Health-related quality of life and work ability among paid and family caregivers: A cross-sectional study in an industrially developing country

Health-related quality of life and work ability among paid and family caregivers: A cross-sectional study in an industrially developing country Jo~ ao Marcos Bernardes, PhD a,b , Laura Ara ujo, MSc a , Rodrigo Chavari de Arruda, MSc a , Adriano Paulo Aparecido Pereira de Oliveira, MSc c , Melissa Spr€ oesser Alonso, MSc a , Carlos Ruiz-Frutos, PhD d,e , Juan Carlos Camacho-Vega, PhD f , Luis El Khoury-Moreno, PhD g , Julio Torrej on-Martínez, PhD g , Juan G omez-Salgado, PhD d,e, *, Adriano Dias, PhD a,b a Graduate Program in Collective/Public Health, S~ ao Paulo State University (UNESP), Medical School, Botucatu, S~ ao Paulo, Brazil b Department of Public Health, S~ ao Paulo State University (UNESP), Medical School, Botucatu, S~ ao Paulo, Brazil. c Undergraduate Medicine Program, S~ ao Paulo State University (UNESP), Medical School, Botucatu, S~ ao Paulo, Brazil d Department of Sociology, Social Work and Public Health, Faculty of Labour Sciences, University of Huelva, Huelva, Spain e Safety and Health Postgraduate Programme, Universidad Espíritu Santo, Guayaquil, Ecuador. f Department of Building Construction II, Higher Technical School of Building Engineering, University of Seville, Seville, Spain g Department of Stomatology, Faculty of Odontology, University of Seville, Seville, Spain ARTICLE INFO Article history: Received 8 September 2025 Received in revised form 30 September 2025 Accepted 20 October 2025 Available online xxx ABSTRACT Caregivers are essential for providing daily care to individuals with functional disabilities, but caregiving can negatively impact physical and mental health. This study assessed the health-related quality of life and work ability of 97 paid caregivers and 91 family caregivers, identifying factors associated with these outcomes. Mann-Whitney U and chi-square tests were used to analyze differences between groups, along with logistic regression models to explore the relationship between caregiver burden, social support, and the outcomes. Results showed family caregivers experienced higher burden, lower social support, worse quality of life, and reduced work ability compared to paid caregivers. Longer caregiving hours were linked to poorer outcomes, while good physical fitness was a protective factor. Moderate to severe caregiver burden strongly correlated with poor outcomes, while social support had a protective effect. The findings highlight the importance of interventions to reduce caregiver burden, enhance social support, and promote physical fitness for caregivers. © 2025 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/) Keywords: Caregivers Home Health Aides Health-Related Quality of Life Work Capacity Evaluation Cross-Sectional Study Occupational Health Key messages Family caregivers have worse health-related quality of life than paid caregivers. Family caregivers also have worse work ability than paid caregivers. Factors associated with health-related quality of life and work ability are similar. Policy implications suggest mixing individual and social interventions for caregivers. Introduction The United Nations estimates that the global life expectancy will reach 77 years by 2048. 1 Latin America and the Caribbean are also experiencing significant growth in their older populations, with approximately 190 million older adults, equaling the youth population (176 million). 2 In Brazil, 2020 statistics show that people aged 60 years and older make up 14% of the population, and this number is expected to match the youth population in a decade. 3 Furthermore, medical advancements have increased survival rates for adults and children with complex health conditions, leading to a rise in people with functional disabilities, requiring care. 4,5 Caregivers play a crucial role in assisting individuals in dependency or semi-dependency, promoting their autonomy and wellbeing. 6 Caregivers can be classified as paid caregivers (also known as formal caregivers, home health aides, home care workers, or personal care attendants) or family caregivers (also known as informal *Corresponding author at: Department of Sociology, Social Work and Public Health. Faculty of Labour Sciences, University of Huelva. Avenida Tres de marzo, s/n, 21007, Huelva, Spain. E-mail address: [email protected] (J. G omez-Salgado). https://doi.org/10.1016/j.gerinurse.2025.103696 0197-4572/$ see front matter © 2025 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/) Geriatric Nursing 66 (2025) 103696 Contents lists available at ScienceDirect Geriatric Nursing journal homepage: www.gnjournal.com caregivers, unpaid caregivers or lay caregivers) based on their relationship with the care recipient. 7-9 It is worth noting, however, that there is currently no updated and widely accepted definition of what constitutes a family caregiver. 9-11 In general, family caregivers provide care without financial compensation, often being family members, friends, or neighbors, and this support may occur either in close proximity to or at a distance from the care recipient. 7,8,12,13 On the other hand, a paid caregiver is someone whose professional occupation is to provide caregiving activities for which they are hired and paid to perform. 8,14-16 It is important to highlight that, unlike some other countries, 17 Brazil does not have policies to financially support family caregivers. Caregiving can result in positive outcomes, such as a sense of purpose, 12,18 but it can also have negative impacts on caregivers’ mental and physical health, leading to burnout, exhaustion, and chronic physical complaints. 12,19-21 Beyond these adverse effects, caregiving can lead to financial toxicity, 22-24 which can subsequently exacerbate negative caregiver health outcomes. 25-27 Given these potential negative consequences, caregiving may affect healthrelated quality of life (HRQoL) and work ability (WA). HRQoL can be defined as the extent to which an individual operates in their life and their perceived wellness across physical, mental, and social health aspects, 28 while WA refers to an individual’s capacity to perform work based on its demands and their health and resources. 29 Both are influenced by numerous personal, job-related, and lifestyle factors. 30-37 To address these challenges effectively, it is essential to understand how these impacts differ between caregiver groups, since caregiving consequences may be influenced by the prior relationship between caregiver and care recipient, and the type of interaction developed during the caregiving process. 19,21 However, few studies have investigated the differential impact on health outcomes between family and paid caregivers comprehensively. A literature search using Google, Google Scholar, and PubMed with the keywords “health-related quality of life,”“work ability,”“caregiver burden,” “paid caregivers,”“formal caregivers,”“home health aides,”“family caregivers,”“informal caregivers,”and “unpaid caregivers”identified only four studies that compared family and paid caregivers 38-41 in terms of mental health, 38,39,41 physical health, 41 quality of life 39,41 and caregiver burden. 38-41 The results of these studies consistently indicate that family caregivers experience worse outcomes across all measured variables. 38-41 Nevertheless, none of these studies specifically evaluated HRQoL or WA. HRQoL has been suggested to be a predictor of various conditions, such as all-cause mortality, 42,43 cardiovascular events, 44,45 dementia, 46 asthma, 47 functional impairment, 48 and disability retirement due to musculoskeletal disorders. 49 Additionally, WA has already been associated with job stress, sickness absence, intention to leave work, and early retirement. 50-53 Thus, promoting caregivers’HRQoL and WA is crucial to addressing challenges related to the increasing population with functional disabilities that require care. Considering the above, this study aimed to evaluate the HRQoL and WA of paid and family caregivers and investigate factors associated with both outcomes. By addressing these understudied outcomes and comparing caregiver groups, this study offers a novel contribution to public health and occupational health. Understanding the differences between paid and family caregivers is critical, as these groups face distinct challenges due to their caregiving roles. Family caregivers often lack formal training, social support and resources, while paid caregivers encounter unique occupational demands and stressors. These differences may have implications for targeted policies and interventions, both to alleviate the growing burden on family caregivers and to improve the working conditions of paid caregivers, ensuring the sustainability and effectiveness of caregiving systems. By clarifying these distinctions, this study provides essential information to guide strategies that promote caregivers’well-being and address the challenges posed by an aging population. Material and methods Design This is a cross-sectional study involving paid and family caregivers working in Itapetininga, a medium-sized city in the southwest region of the State of S~ ao Paulo, Brazil. Itapetininga has a population of approximately 160,000 inhabitants, with 11.5% aged 65 years or older. The city has 4.78 physicians and 2.62 nurses per 1,000 inhabitants. Participants Sample size calculation was conducted using the following parameters: a significance level (alpha) of 5%, a test power (1-beta) of 80%, a mean Work Ability Index (WAI) of 39.7 (SD = 6.6) for paid caregivers, 54 and a mean WAI of 37.2 for family caregivers. 55 Based on these assumptions, the estimated minimum sample size was 110 paid caregivers and 110 family caregivers. In this study, paid caregivers were defined as individuals whose primary activity is caregiving, who receive payment for their work. In contrast, family caregivers were defined as those providing care without financial compensation, typically being family members, friends, or neighbors. The inclusion criteria for paid caregivers were being 18 years or older and having caregiving as their primary work activity. For family caregivers, the inclusion criteria were being 18 years or older, not receiving payment for caregiving, and identifying as the primary caregiver of the care recipient. The exclusion criteria for both groups were the inability to complete the questionnaire, provision of support and assistance as a distance caregiver and not currently providing direct care to an individual with functional limitations. Additionally, family caregivers who were healthcare professionals providing care in a professional capacity were also excluded. Participants were recruited, and data were collected between January and August 2023, resulting in a final sample of 188 caregivers (97 paid and 91 family). Family caregivers were recruited through the Brazilian public health system’s home care service (Servi¸co de Aten¸c~ ao Domiciliar - SAD) and an outpatient physical therapy university clinic. Paid caregivers were recruited from four long-term care facilities and two vocational nursing schools offering practical nursing training (equivalent to Brazilian t ecnico de enfermagem programs). All paid caregivers employed at the facilities, as well as those enrolled in the vocational nursing schools, were invited to participate. Additionally, both paid and family caregivers were recruited through referrals (snowball sampling), messaging app groups, social media, and the offer of a free one-day course on domiciliary oxygen use for caregivers. This multi-faceted recruitment approach has been employed in previous studies with caregivers. 56 While this method can expand a study’s reach, enhance consistency, and reduce bias, it also makes it difficult to determine the exact number of caregivers invited. Consequently, estimating the response rate was not possible. During recruitment, each potential participant was approached individually and provided with a detailed explanation of the study. Those who expressed interest in participating were screened for eligibility based on the inclusion and exclusion criteria. Eligible participants were then asked to provide written informed consent and complete the study questionnaire. 2J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 Data collection The questionnaire collected the following caregiver-related variables: sociodemographic characteristics, self-perceived physical fitness, caregiving activities, caregiver burden, perceived social support, HRQoL, and WA. Additionally, the functional independence level of the care recipients was assessed. To assess caregivers’sociodemographic characteristics, a questionnaire was developed containing the following self-reported variables: sex (female or male), age (in complete years), and skin color. Ethnoracial categorization followed Brazilian census classification, with participants self-identifying according to skin-color-based classifications: White, Black, Mixed, East Asian, or Indigenous. Due to shared socioeconomic marginalization and statistical limitations (low subgroup frequencies and uneven distribution), these categories were aggregated into “white”and “non-white”for regression analyses. This aggregation aligns with Brazil’saffirmative action frameworks, particularly the PPI (“Pretos, Pardos, e Indígenas”- Black, Mixed, and Indigenous) grouping, which is used in public policies to identify and support these groups. Other variables included self-reported body mass index (kg/m 2 ), marital status (single, married or living with a partner, or other), number of children (continuous variable), and household size (number of people in the household). Employment status was classified as unemployed or employed. Personal income was categorized as up to one minimum wage, up to two minimum wages, three or more minimum wages, or “don’t know/prefer not to disclose”(during the data collection period, the Brazilian minimum wage was approximately USD 250 per month). Family income followed the same categories as personal income. Education level was recorded as middle school, high school, or college. Additional variables included religiousness (yes or no), smoking (yes or no), alcohol consumption (yes or no), and medication use (yes or no). Self-perceived physical fitness was assessed with the question: “How do you rate your physical fitness?”Responses were recorded on a six-point ordinal scale, ranging from “precarious” 0 to “excellent”. 5 Following a previous study that used this same assessment method, physical fitness was categorized as “good” for scores greater than three and “precarious”for scores equal to or less than three. 57 Regarding caregiving activities, participants were asked about the length of their caregiving experience (in complete years), the number of care recipients (continuous variable), and the number of daily caregiving hours (continuous variable). The questionnaire also included items to identify which daily caregiving activities were performed. The following activities were listed: bathing, feeding, manual patient handling, dressing/undressing, walking assistance, and toileting. Participants responded to each activity on a five-point Likert scale ranging from “never”to “always.”For statistical analysis, these responses were dichotomized into “yes”or “no”. The response “never”was coded as “no,”while all other responses (“rarely,”“sometimes,” “often,”and “always”) were coded as “yes”. Additionally, caregivers were asked if they performed household activities such as cooking, laundry, and cleaning. Based on these responses, the variable “household tasks”was created and categorized into four levels: none, one, two, or three tasks. In this study, caregiver burden is defined as the perceived adverse effects of caregiving on caregivers’mental and physical health, social relations, financial situation, and work status. 58 Caregiver burden was evaluated using the seven-item version of the Zarit Burden Scale (ZBI-7), 59 which is considered to be efficient for large-scale studies. 60 The ZBI-7 has demonstrated high correlation with the full version of the scale across various caregiving populations, including family caregivers for patients with advanced cancer, dementia, and acquired brain injury. It also shows strong internal consistency, with Cronbach’s alpha ranging from 0.82 to 0.90. 61 In the present study, Cronbach’s alpha for the ZBI-7 was 0.92. The ZBI-7 comprises seven items rated on a five-point Likert scale, with responses ranging from “never”to “almost always”. The total score is obtained by summing the item scores, yielding a range from 7 to 35 points. 59 According to the Brazilian Ministry of Health guidelines, caregiver burden scores were categorized as mild (up to 14 points), moderate (15 to 21 points), and severe (22 points or more). 62 Perceived social support was assessed using the Brazilian version of the MOS Social Support Survey (MOS-SSS). 63 This instrument demonstrated high internal consistency, with Cronbach’s alpha coefficients ranging from 0.83 to 0.92 for all dimensions of the Brazilian version. 64 In the present study, Cronbach’s alpha for the MOS-SSS was 0.97. The MOS-SSS consists of 19 questions rated on a five-point Likert scale, with responses ranging from “never”to “always.”These questions assess four dimensions of social support: material support, emotional support, positive social interaction, and emotional support. The overall scale score was calculated by averaging the scores of all items, with higher scores indicating a better perception of social support. HRQoL was assessed using the Brazilian version of the Nottingham Health Profile (NHP). 65 This instrument demonstrated a Cronbach’s alpha of 0.82, with intra-class correlation coefficients for testretest and inter-rater reliability of 0.96 and 0.92, respectively. 66 In the present study, the NHP had a Cronbach’s alpha of 0.90. The NHP consists of 38 items with dichotomous “yes”or “no” responses. 65 These items are grouped into six domains: energy level (3 items), pain (8 items), emotional reactions (9 items), sleep (5 items), social interaction (5 items), and physical abilities (8 items). The total NHP score was calculated by averaging the item scores, with each item weighted according to the original guidelines. Higher scores indicate a perceived worse state of health. As the NHP lacks an established cutoff, the score was dichotomized at the median (5.493) for analytical purposes, a methodologically robust approach that helps mitigate imbalance in subsequent statistical comparisons. Thus, scores above this value were interpreted as indicating poor HRQoL. WA was evaluated using the Brazilian version of the Work Ability Index (WAI). 67 The Brazilian version of the WAI demonstrated a Cronbach’s alpha of 0.72, 68 while in the present study, Cronbach’s alpha was 0.85. The WAI consists of 10 items distributed across seven dimensions. Each item is scored individually, resulting in a total score ranging from 7 to 49 points. Higher scores indicate a better self-perception of WA. In this study, WA scores below 37 points were considered inadequate. 55,69,70 Finally, the functional independence of care recipients was assessed using the Brazilian version of the Functional Independence Measure (FIM). 71 Although the FIM was translated into Brazilian Portuguese in 2004, no psychometric data on its use have been published to date. In this study, Cronbach’s alpha for the FIM was 0.97. The FIM consists of 18 items divided into two major domains: motor (selfcare, sphincter control, mobility, and locomotion) and cognitive (communication and social cognition). Each item is scored on a scale from 1 (total dependence) to 7 (complete independence). The final score is the sum of all item scores, ranging from 18 to 126. Data analysis Data analysis was performed using IBM SPSS Statistics software v.26.0 (IBM, Armonk, NY, USA). Descriptive analyses included measures of central tendency and dispersion for continuous variables, and simple frequencies for categorical variables. The Kolmogorov-Smirnov test was applied to check the normality assumption for continuous variables. Differences between paid and family caregivers were J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 3 then evaluated. The chi-square test was used for categorical variables, while the Mann-Whitney U test was applied to continuous variables, as none exhibited a normal distribution. A significance level of 5% (p <0.05) was adopted for all statistical tests. To assess factors associated with caregivers’HRQoL and WA, two distinct logistic regression models were employed for each outcome. The first model was hierarchical and evaluated the association between the outcomes and independent variables, grouped according to a two-level hierarchical structure (Fig. 1). The second model examined the relationship between caregiver burden, social support, and the outcomes. The decision to use two distinct regression models was based on the adoption of a hierarchical conceptual framework for the analysis. In this framework, distal-level variables were related to the caregiver, while proximal-level variables pertained to care. Since both the ZBI-7 and the MOS-SSS included variables from both levels, they were excluded from the hierarchical model to avoid potential confounding effects due to multicollinearity, as done in a previous study. 45 The hierarchical models were preceded by bivariate analysis, in which the association of independent variables with each outcome was examined using simple logistic regressions. Subsequently, the independent variables were grouped according to the two-level hierarchical model, and the analysis was conducted sequentially, starting with the distal level (caregiver-related factors) and moving to the proximal level (care-related factors). Distal-level variables were adjusted for each other through logistic regression using a stepwise forward procedure. Variables with p 0.05 were retained in the model. Next, proximal-level variables were adjusted for each other, along with the distal-level variables that remained in the model, using the same procedure and p-value cut-off. Finally, variables that achieved a p 0.05 in the previous step were included in the final multiple hierarchical models. In these final models, variables with a two-tailed p-value of less than 5% were considered significantly associated with the outcomes. The models assessing the association between caregiver burden and social support with HRQoL and WA were initially adjusted using simple logistic regression. Variables with p 0.25 in these models were included in the multiple logistic regression analysis. In the multiple models, variables with a two-tailed p-value below 5% were considered significantly associated with the outcomes Ethical approval The study adhered to Resolution 466/12 concerning ethical guidelines for research involving human subjects. It was conducted with approval from the Research Ethics Committee of the Botucatu Medical School, UNESP (Approval number 5.651.612, CAAE 61013422.0.0000.5411). Participants were informed about the research and its objectives. Those who agreed to participate provided written informed consent by signing the Informed Consent Form. Results A total of 188 caregivers participated in the study, comprising 97 paid caregivers and 91 family caregivers. As shown in Table 1, several sociodemographic characteristics differed significantly between these two groups. Family caregivers reported higher levels of caregiver burden, lower perceived social support, and poorer HRQoL and WA compared to paid caregivers. Notably, despite caring for recipients with higher functional capacity, family caregivers reported dedicating a median of 24 hours per day to caregiving activities. Accordingly, 67.8% of family caregivers were unemployed. Regarding WA, just over a quarter (25.8%) of paid caregivers exhibited inadequate WA, which includes the poor and moderate categories of the WAI. In contrast, this percentage was significantly higher among family caregivers, reaching 51.6% (p <0.001). Notably, only 1.1% of all caregivers were classified as presenting excellent WA; this group comprised two paid caregivers, while no family caregivers achieved this classification. Table 2 presents the results of the simple logistic regression models, grouped according to the two hierarchical levels, as well as the variables associated with inadequate HRQoL in the final multiple hierarchical logistic regression model. Medication use (OR 0.499, 95% CI 0.229-0.880) and good self-perceived physical fitness (OR 0.275, 95% CI 0.140-0.543) were inversely associated with inadequate HRQoL. In contrast, the number of daily caregiving hours (OR 1.057, 95% CI 1.006-1.111) was positively associated with inadequate HRQoL. Table 3 presents the odds ratio estimates and respective 95% confidence intervals for caregiver burden and perceived social support in the simple and multiple logistic regression models. Caregiver burden exhibited a strong dose-response positive association with Fig. 1. Theoretical hierarchical model of the studied variables, according to established levels. 4J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 Table 1 Study participants’characteristics stratified into paid and family caregivers (Brazil, 2024). Variable (n) Total Paid caregivers Family caregivers p-value Median (IQR) or n (%) Median (IQR) or n (%) Median (IQR) or n (%) Sex (188) Female 146 (77.7) 74 (76.3) a 72 (79.1) a 0.641 Male 42 (22.3) 23 (23.7) a 19 (20.9) a Age (174) 37.5 (21.0) 31.0 (16.0) a 43.0 (23.0) b <0.001 Skin color (188) Non-white 66 (35.1) 38 (39.2) a 28 (30.8) a 0.228 White 122 (64.9) 59 (60.8) a 63 (69.2) a Body mass index (179) 27.0 (5.1) 26.8 (5.2) a 27.1 (4.7) a 0.472 Marital status (188) Single 55 (29.3) 37 (38.1) a 18 (19.8) b Married or living with a partner 98 (52.1) 41 (42.3) a 57 (62.6) b 0.010 Others 35 (18.6) 19 (19.6) a 16 (17.6) a Number of children (188) 1.0 (2.0) 0.0 (2.0) a 2.0 (2.0) b <0.001 Household size (186) 2.0 (2.0) 2.0 (2.0) a 2.0 (2.0) a 0.055 Personal income (183) Up to one minimum wage 58 (31.7) 27 (27.8) a 31 (36) a <0.001 Up to two minimum wages 71 (38.8) 55 (56.7) a 16(18.6) b Three or more minimum wages 43 (23.5) 13 (13.4) a 30 (34.9) b Don’t know/prefer not to disclose 11 (6.0) 2 (2.1) a 9 (10.5) b Family income (183) Up to one minimum wage 16 (8.7) 5 (5.3) a 11 (12.5) a 0.365 Up to two minimum wages 37 (20.2) 21 (22.1) a 16 (18.2) a Three or more minimum wages 93 (50.8) 49 (51.6) a 44 (50.0) a Don’t know/prefer not to disclose 37 (20.2) 20 (21.1) a 17 (19.3) a Education level (188) Middle school 22 (11.7) 7 (7.2) a 15 (16.5) b High school 105 (55.9) 66 (68.0) a 39 (42.9) b 0.002 College 61 (32.4) 24 (24.7) a 37 (40.7) b Religiousness (188) Yes 153 (81.4) 79 (81.4) a 74 (81.3) a 0.982 No 35 (18.6) 18 (18.6) a 17 (18.7) a Smoking (188) Yes 33 (17.6) 19 (19.6) a 14 (15.4) a 0.449 No 155 (71.8) 78 (80.4) a 77 (84.6) a Alcohol consumption (188) Yes 80 (42.6) 44 (45,4) a 36 (39.6) a 0.421 No 108 (57.4) 53 (54.6) a 55 (60.4) a Medication use (188) Yes 66 (35.1) 29 (29.9) a 37 (40.7) a 0.122 No 122 (64.9) 68 (70.1) a 54 (59.3) a Self-perceived physical fitness (188) Precarious 66 (35.1) 25 (25.8) a 41 (45.1) b 0.006 Good 122 (64.9) 72 (74.2) a 50 (54.9) b Caregiving experience length (180) 4.0 (6.0) 4.0 (5.0) a 5.0 (6.0) a 0.350 Number of care recipients (186) 1.0 (8.0) 5.0 (19.0) a 1.0 (0.0) b <0.001 Daily caregiving hours (187) 12.0 (12.0) 12.0 (0.0) a 24.0 (14.0) b <0.001 Bathing (188) Yes 152 (80.9) 92 (94.8) a 60 (65.9) b <0.001 No 36 (19.1) 5 (5.2) a 31 (34.1) b Feeding (188) Yes 144 (76.6) 92 (94.8) a 52 (57.1) b <0.001 No 44 (23.4) 5 (5.2) a 39 (42.9) b Manual patient handling (188) Yes 152 (80.9) 92 (94.8) a 60 (65.9) b <0.001 No 36 (19.1) 5 (5.2) a 31 (34.1) b (Un)Dressing (188) Yes 155 (82.4) 95 (97.9) a 60 (65.9) b <0.001 No 33 (17.6) 2 (2.1) a 31 (34.1) b Walking assistance (188) Yes 158 (84.0) 91 (93.8) a 67 (73.6) b <0.001 No 30 (16.0) 6 (6.2) a 24 (26.4) b Toileting (188) Yes 150 (79.8) 93 (95.9) a 57 (62.6) b <0.001 No 38 (20.2) 4 (4.1) a 34 (37.4) b Household tasks (188) None 46 (24.5) 36 (37.1) a 10 (11.0) b <0.001 One 17 (9.0) 14 (14.4) a 3 (3.3) b Two 14 (7.4) 14 (14.4) a 0 (0.0) b Three 111 (59.1) 33 (34.1) a 78 (85.7) b (continued) J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 5 Table 1 (Continued) Variable (n) Total Paid caregivers Family caregivers p-value Median (IQR) or n (%) Median (IQR) or n (%) Median (IQR) or n (%) Care recipients’functional independence (187) 38.0 (54.0) 18.0 (31.0) a 64.5 (69.0) b <0.001 Caregiver burden (186) Mild 102 (54.8) 64 (66.0) a 38 (42.7) b 0.001 Moderate 43 (23.2) 22 (22.7) a 21 (23.6) b Severe 41 (22.0) 11 (11.3) a 30 (33.7) b Perceived social support (181) 4.4 (1.5) 4.8 (1.2) a 4.0 (2.0) b 0.002 Health-related quality of life (187) 5.4 (15.2) 3.9 (10.6) a 7.9 (18.0) b 0.014 Work ability (188) 38.0 (7.0) 40.0 (6.0) a 36.0 (8.0) b 0.001 Notes: 1) Categorical variables compared using the chi-square test. 2) Numerical variables compared using the Mann-Whitney U test. 3) Values followed by different letters in the subindex indicate statistically significant differences between groups (p <0.05). Table 2 Odds ratio (OR), 95% confidence interval (CI) and p-values obtained in the simple hierarchical logistic regression model and in the multiple hierarchical logistic regression model for HRQoL (Brazil, 2024). Variable OR CI p-value Simple hierarchical logistic regression model Distal level Caregiver type (ref.: paid) Family 1.956 1.093-3.501 0.024 Sex (ref.: male) Female 1.792 0.884-3.634 0.106 Age 1.030 1.008-1.053 0.008 Skin color (ref.: non-white) White 0.675 0.369-1.234 0.202 Body mass index 1.144 1.055-1.241 0.001 Marital status (ref.: single) Married or living with a partner 1.432 0.736-2.786 0.290 Others 1.534 0.654-3.597 0.325 Number of children 1.272 1.036-1.562 0.021 Household size 1.111 0.922-1.339 0.267 Personal income (ref.: Up to one minimum wage) Up to two minimum wages 0.637 0.317-1.281 0.206 Three or more minimum wages 1.054 0.475-2.340 0.897 Do not know/prefer not to disclose 1.045 0.287-3.812 0.947 Family income (ref.: Up to one minimum wage) Up to two minimum wages 0.352 0.096-1.294 0.116 Three or more minimum wages 0.319 0.096-1.063 0.063 Do not know/prefer not to disclose 0.203 0.055-0.754 0.017 Education level (ref.: middle school) High school 0.927 0.363-2.367 0.873 College 0.880 0.326-2.374 0.800 Religiousness (ref.: no) Yes 1.433 0.688-3.028 0.332 Smoking (ref.: yes) No 0.548 0.251-1.198 0.132 Alcohol consumption (ref.: yes) No 0.855 0.479-1.528 0.598 Medication use (ref.: yes) No 0.431 0.232-0.798 0.007 Self-perceived physical fitness (ref.: precarious) Good 0.230 0.120-0.442 <0.001 Caregiving experience length 1.041 1.000-1.083 0.050 Proximal level Number of care recipients 1.012 0.978-1.046 0.501 Daily caregiving hours 1.064 1.016-1.113 0.008 Bathing (ref.: no) Yes 0.944 0.453-1.970 0.879 Feeding (ref.: no) Yes 1.216 0.614-2.406 0.575 Manual patient handling (ref.: no) Yes 1.252 0.599-2.616 0.551 (Un)Dressing (ref.: no) Yes 1.178 0.550-2.525 0.674 Walking assistance (ref.: no) Yes 0.792 0.358-1.755 0.566 Toileting (ref.: no) Yes 1.422 0.689-2.936 0.341 Household tasks (ref.: none) One 1.778 0.571-5.538 0.321 Two 2.000 0.592-6.756 0.264 Three 2.723 1.319-5.624 0.007 Care recipients’functional independence 1.000 0.991-1.009 0.964 (continued) 6J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 inadequate HRQoL. In contrast, perceived social support (OR 0.666, 95% CI 0.462-0.960) was inversely associated with inadequate HRQoL. Table 4 presents the results of the simple logistic regression models, grouped into two hierarchical levels, as well as the variables associated with inadequate WA in the final multiple hierarchical logistic regression model. As shown, age (OR 1.057, 95% CI 1.028-1.087), body mass index (OR 1.108, 95% CI 1.001-1.226), and daily caregiving hours (OR 1.057, 95% CI 1.006-1.111) were positively associated with inadequate WA. In contrast, good self-perceived physical fitness (OR 0.257, 95% CI 0.116-0.570) was inversely associated with inadequate WA. Table 5 presents the odds ratio estimates and their respective 95% confidence intervals for caregiver burden and perceived social support in both the simple and multiple logistic regression models. Caregiver burden demonstrated a stronger dose-response positive association with inadequate WA than with HRQoL. In contrast, perceived social support (OR 0.565, 95% CI 0.384-0.832) was inversely associated with inadequate WA. Discussion In this study, we found significant differences between paid and family caregivers. Family caregivers, despite caring for recipients with higher functional capacity, reported greater caregiver burden, lower perceived social support, and poorer HRQoL and WA. Regression analyses also revealed notable similarities in the factors associated with HRQoL and WA among caregivers. Specifically, more daily caregiving hours were associated with worse HRQoL and WA, while better self-perceived physical fitness was associated with improved outcomes. Additionally, caregiver burden showed a strong doseresponse positive association with worse HRQoL and WA, whereas perceived social support was negatively associated with the outcomes. It is important to note that the two groups of caregivers differed in several sociodemographic and caregiving characteristics. Family caregivers were older, dedicated more hours to caregiving, and cared for recipients with higher functional capacity. These differences may have contributed to the higher caregiver burden, lower social support, and poorer health-related quality of life and work ability observed in family caregivers. Therefore, the caregiver type might not be the primary factor leading to the outcomes, but rather a combination of these contextual factors. This is supported by our regression analyses, which identified daily caregiving hours and selfperceived physical fitness as factors associated with both outcomes, regardless of caregiver type. Although few studies have compared paid and family caregivers, 38-41 our findings are supported by the literature. Previous studies suggest that family caregivers experience greater caregiver burden, lower quality of life, and more health problems. 38-41 Regarding WA, previous studies found that 28.8% of paid caregivers 54 and 35.7% of family caregivers 55 presented inadequate WA. These percentages, similar to those found in our study, highlight a worrying scenario. A possible explanation for these findings may relate to the different caregiving activities performed by paid and family caregivers. A previous study divided family caregivers’activities into three major groups: “helping around the house”(preparing meals, shopping, or cleaning), “looking after someone”(emotional support and supervision, as well as transportation and accompanying to appointments), and “nursing care services”(helping with functional activities of daily living like eating, dressing, bathing, or toileting). 72 While “nursing care services”led to significant decreases in self-rated health, “helping around the house”and “looking after someone”were associated with broader negative outcomes, including both poorer self-rated health and increased depressive symptoms. 72 Since paid caregivers usually do not engage, or do so less frequently, in “helping around the house”and “looking after someone”activities, 14 it would be expected that family caregivers would report worse HRQoL, lower WA, and higher caregiver burden. Another explanation for this finding may be the greater number of daily caregiving hours dedicated by family caregivers. Furthermore, over 30% of family caregivers combined caregiving with formal employment. This dual role likely reduced their rest time and made it difficult to access healthcare services, obtain medications or treatments, and recover from their own health issues. These factors, along with the caregiving tasks performed, may have contributed to their poorer outcomes. Regarding HRQoL, numerous studies indicate that assuming the family caregiver role might negatively affect health, with higher prevalence of depressive feelings and lower mental health scores. 73-76 The impact on physical health, however, is mixed. Some studies report negative effects, such as increased medication use and pain that affects daily activities, while others have found positive effects on physical health for specific subgroups of caregivers. 73,74 These differences in physical health outcomes may originate from the assessment methods used. When measured by self-assessed health, caregiving may show short-term positive effects. However, evaluations based on medication use or reported pain often reveal negative effects. 73 Notably, the positive self-assessment may reflect a bias, 77 as caregivers might compare their health to that of the care recipient, masking declines in their own physical health. 77 Although extensive research has explored the relationship between caregiving and health outcomes, few studies have Table 2 (Continued) Variable OR CI p-value Multiple hierarchical logistic regression model Final model Medication use (ref.: yes) No 0.449 0.229-0.880 0.020 Self-perceived physical fitness (ref.: precarious) Good 0.275 0.140-0.543 <0.001 Daily caregiving hours 1.057 1.006-1.111 0.028 Table 3 Odds ratio (OR), 95% confidence interval (CI) and p-values obtained in the simple logistic regression model and in the multiple logistic regression model for HRQoL (Brazil, 2024). Variable OR CI p-value Simple logistic regression model Caregiver burden (ref.: mild) Moderate 5.143 2.382-11.106 <0.001 Severe 14.483 5.505-38.105 <0.001 Perceived social support 0.619 0.450-0.852 0.003 Multiple logistic regression model Caregiver burden (ref.: mild) Moderate 5.908 2.613-13.362 <0.001 Severe 12.502 4.677-33.424 <0.001 Perceived social support 0.666 0.462-0.960 0.029 J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 7 Table 4 Odds ratio (OR), 95% confidence interval (CI) and p-values obtained in the simple hierarchical logistic regression model and in the multiple hierarchical logistic regression model for WA (Brazil, 2024). Variable OR CI p-value Simple hierarchical logistic regression model Distal level Caregiver type (ref.: paid) Family 3.076 1.666-5.680 <0.001 Sex (ref.: male) Female 1.744 0.827-3.679 0.144 Age 1.060 1.034-1.086 <0.001 Skin color (ref.: non-white) White 1.225 0.673-2.338 0.475 Body mass index 1.129 1.045-1.219 0.002 Marital status (ref.: single) Married or living with a partner 1.855 0.894-3.850 0.097 Others 3.905 1.582-9.638 0.003 Number of children 1.379 1.118-1.700 0.003 Household size 1.050 0.569-1.268 0.613 Personal income (ref.: Up to one minimum wage) Up to two minimum wages 1.135 0.540-2.384 0.739 Three or more minimum wages 2.807 1.237-6.371 0.014 Do not know/prefer not to disclose 2.667 0.719-9.890 0.142 Family income (ref.: Up to one minimum wage) Up to two minimum wages 2.955 0.717-12.182 0.134 Three or more minimum wages 2.615 0.696-9.824 0.155 Do not know/prefer not to disclose 2.955 0.717-12.182 0.134 Education level (ref.: middle school) High school 2.267 0.777-6.614 0.134 College 2.361 0.770-7.237 0.133 Religiousness (ref.: no) Yes 1.447 0.661-3.167 0.356 Smoking (ref.: yes) No 1.298 0.588-2.866 0.519 Alcohol consumption (ref.: yes) No 0.967 0.534-1.753 0.913 Medication use (ref.: yes) No 0.386 0.208-0.716 0.003 Self-perceived physical fitness (ref.: precarious) Good 0.257 0.136-0.483 <0.001 Caregiving experience length 1.061 1.018-1.105 0.005 Proximal level Number of care recipients 0.951 0.914-0.990 0.014 Daily caregiving hours 1.087 1.037-1.139 <0.001 Bathing (ref.: no) Yes Feeding (ref.: no) 0.840 0.401-1.760 0.644 Yes Manual patient handling (ref.: no) 0.867 0.435-1.726 0.684 Yes (Un)Dressing (ref.: no) 0.840 0.401-1.760 0.644 Yes Walking assistance (ref.: no) 0.946 0.438-2.042 0.887 Yes Toileting (ref.: no) 0.779 0.353-1.719 0.537 Yes Household tasks (ref.: none) 0.940 0.453-1.949 0.867 One Two 0.480 0.094-2.458 0.378 Three 2.000 0.546-7.327 0.295 Care recipients’functional independence 1.005 0.996-1.014 0.256 Multiple hierarchical logistic regression model Final model Age 1.057 1.028-1.087 <0.001 Body mass index 1.108 1.001-1.226 0.048 Self-perceived physical fitness (ref.: precarious) Good 0.257 0.116-0.570 0.001 Daily caregiving hours 1.079 1.018-1.143 0.011 8J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 investigated its impact on caregivers’occupational issues. 78,79 WA remains particularly underexplored among caregivers in Brazil and worldwide. This may be due to the lack of visibility and recognition of family caregivers by healthcare workers and healthcare systems, 80-83 and even by the caregivers themselves, who may view themselves simply as a spouse, child, or sibling providing practical care, rather than as workers. 84,85 Paid caregivers, on the other hand, often face a negative public image, frequently perceived as uneducated, poorly trained, and unskilled individuals performing unpleasant work. 86-88 Furthermore, they endure precarious work conditions, lack of unions and labor law protections, experience job insecurity, and receive low wages. 86-88 Therefore, the role of both paid and family caregivers as workers often goes unnoticed, which contributes to the scarcity of research on WA in this population despite its clear relevance, since for paid caregivers, caregiving is their remunerated work, and for many family caregivers, it is an unpaid role managed alongside paid employment. Regarding the factors associated with both outcomes, previous studies have shown that more caregiving hours are associated with poorer HRQoL. 89-94 For female nursing personnel, total work hours are also associated with inadequate WA. 95 This may be due to the detrimental effects of long workhours on health and subjective fatigue, 96 and reduced time for activities that promote mental and physical health. Therefore, reducing daily caregiving hours may be essential. Previous research suggests that using assistive devices (e.g., wheelchairs, walkers, canes, magnifiers, hearing aids, and amplifiers) can reduce caregiving hours by promoting care recipient’s independence. 97 These devices alleviate caregivers’physical demands by minimizing hands-on assistance with mobility, personal care, and household tasks. For example, wheelchairs and walkers facilitate safer, independent mobility, while magnifiers and hearing aids improve communication and engagement in daily activities. Adaptive tools for dressing, bathing, and feeding can further streamline caregiving, allowing caregivers more time for rest or other responsibilities. Additionally, adult day services, which offer out-of-home therapeutic activities, health monitoring, socialization, medical care, and transportation, have also been shown to decrease daily caregiving hours. 98 These services lower exposure to care-related stressors and may mitigate the negative effects of long-term caregiving. 99 Concerning physical fitness, previous studies have shown its inverse association with several chronic diseases. 100,101 Physical fitness may optimize neuroendocrine and physiological responses, improving adaptations to physical and psychosocial stressors. This helps protect against chronic stress and buffer against depression and anxiety. 100 Consequently, physical fitness has been associated with better HRQoL across various age groups, 102-106 which corroborates our findings. It is important to note, however, that while our study assessed self-perceived physical fitness, most prior research has relied on objectively assessed fitness. Evidence suggests a moderate to strong correlation between these measures, 107 which makes findings on self-perceived physical fitness a valuable, if indirect, reference for understanding the role of physical fitness in caregiver health. Nonetheless, these constructs are not identical and should be interpreted with caution. There is also evidence for an association between physical fitness and better WA. 108 Importantly, caregivers’ HRQoL and WA are crucial not only for themselves but also for their care recipients. Good physical fitness enables caregivers to deliver care vigorously and safely, minimizing pain and fatigue, thereby providing high-quality care. Consequently, physical activity interventions may help caregivers maintain or improve their HRQoL and WA, enabling them to perform their roles effectively. Previous studies have identified significant improvements in caregivers’cardiorespiratory fitness, 109,110 balance, 109 and BMI 110 in response to physical activity interventions. A systematic review also found a small yet significant improvement in both physical and mental health among family caregivers through exercise training. 111 Regarding WA, high-intensity physical activity during leisure time has been associated with better WA in a dose-response manner among workers in physically demanding jobs. 112 However, two systematic reviews found inconsistent effects of physical activity interventions on caregivers’physical fitness, 113,114 highlighting the need for further research. Moreover, time constraints, fatigue, lack of motivation, and insufficient support often hinder caregiver engagement in physical activities. 115 These barriers should be considered when planning such interventions. Regarding caregiver burden, multiple studies support our findings. 55,116-119 Caregiver burden is also associated with decreased work productivity, 120,121 highlighting the importance of addressing this issue. Improving health literacy may help, as higher health literacy has been associated with lower caregiver burden. 122 Psychological interventions, especially those involving both caregivers and care recipients, have also shown promise in reducing caregiver burden. 123 Additionally, multicomponent interventions that combine strategies such as education, skills training, counselling, support, and stress and mood management are likely to be effective in reducing caregiver burden. 124 It is important to consider the nature of caregiving tasks when designing these interventions, as emotional and personal care tasks tend to be more burdensome than instrumental tasks like household chores. 125,126 The inverse association between social support and poor HRQoL has been previously identified in caregivers. 127-129 Recent studies suggest that social support may act as a protective factor, mediating the relationship between caregiver burden and HRQoL, 130,131 and between psychological distress and quality of life. 132 Furthermore, a positive perception of the support received by caregivers has been shown to improve both mental and physical HRQoL of care recipients. 133-135 Regarding WA, although the impact of social support outside of work on WA has been scarcely investigated, some studies also found that it is inversely associated with WA, 136,137 corroborating our results. These results emphasize the need to expand the focus beyond caregivers and care recipients, highlighting the role of social support in policies and interventions aimed at improving caregiver’s HRQoL and WA. Previous studies have shown that social support provided by trained volunteers was ineffective in improving HRQoL. 138 In contrast, support groups led by experienced caregivers improved caregivers’HRQoL. 138 This suggests that social support from individuals in similar situations may be more effective than support from those without shared experiences. 138 Table 5 Odds ratio (OR), 95% confidence interval (CI) and p-values obtained in the simple logistic regression model and in the multiple logistic regression model for WA (Brazil, 2024). Variable OR CI p-value Simple logistic regression model Caregiver burden (ref.: mild) Moderate 6.670 2.961-15.024 <0.001 Severe 23.925 9.280-61.679 <0.001 Perceived social support 0.512 0.369-0.711 <0.001 Multiple logistic regression model Caregiver burden (ref.: mild) Moderate 6.358 2.677-15.096 <0.001 Severe 21.829 8.174-58.295 <0.001 Perceived social support 0.565 0.384-0.832 0.004 J.M. Bernardes et al. / Geriatric Nursing 66 (2025) 103696 9