Characterization of social frailty domains and related adverse health outcomes in the Asia-Pacific: a systematic literature review

Theoretical framework and conceptual background of the study

The conceptual bases of this study were developed from Gobbens’s and Bunt’s frameworks on the determinants of social frailty (Bunt et al., 2017; Gobbens et al., 2010). Gobbens et al. (2010) suggested three main frailty domains: physical, psychological, and social. Gobbens et al. (2010) highlighted two sub-domains for social frailty, including social network and social support; achieving these domains enhances the quality of life and well-being as people age. Bunt et al. (2017) concept of social frailty developed based on the social needs concept of Social Production Function theory. Based on this theory, Bunt et al. (2017) confirmed that social frailty in old age is affected by four social factors: social needs, social resources (such as social supports and social networks), social behaviors or social activities, and general resources (indirect way of fulfilling social needs, such as living situation, educational level, and income or financial status). As a result, the context of this review was applied to the selected studies and assessment tools to extract data on the suggested social domains for social frailty.

Rationale of the systematic review

Elderly social frailty is an emerging health challenge associated with various health outcomes (Abe et al., 2019). Even though many review studies (Faller et al., 2019; Bunt et al., 2017; Bessa, Ribeiro & Coelho, 2018) conducted on older people’s frailty, the field still lacks comprehensive systematic literature reviews that integrate social frailty domains and their health outcomes. The existing reviews only assessed the social dimension of frailty and did not characterize the relevant health outcomes to each social domain. To address this gap, the current study systematically reviewed 31 journal articles selected from 1,896 text materials to address this gap. It followed a rigorous systematic method to ensure the future replicability of results by employing a clear review protocol and limiting bias in results by only focusing on empirical evidence (Moher et al., 2009). It also systematically analyzed the relevant scientific findings of the selected journal articles and showed whether they are consistent and can be generalized across settings or whether results vary significantly by subsets.

To describe the frailty syndrome in older people, existing review studies focused on reviewing the findings of the frailty assessment tools (Duppen et al., 2019; Bessa, Ribeiro & Coelho, 2018). Despite the unmanageable amounts of information on the frailty assessment tools, there is a need for systematic reviews to integrate efficiently existing knowledge to understand better the adverse health outcomes related to each social domain and contribute to providing data for rational decision-making by healthcare providers and policymakers. There is a significant knowledge gap in the Asia-Pacific context regarding frailty’s social dimensions and the possible health outcomes related to each domain (Faller et al., 2019). There is, therefore, a need to provide a better understanding of social frailty and its domains and health outcomes in Asia-Pacific settings, the world’s largest population of older people. Thus, the current study focused on the frailty tools that measure social frailty and its health outcomes in Asia-Pacific settings.

More specifically, the current systematic review contributed to (1) being one of the early systematic literature reviews that covered a wide range of social domains and determined the related adverse health outcomes to each domain, (2) reviewing the existing frailty screening tools in the Asia Pacific settings and characterize social domains and relevant adverse outcomes, (3) a better understanding of the existing theoretical model of social frailty proposed by existing evidence by identifying their related adverse health outcomes, (4) facilitating healthcare providers and researchers in making informed decisions, (5) contributing to linking social health science to public health and health sciences.

Search strategy

The first online search was conducted in January 2021 for studies over the last 20 years. The search was updated in February 2022 and 2023 to ensure the inclusion of recent studies conducted between January 2021 and January 2023. Two reviewers independently searched three electronic databases: PubMed, Scopus, and ScienceDirect; a manual search was conducted for the references of the complete full-text content using Google Scholar. The following search string and MeSH terms were used in the current systematic review: (“social frailty” OR “frailty social domain” OR “frailty social factor” OR “social vulnerability” OR “social vulnerabilities”) AND (“frailty screening tools” OR “frailty assessment tools” OR instruments OR “assessment tools” OR “screening tools”) AND (“risk factors” OR negative impacts OR “adverse outcomes” OR “adverse health outcomes”) AND (“Aged” (MeSH Terms) OR “older adults” (MeSH Terms) OR elderly (MeSH Terms)) AND (Asia-Pacific (MeSH Terms) OR “Southeastern Asia communit*” (MeSH Terms)).

Truncation, Boolean operators, parentheses, wildcards, quotation marks, and MeSH terms or search terms related to the described keywords were applied whenever possible. The search string was used only for the title, abstract, and keywords. Figure 1 shows the PRISMA flow diagram for the steps of the systematic literature search.

Inclusion and exclusion criteria

The study’s inclusion and exclusion criteria were: (a) publication year and language of the study: including only studies published between 2002 and 2023 in the English language. Studies published before 2002 in a language other than English were excluded. According to Bessa, Ribeiro & Coelho (2018), social frailty is an emerging health-related topic, with most of the related literature evolving in the last 20 years. (b) Study type: including only empirical evidence published in peer-reviewed journals (e.g., quantitative, qualitative, or mixed-methods). Review studies, books, reports, or other text materials not published in peer-reviewed indexed journals were excluded to ensure the quality of the selected studies. (c) Location: The studies’ location must be in the Asia-Pacific region; this is based on the objective of the review. (d) Study scope: selected studies must include the assessment of the social domain of frailty. Studies that did not include any social domain of frailty were excluded to ensure that the review question was addressed. (e) Target group: The sample of the studies must be community-dwelling older adults aged 60 years or older. Studies from the Asia Pacific identified the frailty syndrome of older people aged 60 and over based on retirement possibilities; people aged 60 and above are defined as senior citizens in many Asia Pacific countries (Lian et al., 2020; Pek et al., 2020).

Study selection

In PRISMA’s first and second steps (identification and screening), two reviewers (Author 2 and Author 3) independently reviewed the articles’ titles and abstracts to include relevant articles. Studies written in English between 2002 and 2023 were included, and duplicates were excluded. In the title and abstract screening, studies were included if they described and had the social frailty determinants, assessment tools, and domains or if they included a combination of these variables. Then, all articles selected by either one reviewer or both of them were included in the next step.

In the second step of PRISMA (eligibility), the included articles from the previous step were retrieved in full text from the same reviewers using the eligibility criteria mentioned above. A manual search was conducted on the reference lists of the selected articles to avoid any missing articles in the search process during the previous step. Ultimately, the two independent reviewers engaged in a comprehensive discussion regarding the articles and reached a consensus on their inclusion. If an agreement could not be achieved, the ultimate determination was made by a third impartial reviewer (Author 1) using the Newcastle-Ottawa Scale (NOS). After selecting the included studies, a summary table was reported to present data from the studies, see Document S2. The included studies were then analyzed and synthesized by the two authors using Atlas.ti.9. All five authors had approved the steps of the search protocol.

Data extraction and synthesis

The two reviewers extracted all data from the selected articles that could be related to social frailty, its domains, health outcomes, and assessment tools. To include a broad scope of factors related to social frailty, the data extracted is based on the definition and conceptual framework of social frailty. For example, each criterion included in the existing frailty index and showed an effect on social frailty was reviewed, such as socioeconomic status, loneliness, or isolation. Gobbens et al.’s (2010) and Bunt et al.’s (2017) frameworks were used to synthesize the criteria found and based on the factors highlighted in the conceptual framework of the current study. This conceptual background was used to have a comprehensive framework that integrates all domains and assessment tools associated with social frailty. Therefore, three themes to extract the data were (a) frailty assessment tools, (b) social domains of frailty, and (c) adverse outcomes related to social frailty and its domains. The three themes were identified based on the systematic review questions and objectives. For the first theme (frailty assessment tools), the data was synthesized and analyzed in sixteen categories describing each assessment tool’s main definition and social frailty domains. The data was synthesized in six codes and four categories in the second theme (social domains of frailty). The third theme (adverse outcomes related to social frailty and its domains) included eleven codes (health outcome) and four categories. The four categories of the second and third themes included (1) social resources, (2) social needs, (3) social behaviors (or social activities), and (4) general resources, see Tables 1 and 2.

Study quality assessment

The two reviewers (Author 2 & Author 3) independently carried out a quality assessment of the included studies using the Newcastle-Ottawa Scale (NOS) (Wells et al., 2000). Any disagreements regarding the included studies were resolved in an online meeting among the authors. The impartial reviewer (Author 1) did the final review using the Newcastle-Ottawa Scale (NOS) to solve the disagreement and decide on the included study. The NOS is a risk-of-bias tool for nonrandomized studies, which assesses three domains, including study selection (four factors), study comparability (one factor), and exposure and outcomes (two factors). In general, studies with a low risk of bias (NOS score from 6 to 7 points) were included in the current review. Studies with a high risk of bias (NOS score from 4 to 5 points) and a very high risk of bias (NOS score from 0 to 3 points) were excluded (Wells et al., 2000), see Document S3.

Selected study characteristics

For the individual studies, 11 were cross-sectional studies (secondary data analysis), 13 were longitudinal studies (secondary data analysis), six were cross-sectional surveys (primary data analysis), and one was the experts’ interview method (preliminary data analysis). These studies were conducted in five Asia-Pacific countries, including Japan (n = 17), China (n = 6), South Korea (n = 4), Singapore (n = 2), and Thailand (n = 1). Besides that, one of the selected studies was conducted in a general Asia-Pacific community of older adults without naming the country where the study was conducted. Of these 31 selected articles, 16 frailty assessment instruments with relevant social indicators were reviewed. Eight frailty instruments were published/designed in Chinese, five published in Japanese, two in Korean, and one in English. For more details of the individual study, Document S2 provided a detailed summary of the included study.

Frailty assessment tools

This section describes the first theme regarding the assessment tools designed to measure social frailty syndrome in older adults in Asia-Pacific region. It reports sixteen frailty assessment tools (n = 16 tools) and six social domains of frailty (six codes), see Table 1. The most used tool (n = 13 studies) was the Social Frailty Questionnaire (SFQ) by Makizako et al. (2015), a five-item self-reported questionnaire used to assess three social domains: social activities (n = 2), social role (n = 1), and social network (n = 2). Three studies mentioned the Social Frailty Questionnaire (SFQ) by Teo et al. (2017) that measured five social domains: social network (n = 2), social support (n = 2), social activity (n = 1), socioeconomic (n = 2), and financial difficulty (n = 1). Three studies also used the Social Frailty Questionnaire (SFQ) by Yamada & Arai (2018), a four-question screening tool that consisted of three domains: social activity (n = 1), social network (n = 2), and financial difficulty (n = 1).

The following assessment tools were mentioned in one study each. The Social Frailty Questionnaire (SFQ) by Lian et al. (2020) assessed two domains (daily social activities and social network). Pek et al. (2020) also designed a nine-item Social Frailty Questionnaire (SFQ) that determined five domains of social frailty, including support (n = 3), social network (n = 2), social activities (n = 2), social role (n = 1), and financial difficulty (n = 1). Armstrong et al. (2015) designed the Social Vulnerability Index (SVI) that measures five social domains, including social network (n = 7), social support (n = 6), social activity (n = 2), social role (n = 1), and financial difficulty (n = 1). The Chinese version of the Lubben Social Network Scale (LSNS-6) by Lubben et al. (2006) was a six-item self-reported scale translated by Kuo et al. (2019). It comprised two domains, namely social network (n = 2) and social support (n = 4). The Social Frailty Screening Questionnaire (HALFT) tool was designed by Ma, Sun & Tang (2018) and consisted of four social domains: social networks (n = 2), social role (n = 1), social activity (n = 1), and financial difficulty (n = 1). The Social Deficits/Social Frailty Questionnaire (SD-SF) was a ten-item scale designed by Lee et al. (2020) to measure five domains, including social networks (n = 4), social support (n = 3), socioeconomic (n = 1), financial difficulty (n = 1), and social activity (n = 1).

Kwan, Lau & Cheung (2015) developed the Comprehensive Model of Frailty (CMF), which consisted of a 44-item frailty scale measuring frailty syndrome among Chinese older adults. The social frailty domain included three social items: social networks (n = 1), social activities (n = 1), and social support (n = 1). The Comprehensive Geriatric Assessment-Frailty Index (CGA-FI) was a 14-item scale measuring five domains of frailty developed by Ma et al. (2017); its social domain measured a dichotomous item related to social support (n = 1). Qiao et al. (2018) developed the Chinese-Comprehensive Frailty Assessment Instrument (CFAI), which is a 21-item self-report frailty tool. Kim et al. (2021) developed the Frailty Assessment Scale (FAS); this 14-item self-report frailty assessment tool assessed three domains of frailty, the social domain measured three items regarding social activities (n = 1), social support (n = 1), and social networks (n = 1). Meanwhile, Abe et al. (2019) designed the Kaigo-Yobo Checklist (KYCL), a 15-item dichotomous scale. Its social domain provided five questions measuring social activities (n = 3) and social networks (n = 2). The total social domain score ranged from 0 to 5, with a higher score indicating more social activity. The Kihon Checklist (KCL) was another dichotomous tool developed by the Japanese Ministry of Health, Labour, and Welfare reported by Watanabe et al. (2020). The social domain provided four items only assessing social activities (n = 3) and social support (n = 1). The Tilburg Frailty Indicator (TFI) was a 25-item self-report questionnaire reported by Song et al. (2020). Its social frailty domain consisted of a dichotomous scale for social networks (n = 2) and social support (n = 1).

Social domains of frailty

Regarding the individual domains of social frailty, the current review yielded six potential domains (codes) for social frailty within the four categories, see Table 1. The social network was the most utilized domain in the reviewed scales (n = 13 screening tools; n = 29 study), followed by social activities (n = 12 screening tools; n = 28 study), social support (n = 11 screening tools; n = 13 analysis), financial difficulty (n = 6 screening tools; n = 10 studies), social role (n = 4 screening tools; n = 16 studies), and sociodemographic (n = 2 screening tools; n = 4 studies), such as education status and housing type. Therefore, social resources and social needs were the most frequent category in the reviewed tools (n = 13 tools each), followed by social behaviors (n = 12 tools) and general resources (n = 6 tools).

The social network domain refers to the availability of physical, emotional, or virtual (e.g., online) frequent contact with someone. It consisted of a wide range of questions about frequent contact with others, feeling close, or living with someone. Social activities refer to the regular (group) activities of older adults, and they include items related to the frequency of going out, doing a workout, or having a hobby (Ma, Sun & Tang, 2018; Watanabe et al., 2020).

Furthermore, social support represents confiding or turning to others for advice, help, care, and support (Armstrong et al., 2015; Pek et al., 2020). The social role included items related to feeling helpful to others (Makizako et al., 2018b; Tsutsumimoto et al., 2018). As for financial difficulty, the questions were about receiving enough income or feeling satisfied with the economic status (Lee et al., 2020; Pek et al., 2020), see Table 1. Four studies only showed that females have significantly higher levels of social frailty related to social activity, social support, and social network domains than males (Abe et al., 2019; Kuo et al., 2019; Lee et al., 2020; Song et al., 2020). However, the other studies didn’t provide significant results regarding social frailty domains based on gender differences.

Adverse outcomes related to social frailty and its domains

This section investigates the adverse health outcomes related to the severity of social frailty domains in community-dwelling Asia-Pacific older adults. The study revealed eleven adverse health outcomes (n = 11 codes) associated with the social domains of frailty: mortality (n = 6 domains), physical frailty or mobility issues (n = 6 domains), disability incidence (n = 6 domains), cognitive frailty (n = 5 domains), mental illness or depression (n = 5 domains), nutritional disorders (n = 5 domains), dementia (n = 4 domains), oral frailty (n = 3 domains), hearing loss (n = 3 domains), BMI or obesity issues (n = 3 domains), and chronic pain (n = 3 domains). Thus, all six mentioned domains of social frailty predicted mortality, physical frailty, and disability incidence, see Table 2. Therefore, social resources, social needs, and social behavior predicted all eleven health outcomes, while general resources predicted only seven health outcomes, as shown in Table 2.

Social network and social activities domains predicted all eleven adverse health outcomes, see Table 2. Social role frailty predicted six health outcomes: physical, cognitive, and oral frailty, mental illness (depression), nutritional issues, dementia, disability incidence, chronic pain, and mortality. Social support frailty was associated with seven health outcomes, including physical frailty, mental frailty, disability incidence, nutritional disorders, hearing loss, BMI or obesity, and predicted mortality (Armstrong et al., 2015; Watanabe et al., 2020). The financial difficulty domain also predicted seven health outcomes: physical and cognitive frailty, mental illness or depression, malnutrition, dementia, disability incidence, and mortality (Pek et al., 2020). Moreover, the socioeconomic domains were associated with four health outcomes: physical frailty, cognitive frailty, disability incidence, and mortality (Teo et al., 2017); see Table 2. However, skeletal muscle mass had no significant association with these social domains (Makizako et al., 2018a; Kim et al., 2021; Lian et al., 2020). AD risk and EDS were confounding factors significantly associated with some social domains and non-significant associations with others (Pek et al., 2020; Yamada & Arai, 2018). Besides, each assessment tool predicted different health outcomes, see Document S2.

Two main methods were used to measure the reported health outcomes: clinical observation (n = 8 health outcomes) and screening tools (n = 4 health outcomes). Observation using medical diagnosis measured six health outcomes, including physical frailty, oral frailty, dementia, disability incidence, chronic pain, and hearing loss (Abe et al., 2019; Armstrong et al., 2015; Hirase et al., 2019; Ma, Sun & Tang, 2018; Makizako et al., 2015; Tsutsumimoto et al., 2018; Nakakubo et al., 2019; Song et al., 2020; Watanabe et al., 2020). Mortality was measured by observing the number of deaths, and obesity was measured using the observation of BMI (weight by the square of height). Screening tools using a questionnaire checklist measured four health outcomes: physical frailty (e.g., using the frailty index), cognitive frailty (using the National Center for Geriatrics and Gerontology functional assessment tool (NCGG-FAT) or Mini-Mental State Examination (MMSE)), depression (using Geriatric Depression Scale (GDS) or Epidemiological Studies Depression (CESD)), and nutritional disorders (using Mini Nutritional Assessment (MNA) or Nutrition Screening Initiative (NSI)) (Park et al., 2019; Pek et al., 2020; Teo et al., 2019), see Table 3.

Theoretical and practical implication

The current systematic review, based on the existing evidence and theories, such as the Social Production Function theory and social frailty concept by Bunt et al. (2017), fills the gap in the literature by considering a wide range of social domains in the social frailty model and determining the related adverse health outcomes. The current systematic review contributes to the conceptual definition of social frailty and the interpretation and convergence of the concept of multidimensional social frailty. It also provides a better understanding of the role of each social domain in the conception of social frailty. Although the social variables of frailty, such as social activities and social resources, have been reviewed in the older people’s frailty models (Bessa, Ribeiro & Coelho, 2018; Bunt et al., 2017), its related health outcomes are still unexplored, especially in the context of Asia-Pacific settings. Thus, the current review contributed to the concept of social frailty, its domains, and its health outcomes in Asia-Pacific region. Therefore, the significance of this review lies in linking social health science to public health and health sciences.

The current study also proved that each domain of the six reviewed domains of social frailty predicts adverse health outcomes, including mortality. It enhanced our understanding of the expected adverse health outcomes resulting from the frailty of each social domain. This facilitates professionals and researchers in making informed decisions regarding selecting instruments tailored to each study’s specific circumstances and scope. Understanding the social domains related to certain social frailty themes contributes to guiding healthcare professionals to develop appropriate interventions to mitigate or postpone the onset of specific older people’s health outcomes and frailty. For example, proper intervention, such as enhancing social support combined with social activity and social role, can contribute to reducing nutrition disorders, depression, and disability, therefore decreasing frailty in older people. Social-based interventions, such as participation in social activities, may be beneficial for reducing social frailty among older people, especially in the Asia Pacific region. Therefore, professionals and researchers in health and social sciences are encouraged to consider the six mentioned domains to understand better the frailty syndrome and related health consequences. Furthermore, the current review provided a theoretical background for professionals and researchers to understand social frailty syndrome in older people.

Limitations and future research directions

The limitations found in this review were the ambiguity between the adverse outcomes and the social domains, where some studies using the same frailty scale found a significant result about some of the possible outcomes, while others did not. One additional limitation identified in the study pertained to the potential confusion arising from the inconsistent nomenclature employed for specific scales across various research investigations. Moreover, many of the reviewed screening scales from the Asia-Pacific setting lacked or did not disclose clear validity and reliability checks. However, the authors used methodological rigor and only included quality studies by applying the quality assessment framework of Wells et al. (2000). Finally, all the studies reviewed were written in English and published within the last 20 years. Therefore, authors excluded from the review investigations that may have been reported in other (Asian) languages and published before 2002 were excluded.

The current study is limited to social frailty assessed through questionnaires; this aligns with the study’s objective to characterize the social frailty domains and their health outcomes by overviewing the frailty screening tools. However, experimental approaches, such as clinical practice and physical examinations, can contribute to a more objective frailty assessment. Therefore, future research is encouraged to conduct more objective methods for evaluating older people’s frailty.

Based on these limitations, future research is recommended to develop a comprehensive empirical study to test the conceptual model of social frailty and its domains. A new comprehensive social assessment tool is required to contribute to the operationalization framework and a unified conceptual definition of social frailty. Future interventions that target individuals or groups of socially frail older adults should also address the several domains of social frailty mentioned in this study. Besides, future studies must apply a certain level of reliability and validity to the assessment tools.