Individual Longitudinal Trajectories Research Articles

Extending intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) to study individual longitudinal trajectories, with application to mental health in the UK

The intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) approach is gaining prominence in health sciences and beyond, as a robust quantitative method for identifying intersectional inequalities in a range of individual outcomes. However, it has so far not been applied to longitudinal data, despite the availability of such data, and growing recognition that intersectional social processes and determinants are not static, unchanging phenomena. Drawing on intersectionality and life course theories, we develop a longitudinal version of the intersectional MAIHDA approach, allowing the analysis not just of intersectional inequalities in static individual differences, but also of life course trajectories. We discuss the conceptualization of intersectional groups in this context: how they are changeable over the life course, appropriate treatment of generational differences, and relevance of the age-period-cohort identification problem. We illustrate the approach with a study of mental health using United Kingdom Household Longitudinal Study data (2009–2021). The results reveal important differences in trajectories between generations and intersectional strata, and show that trajectories are partly multiplicative but mostly additive in their intersectional inequalities. This article provides an important and much needed methodological contribution, enabling rigorous quantitative, longitudinal, intersectional analyses in social epidemiology and beyond.

OR25-01 Longitudinal Changes in the Hypothalamic-Pituitary-Testicular (HPT) Axis in Men with Longevity

Abstract Disclosure: D. He: None. T. Gao: None. N. Barzilai: None. S. Milman: None. S. Aleksic: None. Background: Aging in men is associated with a reduction in testosterone (T) levels. In men aged 50 - 80, the decline in T mainly occurs due to hypothalamic-pituitary dysfunction, which results from unhealthy aging characterized by obesity and metabolic disorders. The changes in the HPT axis in men with longevity are less understood. Our recent cross-sectional study in men aged 90 and older showed high prevalence of testicular dysfunction which was not associated with metabolic health. We now explore the longitudinal changes in the HPT axis in men with longevity to better understand their relationship with metabolic health and survival. Methods: In an observational study of community-dwelling Ashkenazi Jewish men (n = 68, mean age (SD) at baseline sample 84 ± 4 years), levels of total T (TT, LC/MS), LH, and SHBG were measured in longitudinal serum samples (a median of 3 measures per subject, age at final sample 89 ± 3 years). Men were followed for survival after the final sample. Free T (FT) was calculated per Vermeulen et al. Men treated with T or GnRH agonists were excluded. Metabolic measures included BMI, A1c, and serum lipids. Individual longitudinal trajectories of TT, FT, LH, and SHBG were estimated with linear regression. Biochemical gonadal phenotypes were determined based on the levels of TT and LH. Results: Baseline and final levels of TT, FT, LH, and SHBG were, respectively: 417 ± 173 and 346 ± 187 ng/dL (TT), 6.0 ± 2.4 and 4.7 ± 2.5 ng/dL (FT), 8.0 [5.1, 14.4] and 10.0 [5.9, 20.4] mIU/mL (LH), and 58 ± 20 and 63 ± 23 nmol/L (SHBG), with average change rate/year of −17 ± 47 ng/dL (TT), −0.3 ± 0.5 ng/dL (FT), 0.7 ± 2.2 mIU/mL (LH) and 0.9 ± 4.1 nmol/L (SHBG). At baseline, the most prevalent biochemical gonadal phenotype was eugonadism (54%), followed by compensated testicular dysfunction (24%), testicular dysfunction (10%), and hypothalamic-pituitary dysfunction (12%), while respective prevalence rates at the final sample were 40%, 24%, 29%, and 7%. Baseline TT (p=0.002), age (p=0.02), and history of cancer (p=0.01) predicted more negative trajectory, while the history of cardiovascular disease (p=0.02) predicted more positive trajectory of TT. Neither TT, FT, LH nor SHBG trajectories were significant predictors of final BMI, A1c, or serum lipid levels. Positive TT trajectory was associated with shorter survival, after adjustment for SHBG, age, and comorbidities (HR 6.61 [1.46, 29.8], p=0.01). Conclusions: This is one of the first longitudinal studies of the HPT axis in community-dwelling men with longevity. As men survive into the late 80s, they increasingly develop testicular dysfunction while hypothalamic-pituitary dysfunction is rare. Changes in testosterone levels do not seem to predict metabolic dysfunction; however, an increase in total testosterone with aging may be associated with reduced survival. These findings need to be explored further in larger and ethnically diverse cohorts. Presentation Date: Saturday, June 17, 2023

Developmental Trajectories in Very Preterm Born Children Up to 8 Years: A Longitudinal Cohort Study.

Aim: Long-term outcome data in preterm children is often limited to cross-sectional measurement of neurodevelopmental impairment (NDI) at the corrected age of 24-36 months. However, impairments may only become overt during childhood or resolve with time, and individual trajectories in outcome over time may vary. The primary aim of this study was to describe NDI in very preterm born children at three subsequent ages of 2, 5, and 8 years of age. As a secondary aim, a longitudinal analysis was performed on the individual longitudinal trajectories in NDI from 2 to 8 years of age.Methods: Single-center prospective cohort study including children born between 1990 and 2011 below 30 weeks' gestation and followed into 2019. The outcome measurement was NDI assessed at 2, 5, and 8 years of age. NDI is a composite score that includes cognitive, neurological, visual, and auditory functions, in which problems were categorized as none, mild, moderate, or severe. Cognitive function measured as total DQ/IQ score was assessed by standardized psychometric tests. Neurological, visual, and auditory functions were assessed by the neonatologist.Results: In total, 921 children were eligible for follow-up, of whom 726 (79%) children were assessed. No NDI was seen in 54, 54, and 62%, mild NDI was seen in 31, 36, and 30%, and moderate-to-severe NDI was seen in 15, 9.2, and 8.6% of the children at 2, 5, and 8 years, respectively. From 2 to 8 years, 63% of the children remained in the same NDI category, 20% of the children improved to a better NDI category, and 17% deteriorated toward a worse NDI category. No differences were found in baseline characteristics of infants that improved or deteriorated. Extreme prematurity, male gender and low parental education were associated with worse NDI status at all time points. Although we observed considerable individual variation over time in NDI status, the course of the trajectories in NDI were not associated with gestation, gender, and parental education.Conclusions: Continued follow-up until school life is essential in order to provide optimal and individually focused referrals and care when needed.

Five-year trajectories of multimorbidity patterns in an elderly Mediterranean population using Hidden Markov Models

This study aimed to analyse the trajectories and mortality of multimorbidity patterns in patients aged 65 to 99 years in Catalonia (Spain). Five year (2012–2016) data of 916,619 participants from a primary care, population-based electronic health record database (Information System for Research in Primary Care, SIDIAP) were included in this retrospective cohort study. Individual longitudinal trajectories were modelled with a Hidden Markov Model across multimorbidity patterns. We computed the mortality hazard using Cox regression models to estimate survival in multimorbidity patterns. Ten multimorbidity patterns were originally identified and two more states (death and drop-outs) were subsequently added. At baseline, the most frequent cluster was the Non-Specific Pattern (42%), and the least frequent the Multisystem Pattern (1.6%). Most participants stayed in the same cluster over the 5 year follow-up period, from 92.1% in the Nervous, Musculoskeletal pattern to 59.2% in the Cardio-Circulatory and Renal pattern. The highest mortality rates were observed for patterns that included cardio-circulatory diseases: Cardio-Circulatory and Renal (37.1%); Nervous, Digestive and Circulatory (31.8%); and Cardio-Circulatory, Mental, Respiratory and Genitourinary (28.8%). This study demonstrates the feasibility of characterizing multimorbidity patterns along time. Multimorbidity trajectories were generally stable, although changes in specific multimorbidity patterns were observed. The Hidden Markov Model is useful for modelling transitions across multimorbidity patterns and mortality risk. Our findings suggest that health interventions targeting specific multimorbidity patterns may reduce mortality in patients with multimorbidity.

Emotion regulation habits related to depression: A longitudinal investigation of stability and change in repetitive negative thinking and positive reappraisal

BackgroundThe propensity to use more repetitive negative thinking and less positive reappraisal is theorized to play an important role in depression. To date, little is known about the presumed enduring nature of these emotion regulation habits. Therefore, this study examined individual longitudinal trajectories and within-person sources of change in the habitual use of repetitive negative thinking and positive reappraisal. MethodsParticipants (N=320) completed twenty waves of data collection separated by one-week time intervals. At each wave, participants completed measures of habitual repetitive negative thinking, habitual positive reappraisal, perceived stress, and depression symptoms. ResultsGrowth curve modeling revealed that the prototypical longitudinal trajectories of habitual repetitive negative thinking and positive reappraisal were relatively stable over time. Yet, substantial variation occurred around the prototypical change trajectories. Perceived stress and depressive symptom severity emerged as within-person factors modulating concurrent levels of emotion regulation habits. Moreover, within-person cross-lagged analysis revealed that changes in repetitive negative thinking both predicted and were predicted by changes in perceived stress. No such within-person reciprocal relations were found for positive reappraisal. LimitationsThis study covered only a period of five months, used self-report measures, and was conducted in a community sample. Furthermore, 25% of the participants did not complete all 20 waves of data collection. ConclusionsThese findings help to understand the enduring trait-like processes and within-person factors that may contribute to the longitudinal course of habitual repetitive negative thinking and positive reappraisal.

A hierarchical temporal attention-based LSTM encoder-decoder model for individual mobility prediction

Prediction of individual mobility is crucial in human mobility related applications. Whereas, existing research on individual mobility prediction mainly focuses on next location prediction and short-term dependencies between traveling locations. Long-term location sequence prediction is of great importance for long-time traffic planning and location advertising, and long-term dependencies exist as individual mobility regularity typically occurs daily and weekly. This paper proposes a novel hierarchical temporal attention-based LSTM encoder-decoder model for individual location sequence prediction. The proposed hierarchical attention mechanism captures both long-term and short-term dependencies underlying in individual longitudinal trajectories, and uncovers frequential and periodical mobility patterns in an interpretable manner by incorporating the calendar cycle of individual travel regularities into location prediction. More specifically, the hierarchical attention consists of local temporal attention to identify highly related locations in each day, and global temporal attention to discern important travel regularities over a week. Experiments on individual trajectory datasets with varying degree of traveling uncertainty demonstrate that our method outperforms four baseline methods on three evaluation metrics. In addition, we explore the interpretability of the proposed model in understanding individual daily, and weekly mobility patterns by visualizing the temporal attention weights and frequent traveling patterns associated with locations.

The impact of population dynamics on the population HIV care cascade: results from the ANRS 12249 Treatment as Prevention trial in rural KwaZulu-Natal (South Africa).

IntroductionThe universal test and treat strategy (UTT) was developed to maximize the proportion of all HIV‐positive individuals on antiretroviral treatment (ART) and virally suppressed, assuming that it will lead to a reduction in HIV incidence at the population level. The evolution over time of the cross‐sectional HIV care cascade is determined by individual longitudinal trajectories through the HIV care continuum and underlying population dynamics. The purpose of this paper is to quantify the contribution of each component of population change (in‐ and out‐migration, HIV seroconversion, ageing into the cohort and definitive exit such as death) on the HIV care cascade in the context of the ANRS 12249 Treatment as Prevention (TasP) cluster‐randomized trial, investigating UTT in rural KwaZulu‐Natal, South Africa, between 2012 and 2016.Methods HIV test results and information on clinic visits, ART prescriptions, viral load and CD4 count, migration and deaths were used to calculate residency status, HIV status and HIV care status for each individual on a daily basis. Position within the HIV care continuum was considered as a score ranging from 0 (undiagnosed) to 4 (virally suppressed). We compared the cascade score of each individual joining or leaving the population of resident adults living with HIV with the average score of their cluster at the time of entry or exit. Then, we computed the contribution of each entry or exit on the average cascade score and their annualized total contribution, by component of change.ResultsWhile the average cascade score increased over time in all clusters, that increase was constrained by population dynamics. Permanent exits and ageing into the people living with HIV cohort had a marginal effect. Both in‐migrants and out‐migrants were less likely to be retained at each step of the HIV care continuum. However, their overall impact on the cross‐sectional cascade was limited as the effect of in‐ and out‐migration balanced each other. The contribution of HIV seroconversions was negative in all clusters.ConclusionsIn a context of high HIV incidence, the continuous flow of newly infected individuals slows down the efforts to increase ART coverage and population viral suppression, ultimately attenuating any population‐level impact on HIV incidence.Clinical Trial Number NCT01509508 (clinicalTrials.gov)/DOH‐27‐0512‐3974 (South African National Clinical Trials Register).

Joint Analyses of Longitudinal and Time-to-Event Data in Research on Aging: Implications for Predicting Health and Survival.

Longitudinal data on aging, health, and longevity provide a wealth of information to investigate different aspects of the processes of aging and development of diseases leading to death. Statistical methods aimed at analyses of time-to-event data jointly with longitudinal measurements became known as the “joint models” (JM). An important point to consider in analyses of such data in the context of studies on aging, health, and longevity is how to incorporate knowledge and theories about mechanisms and regularities of aging-related changes that accumulate in the research field into respective analytic approaches. In the absence of specific observations of longitudinal dynamics of relevant biomarkers manifesting such mechanisms and regularities, traditional approaches have a rather limited utility to estimate respective parameters that can be meaningfully interpreted from the biological point of view. A conceptual analytic framework for these purposes, the stochastic process model of aging (SPM), has been recently developed in the biodemographic literature. It incorporates available knowledge about mechanisms of aging-related changes, which may be hidden in the individual longitudinal trajectories of physiological variables and this allows for analyzing their indirect impact on risks of diseases and death. Despite, essentially, serving similar purposes, JM and SPM developed in parallel in different disciplines with very limited cross-referencing. Although there were several publications separately reviewing these two approaches, there were no publications presenting both these approaches in some detail. Here, we overview both approaches jointly and provide some new modifications of SPM. We discuss the use of stochastic processes to capture biological variation and heterogeneity in longitudinal patterns and important and promising (but still largely underused) applications of JM and SPM to predictions of individual and population mortality and health-related outcomes.

Outcome Vector Dependent Sampling with Longitudinal Continuous Response Data: Stratified Sampling Based on Summary Statistics

The analysis of longitudinal trajectories usually focuses on evaluation of explanatory factors that are either associated with rates of change, or with overall mean levels of a continuous outcome variable. In this article, we introduce valid design and analysis methods that permit outcome dependent sampling of longitudinal data for scenarios where all outcome data currently exist, but a targeted substudy is being planned in order to collect additional key exposure information on a limited number of subjects. We propose a stratified sampling based on specific summaries of individual longitudinal trajectories, and we detail an ascertainment corrected maximum likelihood approach for estimation using the resulting biased sample of subjects. In addition, we demonstrate that the efficiency of an outcome-based sampling design relative to use of a simple random sample depends highly on the choice of outcome summary statistic used to direct sampling, and we show a natural link between the goals of the longitudinal regression model and corresponding desirable designs. Using data from the Childhood Asthma Management Program, where genetic information required retrospective ascertainment, we study a range of designs that examine lung function profiles over 4 years of follow-up for children classified according to their genotype for the IL 13 cytokine.

Modeling subject‐specific phase‐dependent effects and variations in longitudinal responses via a geometric Brownian motion process

We address statistical issues regarding modeling a collection of longitudinal response trajectories characterized by the presence of subject-specific phase-dependent effects and variation. To accommodate these two time-varying individual characteristics, we employ a geometric stochastic differential equation for modeling based on a Brownian motion process and develop a two-step paradigm for statistical analysis. This paradigm reverses the order of statistical inference in random effects model. We first extract individual information about phase-dependent treatment effects and volatility parameters for all subjects. Then, we derive the association relationship between the parameters characterizing the individual longitudinal trajectories and the corresponding covariates by means of multiple regression analysis. The stochastic differential equation model and the two-step paradigm together provide significant advantages both in modeling flexibility and in computational efficiency. The modeling flexibility is due to the easy adaptation of temporal change points for subject-specific phase transition in treatment effects, whereas the computational efficiency benefits in part from the independent increment property of Brownian motion that avoids high-dimensional integration. We demonstrate our modeling approach and statistical analysis on a real data set of longitudinal measurements of disease activity scores from a rheumatoid arthritis study.

Reproduction is adapted to survival characteristics across geographically isolated medfly populations

We propose the hypothesis that individual longitudinal trajectories of fertility are closely coupled to varying survival schedules across geographically isolated populations of the same species, in such a way that peak reproduction takes place before substantial increases in mortality are observed. This reproductive adaptation hypothesis is investigated for medflies through a statistical analysis of biodemographic data that were obtained for female medflies from six geographically far apart regions. The following results support the hypothesis: (i) both survival and reproductive schedules differ substantially between these populations, where early peaks and subsequently fast declining reproduction are observed for short-lived and protracted reproductive schedules for long-lived flies; (ii) when statistically adjusting reproduction for the observed differences in survival, the differences in reproductive schedules largely vanish, and thus the observed differences in fertility across the populations can be explained by differences in population-specific longevity; and (iii) specific survival patterns of the medflies belonging to a specific population predict the individual reproductive schedule for the flies in this population. The analysis is based on innovative statistical tools from functional data analysis. Our findings are consistent with an adaptive mechanism whereby trajectories of fertility evolve in response to specific constraints inherent in the population survival schedules.

Contribution of Residential Relocation and Lifestyle to the Structure of Health Trajectories

Underlining ecological theories of aging, we assessed the impact of relocation, residential type, and individual lifestyle factors on the structure of health status overtime. From the data of Longitudinal Study on Aging II, we included older adults aged 70 and older (N = 5,294). To analyze individual longitudinal trajectories of health outcomes, Latent Growth Curve Modeling (LGCM) was employed. LGCMs supported that older adults' residential relocation and health-related lifestyles were important in preserving better health outcomes. Multiple structural equations corroborated the causal chains in the multidimensionality of health structure. These findings suggest a necessity to design policies for older adults to create a synergy between housing and health care and to translate meaningful health-related lifestyles into diverse long-term care settings.