürk Y., Doğu B.B., Halil M.G., Cankurtaran M., Balci C. A Hand-in-Hand Phenomenon in Older Adults: Increased Risk of Frailty in Geriatric Outpatients Associated with Handgrip Strength Asymmetry and Weakness. Clin. Nutr. 2024;43:2381–2387. doi: 10.1016/j.clnu.2024.09.010.

48.Tang X., Shen D., Zhou T., Ge S., Wu X., Wang A., Li M., Xia Y. Development of a Frailty Prediction Model Among Older Adults in China: A Cross-Sectional Analysis Using the Chinese Longitudinal Healthy Longevity Survey. Nurs. Open. 2024;11:e70070. doi: 10.1002/nop2.70070.

49.Savci C., İlhan N., Yıldırım S. The Relationships between Dysphagia Risk, Frailty, and Depression in Older Adults Living in Nursing Homes. Geriatr. Nurs. 2025;62:116–121. doi: 10.1016/j.gerinurse.2025.02.001.

50.Ceylan S., Ozturk Y., Guner M., Okyar Bas A., Koca M., Balci C., Dogu B.B., Cankurtaran M., Halil M.G. Can Handgrip Strength Alone Detect Individuals Living with Frailty According to the Clinical Frailty Scale? Rev. Assoc. Med. Bras. 2025;71:e20240237. doi: 10.1590/1806-9282.20240237.

51.Ortiz-Navarro B., Losa-Reyna J., Mihaiescu-Ion V., Garcia-Romero J., Carrillo De Albornoz-Gil M., Galán-Mercant A. Identification of Target Body Composition Parameters by Dual-Energy X-Ray Absorptiometry, Bioelectrical Impedance, and Ultrasonography to Detect Older Adults With Frailty and Prefrailty Status Using a Mobile App in Primary Care Services: Descriptive Cross-Sectional Study. JMIR Aging. 2025;8:e67982. doi: 10.2196/67982.

52.Park C., Kim N., Won C.W., Kim M. Predicting Cognitive Frailty in Community-Dwelling Older Adults: A Machine Learning Approach Based on Multidomain Risk Factors. Sci. Rep. 2025;15:18369. doi: 10.1038/s41598-025-00844-3.

53.Tzeng Y.-M., Kao S., Chen W.-S., Chen S.-F., Li S.-R., Chiu Y.-L., Chang Y.-T., Chang Y.-W. Examining Frailty Phenotypes of Community-Dwelling Older Adults in Taiwan Using the Falls Risk for Older People in the Community—Taiwan Version (Tw-FROP-Com) Aging. 2025;17:1011–1025. doi: 10.18632/aging.206231.

54.Hajek A., Brettschneider C., Röhr S., Gühne U., van der Leeden C., Lühmann D., Mamone S., Wiese B., Weyerer S., Werle J., et al. Which Factors Contribute to Frailty among the Oldest Old? Results of the Multicentre Prospective AgeCoDe and AgeQualiDe Study. Gerontology. 2020;66:460–466. doi: 10.1159/000508723.

55.Lewis C., O’Caoimh R., Patton D., O’Connor T., Moore Z., Nugent L.E. Risk Prediction for Adverse Outcomes for Frail Older Persons with Complex Healthcare and Social Care Needs Admitted to a Community Virtual Ward Model. Clin. Interv. Aging. 2020;15:915–926. doi: 10.2147/CIA.S236895.

56.De Breij S., Van Hout H.P.J., De Bruin S.R., Schuster N.A., Deeg D.J.H., Huisman M., Hoogendijk E.O. Predictors of Frailty and Vitality in Older Adults Aged 75 Years and Over: Results from the Longitudinal Aging Study Amsterdam. Gerontology. 2021;67:69–77. doi: 10.1159/000512049.

57.Varan H.D., Kizilarslanoglu M.C., Dogrul R.T., Arik G., Esme M., Kara O., Kilic M.K., Halil M., Cankurtaran M., Yavuz B.B. Comparative Evaluation of Predictive Ability of Comprehensive Geriatric Assessment Components Including Frailty on Long-Term Mortality. Exp. Aging Res. 2021;47:220–231. doi: 10.1080/0361073X.2021.1876395.

58.Rodríguez-Laso Á., García-García F.J., Rodríguez-Mañas L. Transitions Between Frailty States and Its Predictors in a Cohort of Community-Dwelling Spaniards. J. Am. Med. Dir. Assoc. 2022;23:524.e1–524.e11. doi: 10.1016/j.jamda.2021.07.008.

59.Tsay T., Shugrue N., Charles D., Migneault D., McManus R., Gruman C., Robison J. Predictors of Frailty Change in Home and Community-Based Services Populations. J. Am. Med. Dir. Assoc. 2021;23:838–844. doi: 10.1016/j.jamda.2021.07.032.

60.Mielke N., Schneider A., Huscher D., Ebert N., Schaeffner E. Gender Differences in Frailty Transition and Its Prediction in Community-Dwelling Old Adults. Sci. Rep. 2022;12:7341. doi: 10.1038/s41598-022-11358-7.

61.Lee Y., Kim E., Yun J., Chuck K.W. The Influence of Multiple Frailty Profiles on Institutionalization and All-cause Mortality in Community-living Older Adults. J. Cachexia Sarcopenia Muscle. 2022;13:2322–2330. doi: 10.1002/jcsm.13033.

62.Thompson M.Q., Jadczak A.D., Tucker G.R., Theou O., Visvanathan R. Frailty Change Based on Minimally Important Difference in Nursing Home Residents: FIRST Cohort Study Findings. Age Ageing. 2022;51:afac246. doi: 10.1093/ageing/afac246.

63.Yoshizawa Y., Tanaka T., Takahashi K., Fujisaki-Sueda-Sakai M., Son B., Iijima K. Impact of Health Literacy on the Progression of Frailty after 4 Years among Community-Dwelling Older Adults. Int. J. Environ. Res. Public Health. 2022;19:394. doi: 10.3390/ijerph19010394.

64.Rodríguez-Laso Á., García-García F.J., Rodríguez-Mañas L. Predictors of Maintained Transitions Between Robustness and Prefrailty in Community-Dwelling Older Spaniards. J. Am. Med. Dir. Assoc. 2023;24:57–64. doi: 10.1016/j.jamda.2022.10.010.

65.Costenoble A., Knoop V., Debain A., Bautmans I., Van Laere S., Lieten S., Rossi G., Verté D., Gorus E., De Vriendt P., et al. Transitions in Robust and Prefrail Octogenarians after 1 Year: The Influence of Activities of Daily Living, Social Participation, and Psychological Resilience on the Frailty State. BMC Geriatr. 2023;23:485. doi: 10.1186/s12877-023-04178-5.

66.Tchalla A., Cardinaud N., Gayot C., Dumoitier N., Druet-Cabanac M., Laroche M.-L., Rudelle K., Mouret C.L., Boyer S. Patterns, Predictors, and Outcomes of Frailty Trajectories in Community-Dwelling Older Adults: Results from the FREEDOM Cohort Study. Arch. Gerontol. Geriatr. 2023;114:105101. doi: 10.1016/j.archger.2023.105101.

67.Van Der Ploeg T., Gobbens R.J.J. Disability Transitions in Dutch Community-Dwelling Older People Aged 75 Years or Older. Arch. Gerontol. Geriatr. 2024;116:105165. doi: 10.1016/j.archger.2023.105165.

68.Suprawesta L., Chen S.-J., Liang H.-Y., Hwang H.-F., Yu W.-Y., Lin M.-R. Factors Affecting Cognitive Frailty Improvement and Progression in Taiwanese Older Adults. BMC Geriatr. 2024;24:105. doi: 10.1186/s12877-024-04700-3.

69.Tanaka T., Akishita M., Kojima T., Son B., Iijima K. Anticholinergic Burden Quantified Using the Japanese Risk Scale as a Predictor of Frailty and Sarcopenia among Community-dwelling Older Adults: A 9-year Kashiwa Cohort Study. Geriatr. Gerontol. Int. 2025;25:520–527. doi: 10.1111/ggi.70012.

70.Dong Y., Wang Q., Zhang K., Wang X., Liu H., Chen Y., Tang Z., Tan L. Development and Validation of a Prediction Model of Frailty Risk in Community-Dwelling Older Adults: From a National Longitudinal Survey. Public Health. 2025;240:63–70. doi: 10.1016/j.puhe.2024.12.055.

Associated Data

Data Availability Statement

All datasets generated and analyzed during this systematic review are included within this article, its main tables (Table 1, Table 2 and Table 3), and its appendixes ([Table A2, Table A3, Table A4, Table A5, Table A6, Table A7, Table A8 and Table A9]). The search queries and initial screening yields are publicly accessible via the registered PROSPERO record (CRD420251039796). Additional screening records and extraction sheets are available from the corresponding author upon reasonable request.