Data came from the Lausanne cohort 65+ (Lc65+), a longitudinal population-based study of ageing in community-dwelling older adults living in the city of Lausanne, Switzerland. Details of the Lc65+ study are available elsewhere [15, 16]. Briefly, three random samples totaling 4,731 persons aged 65–70 were enrolled in 2004, 2009, and 2014. The following year, they were invited at the study center to complete a frailty phenotype baseline assessment. Follow-up includes annual questionnaires and triennial assessments of the frailty phenotype. The current analyses used 10-year data from the first two cohorts. Supplementary Figure S1 illustrates the selection procedure of study participants. The study protocol and informed consent were approved by the Ethics committee for human research of the canton Vaud (19/04).

Baseline characteristics of non-frail and (pre)frail groups were compared using Pearson Chi-squared and Student’s t tests. Associations between socioeconomic characteristics and baseline frailty were assessed using logistic regression models adjusted for sex, age, cohort, living alone, marital status, and number of children. First, separate models were built for each socioeconomic measure. Then a single model included SES measures at all life stages to estimate their mutually adjusted association with baseline frailty. To avoid overadjustment due to multiple subjective SES measures in early old age (i.e., income, wealth, and financial strain), income and wealth were not entered in this mutually adjusted model.

Frailty trajectories were identified using group-based trajectory modelling (GBTM), which we have extensively applied to the Lc65+ data [20]. In brief, GBTM is based on a finite set of polynomial functions of time to model latent groups with similar trajectories. The number and shapes of frailty trajectories was based on the Bayesian information criterion (BIC) and Bayes factors. As dropouts are generally more frequent in individuals with health and socioeconomic vulnerabilities, the missing at random assumption may generate biased estimates of trajectory group size. Therefore, we used a GBTM extension that models the probability of dropout because of death or illness as a function of time and two prior observed outcomes using a logit distribution [21]. Associations between socioeconomic characteristics and frailty trajectories were tested using logistic regression, also using separate models and a mutually adjusted model (see above) and adjusting for the same variables plus baseline frailty.

A sensitivity analysis was carried out to explore the robustness of the results when including individuals lost to follow-up for reasons other than death or illness [22]. A second sensitivity analysis was performed to examine the potential influence of chronic conditions, smoking history, and problematic alcohol history on the main findings.

In this population of community-dwelling older adults, we found that measures of SES at different points of the life-course are all associated with worse frailty levels. A lower education, occupational class, and worse financial situation in older age are independently associated with higher frailty levels shortly after age 65. In addition, a worse financial situation in early old age is associated with both higher frailty levels as well as worse frailty trajectories after age 65. Findings suggest that SES inequalities in frailty are likely the results of influences across early, mid-life and later life stages, requiring interventions throughout the life-course. In addition, policies aimed at improving the financial situation of older people may help curb inequalities in frailty trajectories.

The associations between socioeconomic characteristics and baseline frailty are consistent with previous studies indicating a particularly high prevalence of frailty in older people with lower education and income [23]. Three competing hypotheses summarize how these associations may change over the life course [24]: The “age-as-leveler” hypothesis posits that socioeconomic inequalities in health narrow in old age after a peak in mid-life or early old age. On the other hand, the “status maintenance” hypothesis suggests that inequalities that develop in early- or mid-life do not increase or decrease with age. Finally, the “cumulative advantage” hypothesis postulates that socioeconomic inequalities in health widen over the life course due to a progressive accumulation of disadvantages. In their literature review, Wang and Hulme compiled substantial evidence of socioeconomic inequalities in frailty [7], and claimed that “age-as-leveler” hypothesis was the prevailing pattern across most of the reviewed studies. While the present study was not specifically designed to formally identify an age pattern in socioeconomic inequalities, it is still interesting to note that SES measures across all life stages were independently associated with baseline frailty, suggesting that socioeconomic inequalities in frailty accumulate and are evident in individuals in their late sixties.

Previous studies aiming to differentiate multiple patterns of frailty trajectories have consistently identified three trajectories as the optimal model fit [10–12, 14, 25]. Our results similarly align with prior research indicating an association between unfavorable frailty trajectories and lower levels of education [10–12, 25] and occupational class [10]. Additionally, studies focusing on a single pattern of frailty change over time have shown that a linear or quadratic increase in frailty is linked with lower education and occupation [24, 26–28]. In line with our findings, Gardiner et al. [10] found that early-life SES measures were not independently associated with frailty trajectories when taking into account later-life SES measures. Other studies have also found that the pathways to the development of frailty begin early in life but are later mediated by socioeconomic factors [29].

In the present study, adjusting for the baseline level of frailty enabled us to isolate the factors specifically linked with the trajectory of frailty, and to determine whether ‘new’ inequalities emerge in older age, in addition to those generated by exposures earlier in life. We found that financial strain and the receipt of financial subsidies in older age were independently associated with the least favorable trajectories of frailty, whereas no associations were observed between educational level or occupational class and frailty trajectories in mutually adjusted models. One possible interpretation is that the associations for educational level and occupational class are mediated by financial measures of SES in later life. The financial situation in later life may affect frailty trajectories in multiple ways [30–32]. For instance, individuals experiencing financial strain may live in lower-quality housing that is less suited to supporting physical function maintenance, thereby creating an environment less favorable for healthy aging. Additionally, limited financial resources can restrict access to paid care and support services, potentially accelerating the decline in physical function. Other factors may include inadequate nutrition, limited access to physical activity opportunities, increased stress, reduced social engagement, and inability to afford preventive healthcare, indicating that financial stability is a key factor in mitigating frailty in older adults.

A major strength of this study lies in its utilization of data from the Lc65+ study, which features a large and representative sample of community-dwelling older people. The Lc65+ study also provides multiple SES measures at various stages throughout the life-course. Frailty trajectories were assessed using a validated tool that combines both self-report and objective measures. However, some important limitations should be considered. First, merging pre-frail and frail individuals was necessary due to the low prevalence of frailty. This may be attributed to a lower occurrence of physical frailty in Switzerland compared to other European countries [33], as well as in studies employing Fried’s phenotype versus the cumulative deficit model [34]. Second, while the focus remained on physical frailty, it is worth noting that other conceptual definitions encompass additional dimensions such as psychological, social, and cognitive frailty. These dimensions may follow distinct trajectories, originating through different pathways. Third, our estimates lack a causal interpretation and solely focus on describing associations between SES and frailty. For example, it is possible that worse frailty trajectories could impact the financial situations of older households, potentially strengthening these associations. Future studies should explore the causal nature of the association between different SES measures and frailty trajectories.

Policies that break down class barriers rather than reinforce them—such as family-friendly social protection, inclusive education, and fair employment—can help reduce health inequalities and prevent their transmission across generations [35]. Research on income-support interventions, including cash transfer programs, income tax credits, and minimum wage policies, has shown modest but meaningful effects at the population level, with consistent long-term benefits [36]. However, further evaluations are needed to determine how these policies can be optimized for greater public health impact. Tackling socioeconomic inequalities early in life can help narrow disparities in frailty trajectories later on, ultimately supporting healthier aging for future generations.

While early-life interventions are crucial for reducing health inequalities, financial protection measures in older age also play a key role in promoting healthy aging. Policies aimed at securing adequate pension schemes, subsidized healthcare, and targeted social assistance can help mitigate the economic vulnerabilities that often intensify in later life [37]. Ensuring stable income and access to essential services can reduce stress, improve healthcare utilization, support social engagement, and promote overall wellbeing among older adults [38]. As populations continue to age, strengthening these financial protection mechanisms will be essential to fostering equity in health outcomes and preventing further socioeconomic-driven disparities in frailty.

This study identified three distinct frailty trajectories in a Swiss population of community-dwelling older adults. We observed socioeconomic inequalities in frailty shortly after age 65, indicated by independent associations with educational level, occupational class, and both objective and subjective measures of financial situation in early old age. In addition, poorer financial situations in early old age were associated with worse frailty trajectories after age 65. Our findings suggest that addressing disparities in frailty requires interventions across the life-course. They also underscore the critical role of policies aimed at improving the financial wellbeing of older people in reducing SES inequalities in frailty trajectories during older age.