Skip to main content
PLOS One logoLink to PLOS One
. 2026 Nov 3;21(11):e0361826. doi: 10.1371/journal.pone.0361826

Sex and educational inequalities across functional ability trajectories in Chile: A sequence analysis of a 17-year national panel

Ignacio Cabib 1,2, Pablo Villalobos Dintrans 3,4,5,*, Jorge Browne Salas 5,6
Editor: Oyelola A Adegboye7
PMCID: PMC13730331  PMID: 42976219

Abstract

Context

Research on disability and functional ability has increasingly moved from cross-sectional descriptions toward longitudinal analyses of intra-individual change. This shift has shown that functional ability follows heterogeneous pathways, encompassing stability, progressive deterioration, fluctuating limitations, partial recovery, and death. However, evidence remains concentrated in high-income countries, and fewer studies have examined long-term functional ability trajectories in Latin American contexts marked by rapid population aging and high socioeconomic inequality. In addition, comparisons across sex and educational groups require trajectory typologies that are constructed in a way that makes group comparisons methodologically comparable.

Methods

We used six observation points from the Chilean Social Protection Survey, a nationally representative longitudinal survey, covering the period 2004–2020. The analytic sample included 2,681 individuals aged 46–65 at baseline, contributing 16,086 person-wave observations. Functional ability was measured through self-reported limitations in activities of daily living and classified into five states: no limitation, low limitation, middle limitation, high limitation, and deceased. We reconstructed individual functional ability sequences over 17 years and applied optimal matching followed by hierarchical clustering. To ensure comparability across sex and educational groups, trajectory typologies were estimated within each baseline age group using the full age-group-specific sample. We then examined how membership in these pooled trajectory types varied by sex and educational attainment, referencing the composition of each trajectory type to the overall composition of its baseline age group.

Findings

Across the four baseline age groups, detailed trajectories were grouped into three broader types: mostly non-declining, progressive-declining, and equivocal-declining. Mostly non-declining trajectories were prevalent in all age groups, although their internal composition changed with baseline age, moving from stable good functional ability in younger age groups toward preserved functional ability before death in older age groups. Progressive-declining trajectories were especially visible among respondents aged 56–60 at baseline and remained substantial among those aged 61–65. Equivocal-declining trajectories were more frequently composed of women, most markedly among respondents aged 61–65 at baseline, and of respondents with primary education or less in the three younger age groups. Sex differences in trajectory membership were statistically significant in all age groups, while educational differences were statistically significant in the first three age groups.

Conclusions

Long-term functional ability trajectories in Chile are heterogeneous and socially patterned. By deriving pooled trajectory typologies within each baseline age group before comparing sex and educational groups, and by referencing group composition to the marginal composition of each age group, this study provides comparable evidence on how functional aging unfolds across socially stratified groups in an understudied Latin American context.

Introduction

Functional ability research has changed substantially over recent decades. Earlier approaches often examined disability or functional limitations as static outcomes observed at one point in time. More recent research has emphasized the dynamic nature of functional ability, conceptualizing it as a process that unfolds through transitions between independence, limitation, recovery, worsening health, and death. This shift has been supported by the availability of longitudinal data and by methodological approaches capable of representing individual trajectories over extended periods. The concept of functional ability has also become increasingly relevant because it moves the analytical focus from disability as a fixed attribute toward people’s capacities and limitations in daily life, consistent with broader changes in the terminology and conceptualization of disability and aging [1,2].

A central contribution of longitudinal disability and functional ability research has been the identification of heterogeneous trajectories. The importance of examining disability as a dynamic process rather than as a single status measured at one point in time has been increasingly emphasized [2,3]. Research has shown that functional aging follows several distinct pathways. Some individuals maintain high functional ability over long periods, others experience gradual deterioration, and others move between lower and higher levels of limitation. Later-life health has been characterized as involving gradual change, homeostasis, and punctuated equilibrium, which draws attention to stability, instability, and non-linear transitions [3]. Previous research in Chile has also shown that functional ability trajectories include stable, progressive-declining, equivocal-declining, and mortality-related patterns [4].

This trajectory-based perspective has also raised questions about the social patterning of functional ability, and evidence from high-income settings has begun to specify the channels through which such patterning operates. Sex differences are consistently documented, with women accumulating more functional limitations and more unstable patterns while surviving longer, a configuration linked to a higher burden of non-fatal disabling conditions, to differences in muscle mass and strength, and to lifetime differences in paid work and caregiving responsibilities [5–8].

Socioeconomic position operates through a broader set of intermediate mechanisms. Educational attainment shapes exposure to physically demanding and hazardous occupations, health literacy and the capacity to navigate health systems, income and material conditions in later life, and health behaviors such as physical activity [9–13]. Chronic conditions, multimorbidity, and sedentary behavior operate both as precipitating events that trigger transitions into limitation and as constraints on the likelihood of regaining function once limitation has occurred, so that advantaged groups more frequently display recovery and less frequently remain in unstable states [14–18].

These mechanisms accumulate across the life course, with early-life socioeconomic circumstances and adulthood employment histories shaping functional status decades later [19]. Evidence from Taiwan, China, Hong Kong, the United States, Japan, and the United Kingdom converges in showing that functional decline is shaped by this configuration of individual and social factors rather than by chronological age alone, although the strength and even the direction of specific associations varies across age groups and cohorts [7,8,12–18].

Research gaps

Despite these advances, two gaps remain relevant. First, most evidence on long-term functional ability trajectories comes from the United States, Europe, and parts of Asia [7,8,12–18]. Evidence from Latin America remains more limited, even though the region is characterized by rapid population aging, high socioeconomic inequality, and institutional arrangements that differ from those observed in countries where most trajectory studies have been conducted. Chile is an especially relevant setting because it combines rapid demographic aging, pronounced social inequalities, and a segmented health and social protection system.

These features carry specific implications for how functional ability trajectories are expected to unfold. Demographic aging in Chile is occurring at a pace that compresses into a few decades a transition that took considerably longer in the settings where most trajectory evidence has been produced, so that successive cohorts reach later life with markedly different educational and occupational endowments. Educational expansion has been rapid, and the proportion of individuals with primary education or less rises steeply across the cohorts examined here, which suggests that educational differences in functional trajectories will be entangled with cohort composition and that the social meaning of a given educational threshold shifts across baseline age groups [20].

The Chilean labor market has been characterized by weak employment protection and persistent informality, conditions that concentrate cumulative exposure to physically demanding and unprotected work among lower-educated individuals and that plausibly produce steeper educational gradients than those observed in settings with stronger labor regulation [19]. The health system is segmented by ability to pay, and long-term care provision remains limited, with families absorbing most of the care burden, which makes the likelihood of regaining function after an episode of limitation dependent on household resources and suggests that socioeconomic differences will be expressed in the stability and reversibility of trajectories as much as in their onset [4,21].

Alongside these context-specific expectations, the basic heterogeneity of functional pathways, encompassing sustained stability, ordered decline, fluctuation, and mortality-related patterns, is expected to resemble that documented elsewhere, since it reflects demographic and biological regularities that are common across settings. Previous studies in Chile have extended functional ability trajectory analysis to this national context, [4] and have examined the association between long-term functional ability trajectories and mental health during the COVID-19 period [22]. However, evidence remains limited on how sex and educational inequalities are distributed across comparable long-term functional ability trajectories.

Second, the methodological basis for comparing trajectories across social groups requires careful attention. If long-term trajectories are estimated separately within sex or educational groups, the resulting clusters may represent different empirical constructs. Under that design, a trajectory labeled as stable, declining, or fluctuating in one group may not be directly equivalent to a trajectory with the same label in another group. Comparative analyses therefore require a common trajectory typology constructed before examining group differences. This approach allows sex and education to be analyzed as dimensions associated with membership in shared trajectory categories rather than as criteria used to create separate and potentially non-equivalent cluster solutions.

The present study extends previous Chilean research on functional ability trajectories in three ways. First, it uses a 17-year observation window to examine functional ability from midlife into older age, building on prior work using the Chilean Social Protection Survey [4,22]. Second, it estimates pooled trajectory typologies within baseline age groups, allowing sex and educational differences to be examined across comparable trajectory categories. Third, it groups detailed trajectories into three broader types (mostly non-declining, progressive-declining, and equivocal-declining), to describe the main ways in which functional ability changes across adulthood and older age in Chile.

The study asks: How are sex and educational attainment associated with membership in long-term functional ability trajectories among Chilean adults observed from midlife into older age? By addressing this question, the article contributes to life-course and aging research by extending trajectory-based evidence to an understudied Latin American context and by clarifying how sex and educational inequalities are expressed across comparable functional ability pathways.

Materials and methods

Data source and study design

We used data from the Encuesta de Protección Social (Social Protection Survey, EPS), a nationally representative Chilean longitudinal survey [23]. The EPS provides publicly available microdata for research on social protection, employment histories, education, income, health, household characteristics, and related dimensions.

The EPS has been fielded in 2002, 2004, 2006, 2009, 2012, 2015, 2018, 2020, and 2024. The present study used six observation points with functional ability information covering the period 2004–2020. The 2002 wave was excluded because questions on functional ability status were incorporated into the survey only from the 2004 wave onwards, and because its sample was restricted to individuals affiliated to the pension system, which limits its representativeness of the adult population [23]. The 2012 wave was published with reservations by the institution responsible for the survey, owing to low response rates during fieldwork and an accompanying recommendation that these data be used with caution for statistical inference, and it was therefore excluded from the analytic window [23]. The 2024 wave had not been released at the time the analyses were conducted and will be incorporated into future extensions of this work. This observation window provides an extended observation period in which functional ability changes can be observed from midlife into older age. The length of the panel is especially relevant for identifying different types of trajectories, including stable functional ability, gradual decline, fluctuating limitation, recovery, and death. Previous research on later-life health trajectories has emphasized that long observation windows are particularly important for detecting non-linear transitions, homeostasis, punctuated change, and mortality-related pathways [3,24].

A further data-quality issue warrants explicit mention. Fieldwork for the 2020 wave began face-to-face in late 2019 and was suspended in March 2020 following the introduction of nationwide COVID-19 lockdown measures. Data collection resumed between September and December of that year by telephone, yielding a smaller achieved sample and a change in survey mode for part of the wave [22]. This wave was included in the analysis, and the implications of this disruption are discussed among the limitations below.

The analytic sample included individuals aged 46–65 in 2004. This baseline age range was selected for two reasons. First, it allows the analysis to begin before most respondents reach older ages, making it possible to observe transitions from midlife into later life. Second, it provides sufficient variation in age at baseline to examine whether trajectory structures differ depending on the stage of the life course at which respondents enter the observation window. Respondents were divided into four five-year baseline age groups: 46–50, 51–55, 56–60, and 61–65 years. These age groups were analyzed separately because functional ability dynamics differ substantially depending on the age at which individuals enter the observation period. Table 1 presents descriptive statistics for the full analytic sample, including distributions by sex, education, and functional ability status at baseline and ending year, across these age groups.

Table 1. Study samples and descriptive statistics at baseline (2004).

Age group 1

N = 743

(27.7%)
Age group 2

N = 711

(26.5%)
Age group 3

N = 657

(24.5%)
Age group 4

N = 570

(21.3%)
Sex
Women 54.6% 55.3% 49.8% 46.3%
Men 45.4% 44.7% 50.2% 53.7%
Educational level
Primary education or none 44.3% 50.1% 55.1% 67.0%
Secondary education or more 55.7% 49.9% 44.9% 33.0%
Functional limitation status in baseline year (2004)
No limitations 95.2% 95.9% 91.9% 89.0%
Low limitations 2.6% 2.7% 4.0% 5.6%
Middle limitations 0.5% 0.1% 0.3% 0.7%
High limitations 1.7% 1.3% 3.8% 4.7%
Functional limitation status in ending year (2020)
No limitations 70.4% 63.0% 41.4% 23.0%
Low limitations 8.1% 7.9% 11.0% 8.8%
Middle limitations 0.7% 0.7% 0.3% 1.2%
High limitations 8.9% 10.1% 11.9% 8.3%
Deceased 12.0% 18.3% 35.5% 58.8%

Note: Age group 1 = Individuals aged 46–50 in 2004. Age group 2 = Individuals aged 51–55 in 2004. Age group 3 = individuals aged 56–60 in 2004. Age group 4 = individuals aged 61–65 in 2004.

From the initial pool of eligible respondents, we retained individuals with no more than one missing functional ability observation across the six waves. This criterion was used to preserve the integrity of sequence reconstruction while retaining a broad analytic sample. For respondents with one missing functional ability observation, missing states were imputed using multiple imputation by chained equations with predictive mean matching, incorporating age, sex, educational attainment, and adjacent functional ability states as predictors. To assess the robustness of our findings to the imputation strategy, we re-estimated the sequence and cluster analysis on the subsample of individuals with fully complete functional ability information across all six waves, excluding all imputed observations (S1 Appendix). As reported there, the imputation of a single missing wave did not materially affect the trajectory typology where its effect could be isolated from sample-size reduction.

The final sample included 2,681 individuals and 16,086 person-wave observations. Individuals who died during the observation period were retained in the analysis, and death was treated as a functional ability state. This decision follows previous trajectory-based research that incorporates mortality into the study of later-life functional dynamics [3,4].

The analytic sample is defined by observability of functional ability status across the six waves and is therefore representative of respondents who remained in the panel rather than of the Chilean population aged 46–65 in 2004. S2 Table compares the analytic sample with the broader sample that includes individuals with missing observations and with a sample excluding deceased individuals. Differences between the analytic sample and the broader sample are modest and consistent in direction. Specifically, the analytic sample contains between 3.4 and 5.0 percentage points more respondents with primary education or less across the four age groups, a somewhat higher proportion of women in the three younger age groups, and a marginally lower baseline prevalence of no limitations in three of the four age groups. Attrition therefore does not appear to have retained a more advantaged or healthier subset of respondents. Retaining individuals who died during the observation period, and treating death as a functional ability state, further limits one of the principal sources of selective loss in long panels of aging populations.

Functional ability measure

Functional ability was measured using self-reported difficulties in activities of daily living available in the EPS. We constructed a five-state categorical measure based on limitations in walking long distances, bathing, dressing, eating, and getting out of bed. The states were classified as follows: no limitation, low limitation, middle limitation, high limitation, and deceased.

The hierarchical classification reflects increasing severity in functional limitation. Respondents reporting no difficulty in the activities considered were classified as having no limitation. Low limitation captured difficulty walking long distances. Middle limitation captured difficulty bathing. High limitation captured difficulty with dressing, eating, or getting out of bed. As noted above, the deceased state was included to avoid treating mortality as missing information and to represent trajectories that end in death as part of the observed functional process.

This classification follows previous work on functional ability trajectories in Chile, where limitations in activities of daily living were used to reconstruct long-term functional ability states [4]. It is also consistent with broader research on functional ability and intrinsic capacity, which has emphasized the relevance of hierarchical and multidimensional approaches to later-life functioning [25]. As a sensitivity analysis, we re-estimated trajectories using the number of reported limitations instead of the hierarchical state variable (S3 Appendix). As reported there, the broad trajectory structure and the age-graded shift from sustained stability toward death-related trajectories were broadly preserved, although concordance decreased with baseline age.

Sex and educational attainment

Sex was measured using the variable available in the EPS and categorized as women and men. Following the Sex and Gender Equity in Research guidelines, we refer to sex when describing this variable because the survey records sex rather than gender identity [26].

Educational attainment was categorized as primary education or less versus secondary education or more. This classification reflects a relevant educational threshold for the cohorts under study and maintains sufficient sample size for trajectory comparisons within age groups. Education is commonly used in functional ability and disability research as an indicator of socioeconomic position and cumulative life-course resources [4,7–12]. In Chile, educational attainment differs substantially across cohorts, with older respondents more frequently concentrated in lower educational categories. This age-related educational composition was considered when interpreting differences across baseline age groups [20].

Sequence analysis and trajectory classification

We used sequence analysis to reconstruct individual functional ability trajectories across six observation points from 2004 to 2020 [27]. Each respondent’s trajectory was represented as an ordered sequence of functional ability states. Sequence analysis is well suited for this study because it captures both the timing and ordering of transitions between functional states, including stability, decline, recovery, fluctuation, and death. This method has been increasingly used in life-course research because it allows researchers to analyze trajectories as whole sequences rather than as isolated transitions or single outcomes [28].

For each baseline age group, we calculated pairwise dissimilarities between individual sequences using optimal matching. Substitution costs were set to 2, and hierarchical clustering using Ward’s method was applied to the resulting dissimilarity matrix. Cluster solutions were evaluated using multiple criteria, including average silhouette width, point biserial correlation, Hubert’s Gamma, and Hubert’s C. Fig 1 presents the normalized scores (Z-scores) for each criterion for each age group. Based on these metrics and guided by theoretical considerations [29], we selected the number of clusters that best captured the heterogeneity of functional ability trajectories within each age group. The gray dashed vertical lines indicate the selected cluster solutions. Specifically, the optimal number of clusters was seven among individuals aged 46–50 at baseline, six among those aged 51–55, eight among those aged 56–60, and six among those aged 61–65.

Fig 1. Cluster cut-off criteria across age groups.

Fig 1

Note: Gray dashed vertical lines indicate the optimal number of representative trajectory clusters need to summarize the diversity of functional ability trajectories. PBC = point biserial correlation; HG = Hubert’s Gamma; ASW = average silhouette width; HC = Hubert’s C. Normalized scores shown. Hogher ASW, PCB, and HG values indicate better cluster solutions. Lower HC values indicate better cluster solutions.

Analytical strategy

The analysis proceeded in four steps. First, we described baseline characteristics of the analytic sample by age group, including sex, educational attainment, and functional ability status. Second, we identified detailed functional ability trajectories within each age group using sequence analysis and cluster analysis. Third, we grouped the detailed trajectories into three broader trajectory types: mostly non-declining, progressive-declining, and equivocal-declining. Fourth, we examined whether membership in these trajectory types differed by sex and educational attainment, using chi-squared tests and referencing the composition of each trajectory type to the overall sex and educational composition of its baseline age group, so that departures from the marginal distribution can be distinguished from compositions that reproduce it. All analyses were conducted in R [30], using the TraMineR [29] and mice [31] packages.

Analyses were conducted without survey weights. This decision follows from the design of the study in three respects. The construction of the typology rests on pairwise dissimilarities between sequences and on hierarchical clustering, procedures into which expansion factors cannot be incorporated in a straightforward manner at the stage where trajectory types are defined. The substantive claims concern the sex and educational composition of shared trajectory categories within each baseline age group, a within-sample comparison for which weighting is less consequential than it would be for national prevalence estimation. Previous research using the same survey and a comparable trajectory design reported highly similar results with and without the expansion factor, which is consistent with an accurate sampling strategy and adequate fieldwork in the EPS [4].

Ethics statement

This study used the EPS, a publicly available secondary dataset. The study respected the original consent and confidentiality procedures under which the EPS data were collected and made available.

Results

Overview of the trajectory typologies

The sequence analysis identified a set of functional ability trajectories within each baseline age group. These trajectories were estimated using the full age-group-specific samples and were then grouped into three broader trajectory types: mostly non-declining, progressive-declining, and equivocal-declining. This analytical strategy allows comparisons by sex and education to be made after trajectory types were defined within the pooled sample of each age group. Importantly, the grouping into three broad types corresponds to an interpretive harmonization based on the qualitative shape of the detailed trajectories, namely the presence, ordering, and timing of transitions toward more severe limitation states or death. This grouping is intended to support cross-age description and does not imply that the underlying detailed clusters are statistically equivalent across age groups.

The mostly non-declining type includes trajectories characterized by sustained absence of functional limitations or preserved functional ability before death. The progressive-declining type includes trajectories marked by a more ordered movement from no or low limitation toward more severe limitation or death. The equivocal-declining type includes trajectories characterized by non-linear, unstable, or fluctuating functional ability patterns, including movements across limitation states and, in some cases, partial recovery. Fig 2 presents the detailed trajectories composing each of the three broader types for each age group using sequence index plots. In these plots, the X-axis represents age and each horizontal line represents one individual trajectory. The plots therefore show intra-individual transitions between functional ability states over time.

Fig 2. Functional ability trajectory types across four age groups.

Fig 2

Functional ability trajectories across age groups

As shown in Fig 2, mostly non-declining trajectories represented a large share of respondents across the four age groups, but their internal composition changed with baseline age. Among individuals aged 46–50 at baseline, 58.7% were classified in mostly non-declining trajectories. This type was composed mainly of the stable good trajectory, which represented 52.6% of the cohort, and the good before mid-period death trajectory, which represented 6.1%. Progressive-declining trajectories represented 10.0% of the cohort and were composed of late low decline and progressive decline before early death, each accounting for 5.0%. Equivocal-declining trajectories represented 31.4%, including early severe decline at 7.3%, late severe decline at 10.4%, and equivocal recovery at 13.7%.

Among individuals aged 51–55 at baseline, mostly non-declining trajectories represented 49.3% of the cohort. This type included stable good at 41.4% and good before mid-period death at 7.9%. Progressive-declining trajectories represented 7.0%, corresponding to the single trajectory progressive decline before early death. Equivocal-declining trajectories represented 43.8%, composed of early severe decline at 5.9%, late severe decline at 11.7%, and equivocal recovery at 26.2%. This age group therefore shows the highest prevalence of equivocal-declining trajectories across the four groups.

Among individuals aged 56–60 at baseline, mostly non-declining trajectories represented 56.7% of the cohort. This type included stable good at 27.5%, good before mid-period death at 6.2%, and good before late death at 23.0%. Progressive-declining trajectories represented 37.4%, making this the age group with the largest progressive-declining component. This broad type included mid-period low decline at 13.9%, progressive severe decline at 5.9%, progressive decline before early death at 9.7%, and progressive decline before late death at 7.9%. Equivocal-declining trajectories represented 5.8%, corresponding to the single trajectory early low decline.

Among individuals aged 61–65 at baseline, mostly non-declining trajectories represented 54.9% of the cohort. This type included stable good at 12.1%, good before mid-period death at 21.2%, and good before late death at 21.6%. Progressive-declining trajectories represented 27.2%, composed of progressive decline before early death at 10.0% and progressive decline before late death at 17.2%. Equivocal-declining trajectories represented 17.9%, corresponding to the trajectory late severe decline.

Sex differences in trajectory types across age groups

As shown in Table 2, sex differences in trajectory membership were statistically significant in all four age groups when considering both the three main trajectory types and the detailed trajectories. Specifically, for the three main types, the association between sex and trajectory membership was statistically significant in age group 1, χ2 = 17.27, p < .001; age group 2, χ2 = 15.14, p < .001; age group 3, χ2 = 8.08, p = .018; and age group 4, χ2 = 24.97, p < .001.

Table 2. Sex composition of main and detailed functional ability trajectory types by age group.

Age group 1

(46-50/62-66)
Age group 2

(51-55/67-71)
Age group 3

(56-60/72-76)
Age group 4

(61-65/77-81)
Men

(%)
Women

(%)
– Men

(%)
Women

(%)
– Men

(%)
Women

(%)
– Men

(%)
Women

(%)
–
Age Group Composition

(All Respondents)
45.4 54.6 – 44.7 55.3 – 50.2 49.8 – 53.7 46.3 –
Main Trajectory Types Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value
Mostly Non-Declining Trajectories 51.6 48.4 <  .001 50.3 49.7 <  .001 54.7 45.3 .018 62.3 37.7 <  .001
Progressive-Declining Trajectories 40.5 59.5 56.0 44.0 43.1 56.9 48.4 51.6
Equivocal-Declining Trajectories 35.2 64.8 36.7 63.3 52.6 47.4 35.3 64.7
Detailed Trajectory Types Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value Men

(%)
Women

(%)
P value
Stable good 49.9 50.1 <  .001 47.6 52.4 <  .001 61.3 38.7 <  .001 46.4 53.6 <  .001
Good before mid-period death 66.7 33.3 64.3 35.7 63.4 36.6 57.9 42.1
Good before late death – – – – 44.4 55.6 75.6 24.4
Mid-period low decline – – – – 38.5 61.5 – –
Late low decline 29.7 70.3 – – – – – –
Progressive severe decline – – – – 20.5 79.5 – –
Progressive decline before early death 51.4 48.6 56.0 44.0 56.2 43.8 66.7 33.3
Progressive decline before late death – – – – 51.9 48.1 37.8 62.2
Early low decline – – – – 52.6 47.4 – –
Early severe decline 22.2 77.8 28.6 71.4 – – – –
Late severe decline 33.8 66.2 38.6 61.4 – – 35.3 64.7
Equivocal recovery 43.1 56.9 37.6 62.4 – – – –

Note: Percentages indicate the composition of each trajectory type by sex. The first row reports the sex composition of all respondents in each baseline age group, so that trajectory-specific compositions can be read as departures from the marginal distribution. Chi-squared tests evaluate whether trajectory membership differs by sex within each age group. P-values of chi-squared tests are reported.

Interpreting these compositions requires reference to the overall sex composition of each baseline age group. Women constitute 54.6% of respondents aged 46–50 at baseline and 55.3% of those aged 51–55, whereas men constitute 50.2% of those aged 56–60 and 53.7% of those aged 61–65 (Table 1). Table 2 reports these marginal distributions alongside the trajectory-specific compositions, and the description below expresses each composition as a departure in percentage points from the marginal proportion of the corresponding age group. A departure is described as reversing the aggregate composition when the majority within a trajectory type differs from the majority in the age group as a whole, and as reinforcing it when the two coincide.

Mostly non-declining trajectories had a higher proportion of men than women in all four age groups, although the margin was minimal in age group 2. In age group 1, mostly non-declining trajectories were composed of 51.6% men and 48.4% women, which reverses the aggregate composition of an age group in which women predominate, departing by 6.2 percentage points from the marginal proportion of men. The same holds in age group 2, where a composition of 50.3% men and 49.7% women, although close to parity in absolute terms, departs by 5.6 percentage points from a marginal of 44.7% men and therefore also reverses the aggregate composition. In age groups 3 and 4, the male majority in this type, 54.7% and 62.3% respectively, reinforces an aggregate composition in which men already predominate, with departures of 4.5 and 8.6 percentage points from the corresponding marginals.

Equivocal-declining trajectories were composed of a higher proportion of women in age groups 1, 2, and 4. In age group 1, this type was composed of 35.2% men and 64.8% women, and in age group 2 of 36.7% men and 63.3% women; both compositions reinforce a female-majority aggregate distribution, with departures of 10.2 and 8.0 percentage points. In age group 4, a composition of 35.3% men and 64.7% women reverses an aggregate composition in which men constitute 53.7% of respondents, and represents the largest single departure in Table 2, at 18.4 percentage points. In age group 3, the composition of this type, 52.6% men and 47.4% women, closely reproduces a nearly balanced marginal, departing by 2.4 percentage points.

Progressive-declining trajectories showed more variation across age groups. Women represented the larger share in age groups 1, 3, and 4, while men represented the larger share in age group 2. In age group 1, the female share of 59.5% reinforces a female-majority aggregate composition, exceeding the marginal by 4.9 percentage points. The remaining three departures reverse the aggregate composition of their age group: 56.0% men in age group 2, exceeding a marginal of 44.7% by 11.3 percentage points; 56.9% women in age group 3, exceeding a marginal of 49.8% by 7.1 percentage points; and 51.6% women in age group 4, exceeding a marginal of 46.3% by 5.3 percentage points. Overall, the concentration of women in equivocal-declining trajectories is the most consistent pattern across age groups, and it is most pronounced in age group 4, where it inverts the aggregate sex composition of the age group.

Educational differences in trajectory types across age groups

Table 3 indicates educational differences in trajectory membership. The association between educational level and the three main trajectory types was statistically significant in the first three age groups: age group 1, χ2 = 11.00, p = .004; age group 2, χ2 = 9.29, p = .010; and age group 3, χ2 = 8.06, p = .018. In age group 4, the association was not statistically significant, although the p-value approached conventional thresholds, χ2 = 5.07, p = .079. For the detailed trajectories, education was significantly associated with trajectory membership in age groups 1, 2, and 3, but not in age group 4.

Table 3. Educational composition of main and detailed functional ability trajectory types by age group.

Age group 1

(46-50/62-66)
Age group 2

(51-55/67-71)
Age group 3

(56-60/72-76)
Age group 4

(61-65/77-81)
Low ed

(%)
High ed

(%)
– Low ed

(%)
High ed

(%)
– Low ed

(%)
High ed

(%)
– Low ed

(%)
High ed

(%)
–
Age Group Composition

(All Respondents)
44.3 55.7 – 50.1 49.9 – 55.1 44.9 – 67.0 33.0 –
Main Trajectory Types Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value
Mostly Non-Declining Trajectories 40.1 59.9 .004 44.3 55.7 .010 50.4 49.6 .018 63.9 36.1 .079
Progressive-Declining Trajectories 40.5 59.5 54.0 46.0 60.6 39.4 74.2 25.8
Equivocal-Declining Trajectories 53.2 46.8 55.9 44.1 65.8 34.2 65.7 34.3
Detailed Trajectory Types Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value Low ed

(%)
High ed

(%)
P value
Stable good 37.9 62.1 <  .001 41.8 58.2 .008 40.3 59.7 <  .001 59.4 40.6 .246
Good before mid-period death 60.0 40.0 57.1 42.9 65.9 34.1 67.8 32.2
Good before late death – – – – 58.3 41.7 62.6 37.4
Mid-period low decline – – – – 58.2 41.8 – –
Late low decline 21.6 78.4 – – – – – –
Progressive severe decline – – – – 61.5 38.5 – –
Progressive decline before early death 59.5 40.5 54.0 46.0 57.8 42.2 75.4 24.6
Progressive decline before late death – – – – 67.3 32.7 73.5 26.5
Early low decline – – – – 65.8 34.2 – –
Early severe decline 59.3 40.7 54.8 45.2 – – – –
Late severe decline 46.8 53.2 62.7 37.3 – – 65.7 34.3
Equivocal recovery 54.9 45.1 53.2 46.8 – – – –

Note: “Low ed” = primary education or less. “High ed” = secondary education or more. Percentages indicate the composition of each trajectory type by education. The first row reports the educational composition of all respondents in each baseline age group, so that trajectory-specific compositions can be read as departures from the marginal distribution. Chi-squared tests evaluate whether trajectory membership differs by educational level within each age group. P-values of chi-squared tests are reported.

As with sex, the educational composition of the trajectory types requires reference to the marginal distribution of each age group. Respondents with primary education or less constitute 44.3% of age group 1, 50.1% of age group 2, 55.1% of age group 3, and 67.0% of age group 4 (Table 1). Table 3 reports these marginals alongside the trajectory-specific compositions.

Mostly non-declining trajectories had a higher proportion of individuals with secondary education or more in age groups 1 and 2. In age group 1, this type was composed of 59.9% individuals with secondary education or more and 40.1% individuals with primary education or less, a departure of 4.2 percentage points below the marginal proportion with primary education or less. In age group 2, the corresponding figures were 55.7% and 44.3%, a departure of 5.8 percentage points below the marginal. In age group 3, the mostly non-declining type was nearly balanced by education, with 49.6% individuals with secondary education or more and 50.4% individuals with primary education or less, which departs by 4.7 percentage points below a marginal of 55.1% and therefore reverses a low-education aggregate majority into near parity. In age group 4, mostly non-declining trajectories were composed mainly of individuals with primary education or less, 63.9%, a figure that falls 3.1 percentage points below the marginal of 67.0% and therefore reproduces the educational composition of this older age group.

Progressive-declining trajectories became increasingly composed of lower-educated respondents across age groups. The proportion of individuals with primary education or less within progressive-declining trajectories increased from 40.5% in age group 1 to 54.0% in age group 2, 60.6% in age group 3, and 74.2% in age group 4. Referenced to the marginals, these figures correspond to a departure of 3.8 percentage points below the age-group composition in age group 1 and to departures of 3.9, 5.5, and 7.2 percentage points above it in age groups 2, 3, and 4. The educational gradient in this trajectory type therefore strengthens with baseline age both in absolute terms and relative to the composition of each age group.

Equivocal-declining trajectories were composed of 53.2% respondents with primary education or less in age group 1, 55.9% in age group 2, 65.8% in age group 3, and 65.7% in age group 4. Referenced to the marginals, the departures are 8.9, 5.8, and 10.7 percentage points above the age-group composition in the first three groups, and the composition in age group 1 reverses an aggregate distribution in which respondents with secondary education or more constitute 55.7% of the age group. In age group 4, by contrast, the composition falls 1.3 percentage points below the marginal, so that the apparent concentration of lower-educated respondents in this trajectory type in the oldest age group reflects the educational composition of the age group itself. More generally, all three trajectory types in age group 4 fall within 7.2 percentage points of the marginal educational distribution, which is consistent with the absence of a statistically significant association in that age group. Taken together, educational differences are most visible in the composition of equivocal-declining trajectories in the three younger age groups and in progressive-declining trajectories in the two older age groups.

Discussion

This study examined long-term functional ability trajectories among Chilean adults aged 46–65 at baseline over a 17-year period. Its contribution is both substantive and methodological. Substantively, the findings show that functional aging in Chile unfolds through heterogeneous pathways that can be summarized as mostly non-declining, progressive-declining, and equivocal-declining trajectories. Methodologically, the analysis derives trajectory typologies within pooled age-group samples before comparing sex and educational groups, and references the composition of each trajectory type to the marginal composition of its age group.

The findings reinforce the value of treating functional ability as a dynamic process, since capturing functional aging requires more than a single contrast between limitation and no limitation at one point in time. The three broad trajectory types identified in this study reflect different temporal organizations of functional ability: sustained or preserved functioning, ordered functional deterioration, and unstable or equivocal transitions between functional states. This is consistent with the broader shift in disability and aging research from static descriptions toward longitudinal analyses of functional change.

One important implication of the results is that the same broad trajectory type can have different meanings depending on baseline age, emphasizing the benefits of early prevention [32]. Mostly non-declining trajectories were prevalent across the four age groups, but their internal composition changed substantially. In younger age groups, mostly non-declining trajectories mainly captured sustained functional stability. In older age groups, this category increasingly captured preserved functional ability before death. Both variants can be read as forms of healthy aging, understood as preserving function for as long as possible, but they are not interchangeable. In younger cohorts, it often describes uninterrupted functional stability over the observation window. In older cohorts, it may describe a pathway in which individuals remain functionally able for several waves before mortality becomes part of the trajectory. Longitudinal data are therefore essential for distinguishing sustained functional stability from preserved functioning before death.

The progressive-declining trajectories were especially visible from the late-fifties baseline group onward. This suggests that individuals aged 56–60 at baseline occupy a relevant position for observing ordered functional deterioration over a 17-year period. At this stage of the life course, respondents are old enough for functional limitations and mortality to become visible during follow-up, while still young enough for the ordering and timing of decline to be observed before the end of the panel. This finding contributes to the literature by showing that the visibility of progressive decline depends on where individuals are in the life course when longitudinal observation begins. The result also supports the value of sequence analysis for aging research because it captures the timing and order of functional transitions rather than reducing functional ability to a baseline status or an end-point outcome. Additionally, results support the importance of rehabilitation and early prevention as relevant strategies to improve functional ability in the population [33,34].

Sex differences in trajectory membership were evident across all age groups, but the form of these differences varied by baseline age. Men were generally more represented in mostly non-declining trajectories, especially in age groups where preserved functioning before death formed part of this category. Women were more represented in declining trajectories, particularly equivocal-declining trajectories in age groups 1, 2, and 4, and progressive-declining trajectories in age group 3. This suggests that sex differences in functional aging are expressed through the form and temporal organization of decline. Women’s higher representation in equivocal- and progressive-declining trajectories indicates greater concentration in pathways involving instability, severe limitation, or ordered deterioration, whereas men’s higher representation in mostly non-declining trajectories partly reflects pathways of sustained functioning and, in older cohorts, preserved functioning before death.

Referencing these compositions to the marginal distribution of each age group clarifies which of them constitute genuine departures from the aggregate composition and which reproduce it. Two configurations stand out. In the two younger age groups, where women form the majority of respondents, the male majority in mostly non-declining trajectories reverses the aggregate composition, which makes the association between sex and sustained functioning most clearly visible precisely where the aggregate distribution works against it. In the oldest age group, where men form the majority, the concentration of women in equivocal-declining trajectories reverses the aggregate composition by 18.4 percentage points, the largest departure observed. Conversely, the male majority in mostly non-declining trajectories in the two older age groups reinforces an already male-majority distribution, so that its interpretation depends on the magnitude of the departure rather than on the direction of the difference alone. The educational findings follow a comparable logic and require the same caution. Specifically, the concentration of respondents with primary education or less in equivocal-declining trajectories is a substantive departure in the three younger age groups, whereas in the oldest age group all three trajectory types track the marginal educational distribution closely, which accounts for the absence of a statistically significant association without implying an absence of educational disadvantage in that cohort.

Educational differences also contribute to the social patterning of functional ability trajectories. In the first three age groups, education was significantly associated with grouped trajectory membership, while in the oldest age group the association was neither statistically significant nor discernible once composition is referenced to the age-group marginal. This pattern suggests that education differentiates functional trajectories from midlife through the transition into older age, while in the oldest cohort the wider spread of mortality and functional limitation across educational groups, combined with a marginal distribution in which two thirds of respondents have primary education or less, makes the association less statistically distinguishable. The educational gradient was especially evident in declining trajectories. Respondents with primary education or less were more concentrated in progressive- and equivocal-declining trajectories across most age groups, while respondents with secondary education or more were more represented in mostly non-declining trajectories in the two youngest age groups.

These findings are consistent with international evidence showing that education is associated with disability, physical functioning, and later-life health trajectories [7–13]. Education can be interpreted as a life-course marker of accumulated social, material, and cognitive resources. In the Chilean context, where older cohorts experienced unequal access to schooling and where health and social protection resources remain stratified, educational attainment is a relevant marker for understanding inequality in functional aging.

These results can also be read against the contextual expectations set out at the outset. The concentration of women and of respondents with primary education or less in equivocal-declining trajectories is consistent with a setting in which recovery from an episode of limitation depends substantially on household resources, given a segmented health system and limited formal long-term care provision [21]. The strengthening of the educational gradient in progressive-declining trajectories with baseline age is consistent with cumulative exposure to physically demanding and unprotected work among lower-educated individuals in a labor market characterized by weak employment protection and persistent informality [19].

At the same time, the educational composition of the analytic sample shifts markedly across baseline age groups, rising from 44.3% with primary education or less in the youngest group to 67.0% in the oldest, so that educational differences observed here operate alongside a rapidly changing cohort composition and cannot be read as a pure period or age effect [20]. The heterogeneity of the trajectories themselves, encompassing sustained stability, ordered decline, and fluctuating patterns, resembles that reported in higher-income settings, which suggests that the distinctiveness of the Chilean case lies in how social groups are distributed across these pathways rather than in the existence of the pathways themselves.

The results also invite a careful interpretation of functional independence in later life. Cross-sectional estimates in Chile suggest that a large proportion of older adults live without functional dependence [21]. The longitudinal findings presented here add a different perspective. A person observed without functional limitations at one point in time may later remain stable, experience progressive decline, move through fluctuating states, or die after a period of preserved functioning. Cross-sectional estimates of functional independence therefore describe the prevalence of independence at a given moment while leaving sustained independence and transitional independence within a longer trajectory indistinguishable. This distinction is relevant for interpreting the demand for care and support because longitudinal trajectories reveal movements between functional states that cross-sectional measures cannot fully capture.

Limitations

Some limitations should be acknowledged. First, although the 17-year observation window is a major strength, the intervals between survey waves may miss short-term transitions in functional ability. More frequent observations would allow for a more detailed understanding of the timing, duration, and reversibility of functional changes [24]. This is particularly relevant for equivocal-declining trajectories, where shorter-term movements between limitation states may be partially unobserved.

Second, the inclusion criteria required respondents to have no more than one missing functional ability observation. This strategy supports robust sequence reconstruction, but it may exclude individuals with more unstable survey participation. The analytic sample is therefore defined by observability across the six waves and reflects the cumulative retention of a long panel. Its composition is broadly comparable to that of the wider eligible sample and, if anything, slightly less advantaged in educational and functional terms, which is reassuring for a study concerned with functional decline. Nonetheless, individuals lost to follow-up for reasons correlated with unobserved health deterioration cannot be recovered, and some selection bias cannot be ruled out.

Third, the final observation point requires a specific caveat. Fieldwork for the 2020 wave was interrupted by the COVID-19 pandemic and completed under a mixed-mode design, with part of the sample interviewed by telephone. Mode effects on the self-report of difficulties in activities of daily living cannot be ruled out, and the reduced achieved sample may have introduced additional selectivity at the endpoint of the observation window.

Three features of our design limit the leverage of this issue on our conclusions. First, trajectories are constructed from whole sequences of six observations, so that any single wave exerts bounded influence on cluster assignment. Second, the functional ability measure is a coarse hierarchical classification that is less sensitive to fine-grained reporting differences than a continuous score. And third, the sensitivity analysis based on the number of reported limitations reproduced the broad trajectory structure. Results concerning the final period should nonetheless be interpreted with this caveat in mind.

Fourth, analyses were conducted without survey weights, for the reasons stated in the analytical strategy. The prevalence figures reported for each trajectory type therefore describe the analytic sample and should not be read as population estimates for Chilean adults aged 46–65 in 2004. The compositional comparisons that support our substantive claims are internal to each age group and are less affected by this decision than prevalence estimation would be.

Fifth, functional ability was classified using a hierarchical index based on self-reported limitations in selected activities of daily living. This approach follows prior research on functional ability trajectories in Chile [4], but it does not incorporate clinical scales such as the Barthel or Katz Index. This may limit direct comparability with studies using different measurement instruments [35]. At the same time, the use of repeated self-reported functional limitations allows the analysis to reconstruct long-term intra-individual trajectories in a nationally representative panel.

Sixth, a further consideration concerns comparability across baseline age groups. Within each age group, the pooled estimation ensures that sex and educational groups are compared within shared trajectory categories. Across age groups, however, the three broad trajectory types represent an interpretive harmonization of age-group-specific cluster solutions rather than a jointly estimated typology. Comparisons across baseline age groups should therefore be read as descriptive rather than as contrasts between strictly equivalent constructs.

Seventh, educational attainment was categorized as primary education or less versus secondary education or more. This binary classification is useful for maintaining sufficient sample size within age groups, but it may obscure differences within educational strata, especially among respondents with tertiary education. Future research could examine whether more detailed educational categories modify the distribution of individuals across mostly non-declining, progressive-declining, and equivocal-declining trajectories.

Finally, the study focused on sex and education as two central dimensions of social inequality. Other dimensions, including income, occupation, ethnicity, region, caregiving histories, household composition, and access to health services, may also shape functional ability trajectories. Future research could examine how these factors combine across the life course and whether they help explain why some groups are more concentrated in progressive or equivocal forms of decline.

Conclusion

This study examined long-term functional ability trajectories among Chilean adults observed from midlife into older age. By estimating trajectory typologies within pooled age-group samples, the analysis provides a methodologically comparable account of how functional ability unfolds across socially differentiated populations. The findings show that functional aging is organized through multiple temporal patterns of continuity, deterioration, instability, and mortality-related transition. Mostly non-declining, progressive-declining, and equivocal-declining trajectories capture these different forms across the life course.

By focusing on Chile, this study extends functional ability trajectory research to a Latin American context characterized by rapid population aging and marked social inequality. The results contribute to previous Chilean evidence on functional ability trajectories and show that sex and education remain relevant dimensions for understanding how later-life functioning develops over time. They also suggest that aging research benefits from examining the composition and internal meaning of trajectory types, and from referencing that composition to the aggregate composition of the population under study, rather than treating broad categories such as stability or decline as equivalent across age groups.

The ways aging is described shape how aging is understood [36]. The trajectories documented here offer a longitudinal description of aging as heterogeneous and socially patterned. Functional ability changes through diverse pathways, and these pathways are connected to social position across the life course. Recognizing this heterogeneity allows research on aging to move beyond static contrasts between dependence and independence and toward a more dynamic understanding of functional ability, inequality, and later-life health.

Declaration of Generative AI and AI-assisted technologies in the writing process: We acknowledge the use of artificial intelligence to improve grammar, clarity, and the flow of writing.

Supporting information

S1 Appendix. Sensitivity Analysis 1.

(DOCX)

pone.0361826.s001.docx (226KB, docx)
S2 Table. Comparison of key variables in the study sample, a sample with missing observations, and a sample excluding deceased individuals.

(DOCX)

pone.0361826.s002.docx (30.9KB, docx)
S3 Appendix. Sensitivity Analysis 2.

(DOCX)

pone.0361826.s003.docx (471.7KB, docx)

Data Availability

The secondary longitudinal data used in this study is publicly available in the following repository of the Office of Social Security, Ministry of Labor and Social Security, Government of Chile: https://previsionsocial.gob.cl/datos-estadisticos/descargar-bases-de-datos-eps/.

Funding Statement

This work was supported by the following grants of the Agencia Nacional de Investigación y Desarrollo (ANID), Chile: ANID/ FONDECYT/ REGULAR N°1260169; ANID/ Millennium Science Initiative Program/ ICS2019_024. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.World Health Organization. Global report on ageism. Geneva: World Health Organization. 2021. [Google Scholar]
  • 2.Institute of Medicine. The Dynamics of Disability. Measuring and Monitoring Disability for Social Programs. Washington, DC: National Academy Press. 2002. [PubMed] [Google Scholar]
  • 3.Engelman M, Jackson H. Gradual Change, Homeostasis, and Punctuated Equilibrium: Reconsidering Patterns of Health in Later Life. Demography. 2019;56(6):2323–47. doi: 10.1007/s13524-019-00826-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Madero-Cabib I, Villalobos Dintrans P, Browne Salas J. Extending the Analysis of Functional Ability Trajectories to Unexplored National Contexts: The Case of Chile. J Gerontol B Psychol Sci Soc Sci. 2022;77(7):1280–93. doi: 10.1093/geronb/gbab116 [DOI] [PubMed] [Google Scholar]
  • 5.Oksuzyan A, Juel K, Vaupel JW, Christensen K. Men: good health and high mortality. Sex differences in health and aging. Aging Clin Exp Res. 2008;20(2):91–102. doi: 10.1007/BF03324754 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Crimmins EM, Kim JK, Solé-Auró A. Gender differences in health: results from SHARE, ELSA and HRS. Eur J Public Health. 2011;21(1):81–91. doi: 10.1093/eurpub/ckq022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Chiu C-J, Wray LA. Physical disability trajectories in older Americans with and without diabetes: the role of age, gender, race or ethnicity, and education. Gerontologist. 2011;51(1):51–63. doi: 10.1093/geront/gnq069 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Verbrugge LM, Latham K, Clarke PJ. Aging With Disability for Midlife and Older Adults. Res Aging. 2017;39(6):741–77. doi: 10.1177/0164027516681051 [DOI] [PubMed] [Google Scholar]
  • 9.Amemiya A, Kondo N, Saito J, Saito M, Takagi D, Haseda M, et al. Socioeconomic status and improvement in functional ability among older adults in Japan: a longitudinal study. BMC Public Health. 2019;19(1):209. doi: 10.1186/s12889-019-6531-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Serper M, Patzer RE, Curtis LM, Smith SG, O’Conor R, Baker DW, et al. Health literacy, cognitive ability, and functional health status among older adults. Health Serv Res. 2014;49(4):1249–67. doi: 10.1111/1475-6773.12154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Awuviry-Newton K, Amponsah M, Amoah D, Dintrans PV, Afram AA, Byles J, et al. Physical activity and functional disability among older adults in Ghana: The moderating role of multi-morbidity. PLOS Glob Public Health. 2023;3(3):e0001014. doi: 10.1371/journal.pgph.0001014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Kingston A, Davies K, Collerton J, Robinson L, Duncan R, Kirkwood TBL, et al. The enduring effect of education-socioeconomic differences in disability trajectories from age 85 years in the Newcastle 85+ Study. Arch Gerontol Geriatr. 2015;60(3):405–11. doi: 10.1016/j.archger.2015.02.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lai ETC, Ho HC, Ho SC, Woo J. Socioeconomic Status, Physical Functioning and Mortality: Results From a Cohort Study of Older Adults in Hong Kong. J Am Med Dir Assoc. 2022;23(5):858-864.e5. doi: 10.1016/j.jamda.2021.08.034 [DOI] [PubMed] [Google Scholar]
  • 14.Gill TM, Han L, Gahbauer EA, Leo-Summers L, Murphy TE. Risk Factors and Precipitants of Severe Disability Among Community-Living Older Persons. JAMA Netw Open. 2020;3(6):e206021. doi: 10.1001/jamanetworkopen.2020.6021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Yu H-W, Chen D-R, Chiang T-L, Tu Y-K, Chen Y-M. Disability trajectories and associated disablement process factors among older adults in Taiwan. Arch Gerontol Geriatr. 2015;60(2):272–80. doi: 10.1016/j.archger.2014.12.005 [DOI] [PubMed] [Google Scholar]
  • 16.Pan C, Kelifa MO, Liang J, Wang P. Joint trajectories of disability and related factors among older adults in China. Public Health. 2021;199:96–102. doi: 10.1016/j.puhe.2021.08.018 [DOI] [PubMed] [Google Scholar]
  • 17.Zimmer Z, Martin LG, Nagin DS, Jones BL. Modeling disability trajectories and mortality of the oldest-old in China. Demography. 2012;49(1):291–314. doi: 10.1007/s13524-011-0075-7 [DOI] [PubMed] [Google Scholar]
  • 18.Ide K, Tsuji T, Kanamori S, Jeong S, Nagamine Y, Kondo K. Social Participation and Functional Decline: A Comparative Study of Rural and Urban Older People, Using Japan Gerontological Evaluation Study Longitudinal Data. Int J Environ Res Public Health. 2020;17(2):617. doi: 10.3390/ijerph17020617 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Cabib I, Valencia R, Olea-Durán B. Childhood Socioeconomic Position, Adulthood Employment Trajectories, and Later Life Functional Limitations in Chile. J Appl Gerontol. 2023;42(3):474–86. doi: 10.1177/07334648221137055 [DOI] [PubMed] [Google Scholar]
  • 20.World Bank. World Development Indicators. 2024. https://databank.worldbank.org/source/world-development-indicators [Google Scholar]
  • 21.Villalobos Dintrans P. Panorama de la dependencia en Chile: avances y desafíos. Rev méd Chile. 2019;147(1):83–90. doi: 10.4067/s0034-98872019000100083 [DOI] [PubMed] [Google Scholar]
  • 22.Cabib I, Olea-Durán B, Villalobos Dintrans P, Browne Salas J. Long-term functional ability trajectories and mental health among older people before and after the COVID-19 pandemic onset in Chile. Aging Ment Health. 2023;27(8):1534–43. doi: 10.1080/13607863.2023.2188170 [DOI] [PubMed] [Google Scholar]
  • 23.Subsecretaría de Previsión Social. Bases de datos EPS. Santiago, Chile: Subsecretaría de Previsión Social. https://previsionsocial.gob.cl/datos-estadisticos/descargar-bases-de-datos-eps/ [Google Scholar]
  • 24.Wolf DA, Gill TM. Modeling transition rates using panel current-status data: how serious is the bias?. Demography. 2009;46(2):371–86. doi: 10.1353/dem.0.0057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gonzalez-Bautista E, Andrieu S, Gutiérrez-Robledo LM, García-Chanes RE, de Souto Barreto P. In the quest of a Standard Index of Intrinsic Capacity. A Critical Literature Review. J Nutr Health Aging. 2020;24(9):959–65. doi: 10.1007/s12603-020-1394-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Heidari S, Babor TF, De Castro P, Tort S, Curno M. Sex and Gender Equity in Research: rationale for the SAGER guidelines and recommended use. Res Integr Peer Rev. 2016;1:2. doi: 10.1186/s41073-016-0007-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Abbott A. Sequence Analysis: New Methods for Old Ideas. Annu Rev Sociol. 1995;21(1):93–113. doi: 10.1146/annurev.so.21.080195.000521 [DOI] [Google Scholar]
  • 28.Baumann I, Madero-Cabib I. Retirement Trajectories in Countries with Flexible Retirement Policies but Different Welfare Regimes. J Aging Soc Policy. 2021;33(2):138–60. doi: 10.1080/08959420.2019.1685358 [DOI] [PubMed] [Google Scholar]
  • 29.Gabadinho A, Ritschard G, Müller NS, Studer M. Analyzing and Visualizing State Sequences in R with TraMineR. J Stat Soft. 2011;40(4). doi: 10.18637/jss.v040.i04 [DOI] [Google Scholar]
  • 30.R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. 2023. https://www.R-project.org/ [Google Scholar]
  • 31.van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations inR. J Stat Soft. 2011;45(3). doi: 10.18637/jss.v045.i03 [DOI] [Google Scholar]
  • 32.Heckman J, Pinto R, Savelyev P. Understanding the Mechanisms Through Which an Influential Early Childhood Program Boosted Adult Outcomes. Am Econ Rev. 2013;103(6):2052–86. doi: 10.1257/aer.103.6.2052 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Villalobos Dintrans P, Izquierdo C, Guzmán R, Gálvez MJ, Santander S. Defining “older people” in Chile: challenges in planning policies for ageing populations. Health Policy Plan. 2021;35(10):1347–53. doi: 10.1093/heapol/czaa113 [DOI] [PubMed] [Google Scholar]
  • 34.Villalobos Dintrans P, Fabiani deLeva B. Challenges in reducing long-term care needs: insights from China. Lancet Healthy Longev. 2025;6(4):100705. doi: 10.1016/j.lanhl.2025.100705 [DOI] [PubMed] [Google Scholar]
  • 35.Echeverría A, Astorga C, Fernández C, Salgado M, Villalobos Dintrans P. Funcionalidad y personas mayores: ¿dónde estamos y hacia dónde ir?. Revista Panamericana de Salud Pública. 2023;46:e34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Nussbaum MC. Aging and control in King Lear—and the danger of generalization. In: Nussbaum MC, Levmore S, editors. Aging thoughtfully: Conversations about retirement, romance, wrinkles, & regret. Oxford: Oxford University Press. 2017. p. 8–23. [Google Scholar]

Decision Letter 0

Oyelola A Adegboye

6 Jun 2026

Dear Dr. Villalobos Dintrans,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

==============================

ACADEMIC EDITOR:

==============================

Please submit your revised manuscript by Jul 21 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

  • A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

As the corresponding author, your ORCID iD is verified in the submission system and will appear in the published article. PLOS supports the use of ORCID, and we encourage all coauthors to register for an ORCID iD and use it as well. Please encourage your coauthors to verify their ORCID iD within the submission system before final acceptance, as unverified ORCID iDs will not appear in the published article. Only  the individual author can complete the verification step; PLOS staff cannot  verify ORCID iDs on behalf of authors.

We look forward to receiving your revised manuscript.

Kind regards,

Oyelola A. Adegboye, PhD

Academic Editor

PLOS One

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please note that PLOS One has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, we expect all author-generated code to be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

3. Thank you for stating the following financial disclosure:

“This work was supported by the following grants of the Agencia Nacional de Investigación y Desarrollo (ANID), Chile: ANID/ FONDECYT/ REGULAR N°1220080; ANID/ FONDAP Nº15130009; ANID/ Millennium Science Initiative Program ICS2019_024.”

Please state what role the funders took in the study.  If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

If this statement is not correct you must amend it as needed.

Please include this amended Role of Funder statement in your cover letter; we will change the online submission form on your behalf.

4. In the online submission form you indicate that your data is not available for proprietary reasons and have provided a contact point for accessing this data. Please note that your current contact point is a co-author on this manuscript. According to our Data Policy, the contact point must not be an author on the manuscript and must be an institutional contact, ideally not an individual. Please revise your data statement to a non-author institutional point of contact, such as a data access or ethics committee, and send this to us via return email. Please also include contact information for the third party organization, and please include the full citation of where the data can be found.

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

6. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

**********

Reviewer #1: This manuscript uses a rich 17-year Chilean panel dataset and applies sequence analysis in a technically manner to explore functional ability trajectories. The use of optimal matching, multiple clustering quality criteria, and clear visualizations represent important methodological strengths. The topic is relevant for life-course and aging research, and the dataset provides a opportunity to examine long-term functional dynamics in an understudied context.

However, several important conceptual and methodological issues limit the validity of the manuscript’s central claims.

The study performs a technically sound and sophisticated sequence analysis; however, the way trajectory clusters are constructed separately within each stratified subgroup (by sex, education, and age) makes the resulting trajectories non-comparable across groups, which undermines the central comparative claims of the manuscript.

The introduction lacks a clearly articulated research gap and does not sufficiently position the study within the existing literature on functional ability trajectories and sequence analysis methodology. Forever, the theoretical background and anchoring the present research in the current state of empirical results of other studies are somewhat weak. I should consider identifying, reviewing Relevant literature on the analyzed topic and strengthen this part of the paper.

Building methodological section, it should be provided the source of the data as the reference in the paper. Also, the reason the author decided to analyze should be explained: they thought we jar very old from 2004 to 2010.

By generating different cluster solutions within each stratified subgroup, the study creates group-specific trajectory types and then compares their prevalence as if they were equivalent constructs, which constitutes a methodological inconsistency.

When constructing the text, the ideas should be consolidated in Francis and paragraphs, not on short sentences and paragraph with 1 1/2 or two lines. Moreover, should be avoided structuring the text in very short bullet points, like on lines 208-212, 222-225, 231-234,

Considering above consideration, the paper should undergo a major revision before being considered for publication.

**********

what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review?  For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

PLoS One. 2026 Nov 3;21(11):e0361826. doi: 10.1371/journal.pone.0361826.r002

Author response to Decision Letter 1


7 Jul 2026

Response Letter

Dear Academic Editor and Reviewer,

We sincerely thank you for the careful and constructive evaluation of our manuscript, now titled “Sex and Educational Inequalities Across Functional Ability Trajectories in Chile: A Sequence Analysis of a 17-Year National Panel”. We are grateful for the opportunity to revise and resubmit the manuscript.

We have undertaken a major revision in response to the comments. The most important change concerns the analytical strategy. In the previous version, trajectory clusters were estimated separately within sex and educational subgroups, which limited the comparability of the central claims.

In the revised manuscript, trajectory typologies are now estimated within each baseline age group using the full pooled age-group-specific sample, and sex and educational differences are examined within these shared trajectory categories. This revised strategy is stated in the Abstract, Materials and Methods, Results, Discussion, and Conclusion.

We also rewrote the Introduction to articulate the research gaps more clearly, strengthened the positioning of the study within the functional ability trajectory literature, expanded the description of the data source and observation period, clarified the handling of missing data and sensitivity analyses, reorganized the Results section around three comparable trajectory types, and revised the Discussion and Conclusion to better distinguish empirical description from interpretation.

Below, we provide a point-by-point response to each comment.

Yours sincerely,

The authors

Academic Editor

1. The manuscript would be strengthened by clearer reporting of the study characteristics and extent of missing data, ideally in Table 1 or a supplementary table, so that readers can better assess the amount of information addressed through imputation.

Response: We thank the Academic Editor for this suggestion. We revised the Materials and Methods section to provide clearer information about the study characteristics, sample selection, and missing-data criteria. The revised manuscript now states that the study used six observation points from the Chilean Social Protection Survey covering the period 2004–2020, and that the analytic sample included 2,681 individuals aged 46–65 at baseline, contributing 16,086 person-wave observations.

We also clarified that respondents were retained if they had no more than one missing functional ability observation across the six waves. For respondents with one missing functional ability observation, missing states were imputed using multiple imputation by chained equations with predictive mean matching, incorporating age, sex, educational attainment, and adjacent functional ability states as predictors.

These revisions appear in the Materials and Methods section, especially in the subsection “Data source and study design”, where we now describe the data source, observation period, analytic sample, age-group structure, missing-data criterion, and imputation procedure in a more explicit way. The revised Table 1 also reports descriptive characteristics of the analytic sample by baseline age group, including sex, educational attainment, baseline functional ability status, and functional ability status in 2020. In addition, Supplementary Table 1 reports sensitivity analyses comparing the study sample with alternative samples, including the broader sample with missing observations and the sample excluding deceased individuals.

2. In addition, a sensitivity analysis examining the impact of the imputation approach would be valuable, as this would help demonstrate the robustness of the findings to assumptions made about missing data.

Response: We thank the Academic Editor for this recommendation. To directly assess the robustness of our findings to the imputation strategy, we conducted an additional sensitivity analysis in which we re-estimated the sequence and cluster analysis on the subsample of individuals with fully complete functional ability information across all six waves, excluding all imputed observations. Because each complete-case individual carries both an original (imputed-sample) and a complete-case trajectory assignment, we cross-classified the two at the individual level and computed the percentage agreement, Cohen’s κ, and the adjusted Rand index for each baseline age group. The results are reported in Supplementary Table 2, and the complete-case typology is shown in Supplementary Figure 1.

In the two younger age groups, the two solutions assigned 100% of individuals to the same broad trajectory type (Cohen’s κ = 1.00), indicating that the imputation of a single missing wave did not alter the typology. In the two older age groups, concordance was lower (agreement of 60.6% and 75.9%). We interpret this difference cautiously. Because mortality is high in these cohorts, the complete-case restriction removed a large share of the sample, and the cluster structure was therefore re-estimated on a substantially smaller and differently composed subsample. The reduced concordance thus reflects the combined effect of sample-size reduction and cluster re-estimation rather than the imputation procedure alone. Importantly, the same three broad trajectory types remained identifiable in all four age groups (Supplementary Figure 1).

In addition to this analysis, and to further probe the robustness of the trajectory typology, we conducted a second sensitivity analysis in which functional ability was operationalized using the number of reported limitations instead of the hierarchical state variable. Cross-classifying each individual's broad-type membership across the two operationalizations showed very high concordance in the younger age groups (agreement of 96.0% and 98.0%; Cohen’s κ = 0.93 and 0.96) and lower concordance in the oldest group, consistent with the higher concentration of limitation and mortality states in that cohort. The broad three-type structure and the age-graded shift from sustained stability toward death-related trajectories were nonetheless preserved across operationalizations. This analysis is reported in Supplementary Table 3 and Supplementary Figure 2.

We have added a description of both sensitivity analyses to the Materials and Methods section (Data source and study design, and Functional ability measure subsections, respectively) and report the full results in the revised Supplementary Material.

Reviewer

1. This manuscript uses a rich 17-year Chilean panel dataset and applies sequence analysis in a technically manner to explore functional ability trajectories. The use of optimal matching, multiple clustering quality criteria, and clear visualizations represent important methodological strengths. The topic is relevant for life-course and aging research, and the dataset provides a opportunity to examine long-term functional dynamics in an understudied context. However, several important conceptual and methodological issues limit the validity of the manuscript’s central claims.

Response: We sincerely thank the Reviewer for recognizing the strengths of the dataset, the relevance of the topic, and the methodological value of the sequence-analysis approach. We also appreciate the Reviewer’s identification of conceptual and methodological issues that limited the previous version.

In response, we substantially revised the manuscript. The revised version now includes a clearer research gap, stronger positioning within the literature on functional ability trajectories, a corrected analytical strategy based on pooled age-group-specific trajectory typologies, a rewritten Results section organized around comparable trajectory types, and a revised Discussion that interprets the findings in relation to functional aging, life-course inequality, and methodological comparability.

These changes are reflected throughout the manuscript. The Abstract has been rewritten to summarize the revised analytical strategy and main findings. The Introduction now clarifies the research gaps and contribution. The Materials and Methods section explains the pooled age-group-specific clustering strategy and the treatment of missing data. The Results section is reorganized around three main trajectory types, and the Discussion and Conclusion now interpret the findings while emphasizing the methodological contribution of comparing sex and educational groups within shared trajectory categories.

2. The study performs a technically sound and sophisticated sequence analysis; however, the way trajectory clusters are constructed separately within each stratified subgroup (by sex, education, and age) makes the resulting trajectories non-comparable across groups, which undermines the central comparative claims of the manuscript.

Response: We agree with this comment. This was the central methodological limitation of the previous version. In the original manuscript, clusters were estimated separately within sex and educational subgroups, which meant that trajectories with similar labels were not necessarily equivalent empirical constructs across groups.

We have corrected this issue in the revised manuscript. Trajectory clusters are now estimated within each baseline age group using the full pooled age-group-specific sample. After creating this common typology for each age group, we compare the sex and educational composition of the resulting shared trajectory types. This ensures that comparisons across women and men, and across educational groups, refer to common trajectory categories within each baseline age group.

This revised strategy is introduced in the Abstract and developed in the Materials and Methods section, especially in the subsections “Sequence analysis and trajectory classification” and “Analytical strategy”. The Results section then applies this strategy by presenting trajectory types estimated within pooled age-group samples and by comparing sex and educational composition within those shared types. The Discussion and Conclusion explicitly interpret this methodological change as a way to provide a comparable account of sex and educational differences in long-term functional ability trajectories.

We have also applied the same comparability logic to clarify the status of comparisons made across baseline age groups. Within each age group, the pooled estimation guarantees that sex and educational groups are compared within statistically shared trajectory categories. Across age groups, however, the grouping of the age-group-specific cluster solutions into three broad types (mostly non-declining, progressive-declining, and equivocal-declining) is an interpretive harmonization based on the shape, ordering, and timing of the detailed trajectories, rather than a single jointly estimated cross-age cluster solution. We now state this explicitly so that within-age comparisons (statistically grounded) and cross-age comparisons (descriptive) are not conflated. This clarification is reflected in the Results section, where the three broad types are introduced as a harmonization of the detailed age-group-specific trajectories, and in the Limitations subsection of the Discussion, where we note that comparisons across baseline age groups should be read as descriptive rather than as contrasts between strictly equivalent constructs.

3. The introduction lacks a clearly articulated research gap and does not sufficiently position the study within the existing literature on functional ability trajectories and sequence analysis methodology. Forever, the theoretical background and anchoring the present research in the current state of empirical results of other studies are somewhat weak. I should consider identifying, reviewing Relevant literature on the analyzed topic and strengthen this part of the paper.

Response: We have substantially rewritten the Introduction to address this concern. The revised Introduction now identifies two explicit research gaps. First, most evidence on long-term functional ability trajectories comes from the United States, Europe, and parts of Asia, while evidence from Latin America remains more limited. Second, the revised Introduction explains that comparisons across sex and educational groups require trajectory typologies constructed in a way that allows valid comparison across groups.

We also strengthened the theoretical and empirical framing. The revised Introduction discusses the shift from static disability measures toward longitudinal functional ability trajectories, the relevance of life-course and bridging approaches, previous international evidence on disability and functional ability trajectories, and prior Chilean evidence using the EPS. The revised version also clarifies why education is relevant as a marker of cumulative life-course resources.

These changes appear throughout the revised Introduction. The opening paragraphs now frame functional ability as a dynamic process and situate the study within longitudinal disability and functional ability research. The subsection “Research gaps” explicitly identifies the empirical and methodological gaps addressed by the study. The final paragraphs of the Introduction then clarify how the present study extends previous Chilean research and state the research question guiding the analysis.

4. Building methodological section, it should be provided the source of the data as the reference in the paper. Also, the reason the author decided to analyze should be explained: they thought we jar very old from 2004 to 2010.

Response: We revised the Materials and Methods section to provide a clearer description of the data source and the rationale for the analytic period and baseline age range. The revised manuscript now identifies the Encuesta de Protección Social (Social Protection Survey, EPS) as a nationally representative Chilean longitudinal survey and includes a reference to the official EPS data source.

We also explain why the 2004–2020 observation window was used. The period covers all the surveys available to date. This period provides six observation points with functional ability information and allows us to examine functional ability changes over 17 years, from midlife into older age. We further clarify why individuals aged 46–65 in 2004 were selected: this baseline age range allows the analysis to begin before most respondents reach older ages while still observing transitions into later life. Respondents were then divided into four five-year baseline age groups: 46–50, 51–55, 56–60, and 61–65.

These revisions are concentrated in the Materials and Methods section, particularly in the subsection “Data source and study design”. In that subsection, we now identify the EPS as the data source, cite the official EPS data reference, justify the 2004–2020 observation window, explain the selection of individuals aged 46–65 at baseline, and describe the rationale for analyzing four five-year baseline age groups separately.

5. By generating different cluster solutions within each stratified subgroup, the study creates group-specific trajectory types and then compares their prevalence as if they were equivalent constructs, which constitutes a methodological inconsistency.

Response: We agree with the Reviewer. This methodological inconsistency has been corrected. In the revised manuscript, we no longer generate sex-specific or education-specific cluster solutions. Instead, for each baseline age group, we estimate one pooled trajectory typology using the full age-group-specific sample. We then compare the sex and educational composition of the resulting shared trajectory types.

This revision changes the interpretation of the Results. The manuscript no longer compares the prevalence of non-equivalent clusters across separately stratified subgroup solutions. It now compares the distribution of sex and educational groups within common trajectory categories.

This correction is described in the Materials and Methods section, where we explain the pooled age-group-specific clustering strategy. The Results section then follows this revised design by presenting trajectory types defined before sex and education are compared. The subsections “Sex differences in trajectory types across age groups” and “Educational differences in trajectory types across age groups” report the composition of shared trajectory types by sex and education. Tables 2 and 3 support this analysis by reporting the sex and educational composition of main and detailed trajectory types within each age group. The Discussion explains why this revised strategy strengthens the validity of the comparative claims.

We would like to clarify one remaining aspect of comparability that follows from the Reviewer’s observation. The pooled estimation resolves the comparability problem for sex and educational groups within each baseline age group, because these groups are now compared within statistically equivalent trajectory categories. The grouping of the detailed, age-group-specific clusters into three broad types, however, is an interpretive harmonization rather than a jointly estimated typology spanning all age groups. We have made this distinction explicit to avoid implying that the underlying detailed clusters are statistically equivalent across age groups. Accordingly, the Results section now presents the three broad types as an interpretive grouping of the age-group-specific trajectories, intended to support cross-age description, and the Limitations subsection notes that comparisons across baseline age groups are descriptive in nature, while comparisons by sex and education within each age group rest on the shared, pooled trajectory categories.

6. When constructing the text, the ideas should be consolidated in Francis and paragraphs, not on short sentences and paragraph with 1 1/2 or two lines. Moreover, should be avoided structuring the text in very short bullet points, like on lines 208-212, 222-225, 231-234,

Response: We revised the manuscript’s writing style to address this comment. The previous Results section included short bullet-style descriptions of trajectory subtypes. In the revised manuscript, these descriptions have been rewritten as continuous prose. We also consolidated short paragraphs throughout the Introduction, Materials and Methods, Results, and Discussion sections to improve flow and readability.

This revision is especially visible in the Results section. The subsection “Functional ability trajectories across age groups” now describes the detailed trajectories in integrated paragraphs rather than short bullet points. The sex and education subsections also present group differences in narrative prose. The Discussion has been reorganized into interpretive paragraphs that synthesize the findings and reduce unnecessary repetition with the Results section.

7. Considering above consideration, the paper should undergo a major revision before being considered for publication.

Response: We agree that a major revision was necessary. The manuscript has been substantially revised in response to the Reviewer’s comments. The revised version includes a corrected analytical strategy, a clearer research gap, a stronger methodological explanation, updated figures and tables, a rewritten Results section based on pooled age-group-specific trajectory typologies, and a revised Discussion and Conclusion.

These changes affect the manuscript as a whole. The revision corrects the methodological inconsistency identified by the Reviewer, strengthens the Introduction and Methods, reorganizes the Results around comparable trajectory categories, and clarifies the interpretation of sex and educational differences in the Discussion and Conclusion. We believe these changes directly address the Reviewer’s concerns and substantially improve the manuscript’s methodological validity, clarity, and contribution.

Attachment

Submitted filename: Response to Reviewers (2).docx

pone.0361826.s005.docx (32.6KB, docx)

Decision Letter 1

Oyelola A Adegboye

28 Aug 2026

Dear Dr. Villalobos Dintrans,

Thank you for submitting your manuscript to PLOS One. After careful consideration, we feel that it has merit but does not fully meet PLOS One’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Oct 12 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

  • A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS One offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

As the corresponding author, your ORCID iD is verified in the submission system and will appear in the published article. PLOS supports the use of ORCID, and we encourage all coauthors to register for an ORCID iD and use it as well. Please encourage your coauthors to verify their ORCID iD within the submission system before final acceptance, as unverified ORCID iDs will not appear in the published article. Only  the individual author can complete the verification step; PLOS staff cannot  verify ORCID iDs on behalf of authors.

We look forward to receiving your revised manuscript.

Kind regards,

Oyelola A. Adegboye, PhD

Academic Editor

PLOS One

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #2: Yes

**********

Reviewer #2: I thank the authors for their very interesting contribution, emphasising the dynamic nature of functional ability in Chile across time, and the potential heterogeneity of this across age, sex and education levels. I have the following comments on the different sections.

- Introduction:

o It is stated that previous research on high-income countries has shown that individual and social factors shape functional decline, but how this operates is not made explicit (i.e., what do we already know from these contexts?)

- Research gaps:

o It is argued that a main gap is that evidence so far comes mainly from the US, Europe and Asia, and that evidence from Latin America is scarce, even though the region faces rapid population ageing, high inequality and different institutional arrangements. I think the argument could be pushed a little further to say why these particularities make this research gap relevant (e.g., what should we expect to be different from/similar to what we already know from previous research?). This could also be further re-emphasised within the discussion or conclusion sections.

- Materials and methods:

o EPS waves selection: It is stated that the period 2004-2020 covers all EPS available waves. However, there is also a 2002 wave and a 2024 wave. I assume the former wasn’t used as it is not representative of the adult population but only of those affiliated to the pension system (or perhaps because of the lack of any relevant variable?) and the latter as it was probably not yet available when the analysis was undertaken, but it is worth mentioning whether this is the case.

o EPS waves limitations: Some additional issues with the EPS data might be acknowledged (and/or reflect on whether they might have any implications for the analysis). For instance, EPS 2012 data were published with some reservations due to low response rates during fieldwork and the institution responsible for the survey suggests caution when using them for statistical inference. In addition, fieldwork for the EPS 2020 wave had to be suspended during mid-March 2020 due to Covid-19 lockdown measures, resulting in a smaller sample size, and although the survey was continued from September to December of that year, it changed mode to a phone survey.

o Table 1 description: When describing Table 1, the acronym FA is used (in line 156). I assume it refers to functional ability, but I think it is not previously defined in the text. FA is also used in line 225 when describing the sequence analysis.

o Analytical sample: The analytical sample is composed of 2,681 individuals aged 45-64 in 2004 across the 6 EPS waves. Is this sample representative of the population in that age group? (i.e., how might the attrition of the survey be affecting this, given that the analyses may miss people who are only present in some of the waves?).

o Sampling weights: Does the analysis use any survey sampling weights? This should be stated, and the reason for the choice justified.

- Results and discussion:

o When presenting/discussing the findings of the sex and educational differences in trajectory types across age groups, it will be interesting to relate them to the respective prevalent differences in sex and educational categories by each group and whether or not it is consistent. For example, the mostly non-declining trajectories had a higher proportion of men in age groups 1, 3, and 4 and the progressive-declining trajectories had a higher proportion of women in those same age groups. Given that age groups 1 and 2 overall have more women than men and age groups 3 and 4 more men than women, the disaggregation by trajectory types may be “reverting” or “reinforcing” some of this aggregate composition. The same applies to the education findings discussion.

**********

what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review?  For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

PLoS One. 2026 Nov 3;21(11):e0361826. doi: 10.1371/journal.pone.0361826.r004

Author response to Decision Letter 2


29 Sep 2026

Response Letter

Dear Academic Editor and Reviewer,

We sincerely thank you for the opportunity to revise and resubmit our manuscript, “Sex and Educational Inequalities Across Functional Ability Trajectories in Chile: A Sequence Analysis of a 17-Year National Panel”. We are particularly grateful to the Reviewer for a set of comments that were detailed, constructive, and closely informed by the data source we use. Several of them identified issues that we had not documented adequately and that we have now addressed throughout the manuscript.

The most important change concerns the interpretation of our central findings. Following the Reviewer’s eighth comment, we now reference the sex and educational composition of every trajectory type to the marginal composition of its baseline age group, both in Tables 2 and 3 and in the Results and Discussion. This has materially improved the analysis, as it allows us to distinguish compositional differences that reverse the aggregate distribution of an age group from those that reproduce it, and it revealed that one of our previous claims required qualification. The apparent concentration of respondents with primary education or less in equivocal-declining trajectories in the oldest age group (65.7%) in fact falls slightly below that age group’s marginal (67.0%), and all three trajectory types in that age group track the marginal educational distribution within 7.2 percentage points, which now provides a substantive account of why the chi-squared test is non-significant there.

Beyond this, we have made the following changes. The Introduction now specifies the mechanisms through which individual and social factors shape functional decline in high-income settings. The Research gaps subsection articulates why the Chilean configuration of rapid demographic aging, recent educational expansion, a historically deregulated labor market, and a segmented health and long-term care system leads to specific expectations about how functional trajectories unfold, and the Discussion returns to these expectations. The Materials and Methods section now documents the selection of EPS waves explicitly, including the reasons for excluding the 2002, 2012, and 2024 waves, and acknowledges the disruption of the 2020 fieldwork. We have clarified the representativeness of the analytic sample, stated and justified the decision not to apply survey weights, and removed an undefined acronym. Six new references have been added and the reference list has been renumbered accordingly.

Below we respond to each comment in turn, indicating where the corresponding changes appear.

Yours sincerely,

The authors.

REVIEWER

Comment 1. It is stated that previous research on high-income countries has shown that individual and social factors shape functional decline, but how this operates is not made explicit (i.e., what do we already know from these contexts?)

Response: We thank the Reviewer for this observation. The previous version listed the factors associated with functional decline without specifying how they operate. We have rewritten the third paragraph of the Introduction to describe the mechanisms documented in high-income settings, distinguishing sex-related, socioeconomic, and clinical pathways.

The revised paragraph specifies the configuration underlying sex differences, namely a higher burden of non-fatal disabling conditions, differences in muscle mass and strength, and lifetime differences in paid work and caregiving responsibilities, alongside longer survival among women. It then specifies the intermediate channels through which educational attainment operates, including exposure to physically demanding and hazardous occupations, health literacy and the capacity to navigate health systems, income and material conditions in later life, and health behaviors. It also clarifies the dual role of chronic conditions, multimorbidity, and sedentary behavior, which act both as precipitating events that trigger transitions into limitation and as constraints on the likelihood of recovering function once limitation has occurred, so that advantaged groups more frequently display recovery and less frequently remain in unstable states. We further note that these mechanisms accumulate across the life course, with early-life circumstances and adulthood employment histories shaping functional status decades later, and that the strength and direction of specific associations varies across age groups and cohorts.

Two references on sex differences in health and survival and one on childhood socioeconomic position and later-life functional limitations in Chile have been added to support these mechanisms.

Comment 2. It is argued that a main gap is that evidence so far comes mainly from the US, Europe and Asia, and that evidence from Latin America is scarce, even though the region faces rapid population ageing, high inequality and different institutional arrangements. I think the argument could be pushed a little further to say why these particularities make this research gap relevant (e.g., what should we expect to be different from/similar to what we already know from previous research?). This could also be further re-emphasised within the discussion or conclusion sections.

Response: We agree that the argument required further development. The previous version stated that the region differs from the settings where most evidence has been produced without specifying what those differences imply for the phenomenon under study. We have added a new paragraph to the Research Gaps subsection that articulates why the Chilean configuration matters and what it leads us to expect.

Specifically, we now discuss four features and their implications. The compressed pace of demographic aging means that successive cohorts reach later life with markedly different educational and occupational endowments. Rapid educational expansion produces steep cohort gradients in educational attainment, which entangles educational differences with cohort composition and shifts the social meaning of a given educational threshold across baseline age groups. A labor market characterized by weak employment protection and persistent informality concentrates cumulative occupational exposure among lower-educated workers, which plausibly produces steeper educational gradients than in settings with stronger labor regulation. A segmented health system and limited long-term care provision make recovery from an episode of limitation dependent on household resources, which suggests that socioeconomic differences will be expressed in the stability and reversibility of trajectories as much as in their onset. We also state which features we expect to resemble those documented elsewhere, namely the basic heterogeneity of functional pathways.

Following the Reviewer's suggestion, we return to this argument in the Discussion with a dedicated paragraph that assesses our findings against these expectations. First, we note that the concentration of women and of lower-educated respondents in equivocal-declining trajectories is consistent with recovery being conditional on household resources. Second, we observe that the strengthening of the educational gradient in progressive-declining trajectories with baseline age is consistent with cumulative occupational exposure. Finally, we argue that the heterogeneity of the trajectories themselves resembles that reported in higher-income settings, so that the distinctiveness of the Chilean case lies in how social groups are distributed across these pathways rather than in the existence of the pathways themselves. The Conclusion has been adjusted accordingly.

Comment 3. EPS waves selection: It is stated that the period 2004-2020 covers all EPS available waves. However, there is also a 2002 wave and a 2024 wave. I assume the former wasn’t used as it is not representative of the adult population but only of those affiliated to the pension system (or perhaps because of the lack of any relevant variable?) and the latter as it was probably not yet available when the analysis was undertaken, but it is worth mentioning whether this is the case.

Response: We thank the Reviewer for noting this, and we confirm that both assumptions are correct. The statement that the study used all available EPS waves was inaccurate and has been removed.

The revised Data source and study design subsection now lists all waves fielded by the EPS (2002, 2004, 2006, 2009, 2012, 2015, 2018, 2020, and 2024) and explains the selection of the analytic window. The 2002 wave was excluded for the two reasons the Reviewer anticipates, that is, questions on functional ability status were incorporated into the survey only from the 2004 version onwards, and the 2002 sample was restricted to individuals affiliated to the pension system, which limits its representativeness of the adult population. The 2024 wave had not been released at the time the analyses were conducted, and we now state that it will be incorporated into future extensions of this work. Additionally, the 2012 wave was also excluded, for reasons of data quality that we detail in our response to Comment 4 below.

Comment 4. EPS waves limitations: Some additional issues with the EPS data might be acknowledged (and/or reflect on whether they might have any implications for the analysis). For instance, EPS 2012 data were published with some reservations due to low response rates during fieldwork and the institution responsible for the survey suggests caution when using them for statistical inference. In addition, fieldwork for the EPS 2020 wave had to be suspended during mid-March 2020 due to Covid-19 lockdown measures, resulting in a smaller sample size, and although the survey was continued from September to December of that year, it changed mode to a phone survey.

Response: We are grateful for this observation, which led us to document two data-quality issues that the previous version did not acknowledge.

The revised Data source and study design subsection now includes a dedicated paragraph stating that the 2012 wave was published with reservations by the institution responsible for the survey, due to low response rates during fieldwork and an accompanying recommendation that these data be used with caution for statistical inference, and that it was therefore excluded from the analytic window. The same paragraph describes the disruption of the 2020 fieldwork, which began face-to-face in late 2019, was suspended in March 2020 following the introduction of nationwide lockdown measures, and resumed between September and December of that year by telephone, yielding a smaller achieved sample and a change in survey mode for part of the wave.

We have also added a new limitation (now the third) reflecting on the implications for our analysis. We acknowledge that mode effects on the self-report of difficulties in activities of daily living cannot be ruled out, and that the reduced achieved sample may have introduced additional selectivity at the endpoint of the observation window.

We note three features of our design that limit the leverage of this issue on our conclusions. First, trajectories are constructed from whole sequences of six observations, so that any single wave exerts bounded influence on cluster assignment. Second, the functional ability measure is a coarse hierarchical classification that is less sensitive to fine-grained reporting differences than a continuous score. And third, the sensitivity analysis based on the number of reported limitations reproduced the broad trajectory structure. We state that results concerning the final period should nonetheless be interpreted with this caveat in mind.

Comment 5. Table 1 description: When describing Table 1, the acronym FA is used (in line 156). I assume it refers to functional ability, but I think it is not previously defined in the text. FA is also used in line 225 when describing the sequence analysis.

Response: We thank the Reviewer for identifying this inconsistency. The acronym was indeed used without prior definition. Since it appeared only twice, we have replaced both occurrences with the full term “functional ability” rather than introducing an abbreviation, which improves readability.

The affected passages are the description of Table 1 in the Data source and study design subsection and the description of the cluster selection procedure in the Sequence analysis and trajectory classification subsection.

Comment 6. Analytical sample: The analytical sample is composed of 2,681 individuals aged 45-64 in 2004 across the 6 EPS waves. Is this sample representative of the population in that age group? (i.e., how might the attrition of the survey be affecting this, given that the analyses may miss people who are only present in some of the waves?).

Response: We thank the Reviewer for raising this point, which we have addressed in two ways.

First, we now state explicitly that the analytic sample is defined by observability of functional ability status across the six waves, and is therefore representative of respondents who remained in the panel rather than of the Chilean population aged 46–65 in 2004.

Second, we have expanded the discussion of Supplementary Table 1, which compares the analytic sample with the broader sample that includes individuals with missing observations. The differences are modest and consistent in direction across age groups. The analytic sample contains between 3.4 and 5.0 percentage points more respondents with primary education or less across the four age groups, a somewhat higher proportion of women in the three younger age groups, and a marginally lower baseline prevalence of no limitations in three of the four age groups. Retention therefore does not appear to have selected the most advantaged or the healthiest respondents, which is reassuring for a study concerned with functional decline. We also note that retaining individuals who died during the observation period, and treating death as a functional ability state, limits one of the principal sources of selective loss in long panels of aging populations.

The second limitation has been expanded accordingly, acknowledging that individuals lost to follow-up for reasons correlated with unobserved health deterioration cannot be recovered and that some selection bias cannot be ruled out.

Comment 7. Sampling weights: Does the analysis use any survey sampling weights? This should be stated, and the reason for the choice justified.

Response: We thank the Reviewer for this request. The analyses are unweighted, and the previous version neither stated nor justified this decision. We have now made it explicit in a new paragraph of the Analytical Strategy subsection and provided three reasons.

First, the construction of the typology rests on pairwise dissimilarities between sequences and on hierarchical clustering, procedures into which expansion factors cannot be incorporated in a straightforward manner at the stage where trajectory types are defined. Second, our substantive claims concern the sex and educational composition of shared trajectory categories within each baseline age group, a within-sample comparison for which weighting is less consequential than it would be for national prevalence estimation. Third, previous research using the same survey and a comparable trajectory design reported highly similar results with and without the expansion factor, which is consistent with an accurate sampling strategy and adequate fieldwork in the EPS.

We have also added a new limitation (now the fourth) stating that, because analyses are unweighted, the prevalence figures reported for each trajectory type describe the analytic sample and should not be read as population estimates, while noting that the compositional comparisons supporting our substantive claims are internal to each age group.

Comment 8. When presenting/discussing the findings of the sex and educational differences in trajectory types across age groups, it will be interesting to relate them to the respective prevalent differences in sex and educational categories by each group and whether or not it is consistent. For example, the mostly non-declining trajectories had a higher proportion of men in age groups 1, 3, and 4 and the progressive-declining trajectories had a higher proportion of women in those same age groups. Given that age groups 1 and 2 overall have more women than men and age groups 3 and 4 more men than women, the disaggregation by trajectory types may be “reverting” or “reinforcing” some of this aggregate composition. The same applies to the education findings discussion.

Response: We are grateful for this suggestion, which substantially improved the interpretation of our findings and led to the most extensive revision in this round. The previous version reported the composition of each trajectory type without anchoring it to the overall composition of the corresponding baseline age group, so that readers could not assess whether a given composition departed from or reproduced the aggregate distribution.

We have addressed this in four ways.

First, Tables 2 and 3 now include a first row reporting the marginal sex and educational distribution of each baseline age group, so that over- and under-representation can be read directly from the tables. The table notes explain this addition.

Second, the sex subsection now opens with an explicit statement of the marginals and of the convention we adopt, following the Reviewer’s own terms, that is, a departure is described as reversing the aggregate composition when the majority within a trajectory type differs from the majority in the age group, and as reinforcing it when the two coincide. Every composition is then expressed as a departure in percentage points. The male majority in mostly non-declining trajectories reverses the aggregate composition in age groups 1 and 2 (departures of 6.2 and 5.6 percentage points from female-majority marginals) and reinforces it in age groups 3 and 4 (4.5 and 8.6 percentage points). The female majority in equivocal-declining trajectories reinforces the aggregate composition in age groups 1 and 2 (10.2 and 8.0 percentage points) and reverses it in age group 4, where it constitutes the largest departure in the table at 18.4 percentage points. In age group 3 this type closely reproduces a nearly balanced marginal. Among progressive-declining trajectories, three of the four departures reverse the aggregate composition.

Third, the educational subsection follows the same logic. Mostly non-declining trajectories fall below the marginal proportion with primary education or less in all four age groups, by 4.2, 5.8, 4.7, and 3.1 percentage points. Progressive-declining trajectories move from 3.8 percentage points below the marginal in age group 1 to 3.9, 5.5, and 7.2 percentage points above it in age groups 2, 3, and 4, so that the educational gradient strengthens with baseline age relative to the composition of each age group and not only in absolute terms. Equivocal-declining trajectories exceed the marginal by 8.9, 5.8, and 10.7 percentage points in the three younger age groups.

A qualification of our earlier claim. Applying this framework revealed that one statement in the previous version required correction. We had described equivocal-declining trajectories as predominantly composed of lower-educated respondents in all four age groups, citing 65.7% in age group 4. Referenced to that age group’s marginal of 67.0%, this composition in fact falls 1.3 percentage points below the aggregate distribution and therefore does not indicate over-representation. We now state this explicitly, and we note that all three trajectory types in age group 4 fall within 7.2 percentage points of the marginal educational distribution, which provides a substantive interpretation of the non-significant chi-squared test in that age group. The Abstract has been amended so that the educational claim is restricted to the three younger age groups.

The Discussion now includes a dedicated paragraph drawing these threads together, and the Conclusion notes that aging research benefits from referencing trajectory composition to the aggregate composition of the population under study.

We thank again the Reviewer for giving us the opportunity to improve our study.

Attachment

Submitted filename: Response_to_Reviewers_(2)_auresp_2.docx

pone.0361826.s006.docx (32.8KB, docx)

Decision Letter 2

Oyelola A Adegboye

16 Oct 2026

Sex and Educational Inequalities Across Functional Ability Trajectories in Chile: A Sequence Analysis of a 17-Year National Panel

PONE-D-26-00167R2

Dear Dr. Villalobos Dintrans,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Oyelola A. Adegboye, PhD

Academic Editor

PLOS One

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #2: Yes

**********

Reviewer #2: (No Response)

**********

what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review?  For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #2: Yes:  Joaquín Mayorga Camus

**********

Acceptance letter

Oyelola A Adegboye

PONE-D-26-00167R2

PLOS One

Dear Dr. Dintrans,

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS One. Congratulations! Your manuscript is now being handed over to our production team.

At this stage, our production department will prepare your paper for publication. This includes ensuring the following:

* All references, tables, and figures are properly cited

* All relevant supporting information is included in the manuscript submission,

* There are no issues that prevent the paper from being properly typeset

You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps.

Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing.

If we can help with anything else, please email us at customercare@plos.org.

Thank you for submitting your work to PLOS One and supporting open access.

Kind regards,

PLOS One Editorial Office Staff

on behalf of

Assoc Prof Oyelola A. Adegboye

Academic Editor

PLOS One

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Appendix. Sensitivity Analysis 1.

    (DOCX)

    pone.0361826.s001.docx (226KB, docx)
    S2 Table. Comparison of key variables in the study sample, a sample with missing observations, and a sample excluding deceased individuals.

    (DOCX)

    pone.0361826.s002.docx (30.9KB, docx)
    S3 Appendix. Sensitivity Analysis 2.

    (DOCX)

    pone.0361826.s003.docx (471.7KB, docx)
    Attachment

    Submitted filename: Response to Reviewers (2).docx

    pone.0361826.s005.docx (32.6KB, docx)
    Attachment

    Submitted filename: Response_to_Reviewers_(2)_auresp_2.docx

    pone.0361826.s006.docx (32.8KB, docx)

    Data Availability Statement

    The secondary longitudinal data used in this study is publicly available in the following repository of the Office of Social Security, Ministry of Labor and Social Security, Government of Chile: https://previsionsocial.gob.cl/datos-estadisticos/descargar-bases-de-datos-eps/.


    Articles from PLOS One are provided here courtesy of PLOS

    RESOURCES