Latent Class Trajectory Modeling Research Articles

Change in Engagement in Cognitive Activity and Risk for Mild Cognitive Impairment in a Cohort of Older Adults: The Monongahela-Youghiogheny Healthy Aging Team (MYHAT) Study.

Engagement in cognitively stimulating activities is gaining prominence as a potential strategy to maintain cognitive functioning in old age. In a population-based cohort of individuals aged 65+ years, we examined patterns of change in frequency of engagement in total cognitive activity (TCA), higher cognitive activity (HCA), and frequent cognitive activity (FCA) based on the Florida Cognitive Activities Scale over an average of 3.62 years, and whether these patterns were associated with incident mild cognitive impairment (MCI) during this same period. Among 867 cognitively normal participants, 129 (15%) progressed to MCI. Latent class trajectory modeling identified high and stable, slowly, and quickly declining patterns for TCA; high and stable, slowly declining, and slowing increasing patterns for FCA; and high and stable, and slowly declining patterns for HCA. Separate, adjusted Cox proportional hazard models, revealed that compared with the high, stable pattern, both slow decline [hazard ratio (HR), 2.5; 95% confidence interval (CI), 1.5-4.0] and quick decline (HR, 11.0; 95% CI, 6.3-19.2) in TCA, and slow decline in the FCA (HR, 8.7; 95% CI, 5.3-14.3) and HCA (HR, 3.4; 95% CI, 2.0-5.6) subscales increased risk for incident MCI. Maintaining engagement in cognitive activities may be protective against progression to MCI, alternatively, declining engagement may be a marker for impending cognitive impairment.

12 Year Trajectories of Depressive Symptoms in Community-Dwelling Older Adults and the Subsequent Risk of Death Over 13 Years.

Populations of depressed persons are typically comprised of individuals with different courses of depression and thus might carry different risks of death. This study aimed to identify different trajectories of depressive symptoms in community-dwelling older adults and study the risk of death across these trajectories. In the population-based Rotterdam Study, depressive symptoms (Center for Epidemiological Studies-Depression scale) at three examination rounds (1993-2004) from 3,325 dementia-free participants (mean age 64.6 ± 6.1 years) were used to identify depression trajectories by latent-class trajectory modeling. Mortality rates by trajectory were calculated over a subsequent 13 year period (2002-2015), that is using 23 years of follow-up data. Five trajectories of depressive symptoms characterized by low (73.4%), decreasing (11.1%), remitting (5.1%), increasing (7.7%), and high (2.7%) depressive symptoms were identified. Compared with persons in the low symptoms trajectory, persons with a trajectory of increasing depressive symptoms (hazard ratio [HR]: 1.21 [95% CI = 1.02, 1.44]) had a higher risk of death, but not those with remitting depressive symptoms, HR: 1.06 (95% CI = 0.85, 1.32). The estimates for the high symptoms trajectory were also suggestive of a higher risk of mortality, HR: 1.20 (95% CI = 0.91, 1.58). Repeated measures of depression can help predict long-term health outcomes in persons with depressive symptoms. Participants with increasing symptoms over time had a higher risk of death than those with low or no depressive symptoms. Transient high depressive symptoms that remitted were not associated with a higher risk compared with those with no symptoms. Our results open avenues for etiological and prognostic research to focus upon risk factors' key to a particular trajectory.

Adiposity across the adult life course and incidence of primary liver cancer: The NIH-AARP cohort.

Obesity relatively late in adulthood has been consistently associated with increased risk of primary liver cancer. However, little is known about the role of early adult adiposity and evolution of adiposity across adulthood in hepatocarcinogenesis. We examined adult body mass index (BMI; kg/m2 ) in relation to hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) in a large prospective cohort. Weight at ages 18, 35, 50 and at study baseline was retrospectively reported by 303,620 participants. BMI trajectories were identified using latent class trajectory modeling. Incidence of HCC and ICC was determined through 2011. Cox proportional hazards modeling was used to calculate adjusted hazard ratios (HRs) and 95% confidence intervals (CIs). A total of 372 HCC cases and 104 ICC cases occurred during follow-up. Being obese (BMI ≥ 30) at ages 18, 35, 50 and at baseline (mean age 62.3 years, range 50.3-71.5 years) was associated with an 86-119% elevated risk of HCC. BMI trajectories that resulted in obesity were associated with ∼80% higher HCC incidence. BMI at age 18, per 5 kg/m2 , was associated with a 34% higher risk of ICC, but the association attenuated for BMI at older ages. In conclusion, our findings suggest that maintaining a healthy BMI throughout the lifetime may reduce liver cancer risk. Future studies with longitudinally collected weight information are warranted to further elucidate the role of life-course adiposity in liver cancer development.

Maternal stress and neonatal anthropometry: the NICHD Fetal Growth Studies

The effect of maternal mood disorders on neonatal measurements is not well-defined. The Fetal Growth Studies-Singletons provide a unique opportunity to evaluate the relationship between perceived maternal stress and neonatal growth measurements. The purpose of this study was to determine whether perceived maternal stress during pregnancy is associated with anthropometric measurements in the neonate. This analysis was based on a prospective, multicenter longitudinal study of fetal growth. Women 18-40 years old with a body mass index of 19.0-29.9 kg/m2 were screened at 8+0 to 13+6 weeks gestation for low-risk status associated with optimal fetal growth (eg, healthy, nonsmoking) and underwent serial sonographic examination at 6 study visits throughout gestation. At each study visit, women completed the Cohen's Perceived Stress Survey, which could have a score that ranges from 0-40. We used a latent class trajectory model to identify distinct groupings (ie, classes) of the Perceived Stress Survey trajectories over pregnancy. Trend analysis was used to determine whether neonatal measurements including birthweight, length, head circumference, and abdominal circumference differed by Perceived Stress Survey class and whether this relationship was modified by maternal race/ethnicity, after adjustment for gestational age at delivery, maternal height, age, and parity. Of the 2334 women enrolled in the study, 1948 women had complete neonatal anthropometry and were included in the analysis. Latent class analysis identified 3 Perceived Stress Survey trajectory classes, with mean Perceived Stress Survey scores of 2.82 (low), 7.95 (medium), and 14.80 (high). Neonatal anthropometric measures of birthweight, length, head circumference and abdominal circumference were similar (P=.78,=.10,=.18, and=.40 respectively), regardless of the participants' Perceived Stress Survey class. There was no effect modification by maternal race/ethnicity. Neonatal measurements did not differ by levels of perceived stress among low-risk pregnant women.

Trajectories of neighborhood attainment after prison

A potentially important but understudied aspect of prisoner reentry is the neighborhood environments experienced by formerly incarcerated people. We know that many formerly incarcerated people return to very disadvantaged neighborhood environments and that returning to disadvantaged neighborhoods after prison increases the risk of recidivism and reduces employment. Yet very little is known about the social, economic, and institutional processes that sort formerly incarcerated people into different neighborhoods after release or their trajectories of neighborhood attainment over time. Motivated by a conceptualization of prisoner reentry and reintegration as a process that unfolds over time, we examine trajectories of neighborhood environments after release. Motivated by the literature on neighborhood attainment, social capital, and the role of criminal justice institutions in structuring the lives of former prisoners, we examine sources of variation in neighborhood attainment. We use administrative data from the Michigan Department of Corrections on formerly incarcerated people paroled in 2003 and followed for two years after release. Descriptive results from a latent class trajectory model show that most white and black formerly incarcerated people experience flat trajectories, with little upward or downward residential mobility over time. Findings from multi-level growth curve models suggest that institutional factors are particularly important for the neighborhood attainment of whites, while human capital and social ties are particularly important for blacks. Among both blacks and whites, pre-prison and first post-prison neighborhood conditions exhibit a strong association with post-prison neighborhood attainment, although these associations are larger for blacks than whites.

Prediagnostic Body Mass Index Trajectories in Relation to Prostate Cancer Incidence and Mortality in the PLCO Cancer Screening Trial.

Evidence suggests that obesity in adulthood is associated with increased risk of "clinically significant" prostate cancer. However, studies of body mass index (BMI) across the adult life course and prostate cancer risks remain limited. In a prospective cohort of 69 873 men in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, we examined associations of prediagnostic BMI across the adult life course with risk of incident prostate cancer and fatal prostate cancer (prostate cancer-specific mortality). At 13 years of follow-up, we identified 7822 incident prostate cancer cases, of which 3078 were aggressive and 255 fatal. BMI trajectories were determined using latent-class trajectory modeling. Cox proportional hazards regression was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). BMI at age 20 years, 50 years, and baseline questionnaire (mean age = 63 years) were associated with increased risks of fatal prostate cancer (HRs = 1.27-1.32 per five-unit increase). In five BMI trajectories identified, fatal prostate cancer risk was increased in men who had a normal BMI (HR = 1.95, 95% CI = 1.21 to 3.12) or who were overweight (HR = 2.65, 95% CI = 1.35 to 5.18) at age 20 years and developed obesity by baseline compared with men who maintained a normal BMI. Aggressive and nonaggressive prostate cancer were not associated with BMI, and modest inverse associations were seen for total prostate cancer. Our results suggest that BMI trajectories during adulthood that result in obesity lead to an elevated risk of fatal prostate cancer.

10-year trajectories of depressive symptoms and risk of dementia: a population-based study

Late-life depressive symptoms have been extensively studied for their relationship with incident dementia, but have been typically assessed at a single timepoint. Such an approach neglects the course of depression, which, given its remitting and relapsing nature, might provide further insights into the complex association of depression with dementia. We therefore repeatedly measured depressive symptoms in a population of adults over a decade to study the subsequent risk of dementia. Our study was embedded in the Rotterdam Study, a population-based study of adults aged 55 years or older in Rotterdam (Netherlands), ongoing since 1990. The cohort is monitored continuously for major events by data linkage between the study database and general practitioners. We examined a cohort of participants who were free from dementia, but had data for depressive symptoms from at least one examination round in 1993-95, 1997-99, or 2002-04. We assessed depressive symptoms with the validated Dutch version of the Center for Epidemiology Depression Scale (CES-D) and the Hospital Anxiety and Depression Scale-Depression. We used these data to identify 11-year trajectories of depressive symptoms by latent class trajectory modelling. We screened participants for dementia at each examination round and followed up participants for 10 years for incident dementia by latent trajectory from the third examination round to 2014. We calculated hazard ratios (HR) for dementia by assigned trajectory using two Cox proportional hazards models (model 1 adjusted for age and sex only, and model 2 adjusted additionally for APOEɛ4 carrier status, educational level, body-mass index, smoking, alcohol consumption, cognitive score, use of antidepressants, and prevalent disease status at baseline). We repeated the analyses censoring for incident stroke, restricting to Alzheimer's disease as an outcome, and accounting for mortality as a competing risk for dementia. From 1993-2004, we obtained data for depressive symptoms from at least one examination round for 3325 participants (median age: 74·88 years [IQR 70·62-80·06], 1995 [60%] women). We identified five trajectories of depressive symptoms in these 3325 individuals, characterised by maintained low CES-D scores (low; 2441 [73%]); moderately high starting scores but then remitting (decreasing; 369 [11%]); low starting scores, increasing, then remitting (remitting; 170 [5%]); low starting scores that steadily increased (increasing; 255 [8%]); and maintained high scores (high; 90 [3%]). During 26 330 person-years, 434 participants developed incident dementia. Only the trajectory with increasing depressive symptoms was associated with a higher risk of dementia compared with the low depressive symptom trajectory, using model 2 (HR 1·42, 95% CI 1·05-1·94; p=0·024). Additionally, only the increasing trajectory was associated with a higher risk of dementia compared with the low trajectory after censoring for incident stroke (1·58, 1·15-2·16; p=0·0041), restricting to Alzheimer's disease as an outcome (1·44, 1·03-2·02; p=0·034), and accounting for mortality as a competing risk (1·45, 1·06-1·97; p=0·019). Risk of dementia differed with different courses of depression, which could not be captured by a single assessment of depressive symptoms. The higher risk of dementia only in the increasing trajectory suggests depression might be a prodrome of dementia. Erasmus Medical Center; ZonMw; the Netherlands Ministry of Education Culture and Science; and the Netherlands Ministry for Health, Welfare and Sports.

Disability Trajectories at the End of Life: A "Countdown" Model.

Studies of late-life disablement typically address the role of advancing age as a factor in developing disability, and in some cases have pointed out the importance of time to death (TTD) in understanding changes in functioning. However, few studies have addressed both factors simultaneously, and none have dealt satisfactorily with the problem of missing data on TTD in panel studies. We fit latent-class trajectory models of disablement using data from the Health and Retirement Study. Among survivors (~20% of the sample), TTD is unknown, producing a missing-data problem. We use an auxiliary regression equation to impute TTD and employ multiple imputation techniques to obtain final parameter estimates and standard errors. Our best-fitting model has 3 latent classes. In all 3 classes, the probability of having a disability increases with nearness to death; however, in only 2 of the 3 classes is age associated with disability. We find gender, race, and educational differences in class-membership probabilities. The model reveals a complex pattern of age- and time-dependent heterogeneity in late-life disablement. The techniques developed here could be applied to other phenomena known to depend on TTD, such as cognitive change, weight loss, and health care spending.

Characterizing long-term patterns of weight change in China using latent class trajectory modeling.

BackgroundOver the past three decades, obesity-related diseases have increased tremendously in China, and are now the leading causes of morbidity and mortality. Patterns of weight change can be used to predict risk of obesity-related diseases, increase understanding of etiology of disease risk, identify groups at particularly high risk, and shape prevention strategies.MethodsLatent class trajectory modeling was used to compute weight change trajectories for adults aged 18 to 66 using the China Health and Nutrition Survey (CHNS) data (n = 12,611). Weight change trajectories were computed separately for males and females by age group at baseline due to differential age-related patterns of weight gain in China with urbanization. Generalized linear mixed effects models examined the association between weight change trajectories and baseline characteristics including urbanicity, BMI category, age, and year of study entry.ResultsTrajectory classes were identified for each of six age-sex subgroups corresponding to various degrees of weight loss, maintenance and weight gain. Baseline BMI status was a significant predictor of trajectory membership for all age-sex subgroups. Baseline overweight/obesity increased odds of following ‘initial loss with maintenance’ trajectories. We found no significant association between baseline urbanization and trajectory membership after controlling for other covariates.ConclusionTrajectory analysis identified patterns of weight change for age by gender groups. Lack of association between baseline urbanization status and trajectory membership suggests that living in a rural environment at baseline was not protective. Analyses identified age-specific nuances in weight change patterns, pointing to the importance of subgroup analyses in future research.

Impact of variability in adherence to HIV antiretroviral therapy on the immunovirological response and mortality.

BackgroundSeveral previous studies have shown relationships between adherence to HIV antiretroviral therapy (ART) and the viral load, the CD4 cell count, or mortality. However, the impact of variability in adherence to ART on the immunovirological response does not seem to have been investigated yet.MethodsMonthly adherence data (November 1999 to April 2009) from 317 HIV-1 infected patients enrolled in the Senegalese ART initiative were analyzed. Latent-class trajectory models were used to build typical trajectories for the average adherence and the standardized variance of adherence. The relationship between the standardized variance of adherence and each of the change in CD4 cell count, the change in viral load, and mortality were investigated using, respectively, a mixed linear regression, a mixed logistic regression, and a Cox model with time-dependent covariates. All the models were adjusted on the average adherence.ResultsThree latent trajectories for the average adherence and three for the standardized variance of adherence were identified. The increase in CD4 cell count and the increase in the percentage of undetectable viral loads were negatively associated with the standardized variance of adherence but positively associated with the average adherence. The risk of death decreased significantly with the increase in the average adherence but increased significantly with the increase of the standardized variance of adherence.ConclusionsThe impacts of the level and the variability of adherence on the immunovirological response and survival justify the inclusion of these aspects into the process of patient education: adherence should be both high and constant.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2288-15-10) contains supplementary material, which is available to authorized users.

Characterizing Long-Term Patterns of Weight Change in China Using Latent Class Trajectory Modeling

Characterizing Long-Term Patterns of Weight Change in China Using Latent Class Trajectory Modeling

Modeling the seasonality of Anopheles gambiae s.s. biting rates in a South Benin sanitary zone

Efficient malaria vector control requires knowledge of spatio-temporal vector dynamics. We have classified village groups according to the biting rate profiles of both Anopheles coluzzii and An. gambiae, the major malaria vectors in these villages. Mosquitoes were captured by human bait in 28 South Benin villages during 2009. Both An. coluzzii and An. gambiae counts in each village were standardized to focus on changes in the vector biting rate over time. Latent class trajectory modeling, allowing for random intercept at the 'village' level, was adjusted to standardized values. The villages could be classified into two groups with distinct vector biting rate profiles (continuous/transient). This classification helped creating a map of vector biting rates in the area. The biting rate profiles were found to be significantly correlated with mean rainfall, altitude, average number of larval sites, and average normalized difference vegetation index. In highly malaria-prone regions, knowledge of vector biting rate profiles is important to improve vector control interventions. A similar methodology may be applied to study the biting rate profiles of other vector-borne infections.

Trajectories of health‐related quality of life in children with epilepsy: A cohort study

Little is known about subgroups of children with epilepsy who may experience less favorable outcomes over time. The objectives of this study were to document trajectories of health-related quality of life (HRQL) and to identify predictors of the trajectory group in children with new-onset epilepsy. Data were obtained from the Health Related Quality of Life in Children with Epilepsy Study, a prospective multisite study of children 4-12 years old with new-onset epilepsy followed for 24 months. Health-related quality of life was measured using the Quality of Life in Childhood Epilepsy questionnaire. Trajectories of HRQL were investigated using latent class trajectory modeling. Multinomial logistic regression was used to identify child, parent, and family predictors of HRQL trajectories. A total of 374 families responded at baseline and 283 (76%) completed the study. Five HRQL trajectories were observed: low-increasing (4%), moderate-decreasing (12%), moderate-increasing (22%), high-increasing (32%), and high-stable (30%). Many children in the low-increasing, moderate-increasing, high-increasing, and high-stable had clinically meaningful improvements in HRQL: 82%, 47%, 63%, and 44%, respectively. In contrast, the majority of children in the moderate-decreasing group (56%) experienced clinically meaningful declines in their HRQL. Factors predicting trajectories were number of antiepileptic drugs prescribed, presence of comorbid behavior or cognitive problems, parent depression, and family functioning and demands. Results suggested that children with epilepsy are not homogenous but rather consist of groups with different trajectories and unique predictors of HRQL. Problems associated with child behavior and cognition were the strongest predictors identified. Given that several risk factors are modifiable, it is important to examine these as potential targets within a family-centered framework to improve HRQL of children with new-onset epilepsy.

Obesity and Mortality Risk: New Findings From Body Mass Index Trajectories

Little research has addressed the heterogeneity and mortality risk in body mass index (BMI) trajectories among older populations. Applying latent class trajectory models to 9,538 adults aged 51 to 77 years from the US Health and Retirement Study (1992-2008), we defined 6 latent BMI trajectories: normal weight downward, normal weight upward, overweight stable, overweight obesity, class I obese upward, and class II/III obese upward. Using survival analysis, we found that people in the overweight stable trajectory had the highest survival rate, followed by those in the overweight obesity, normal weight upward, class I obese upward, normal weight downward, and class II/III obese upward trajectories. The results were robust after controlling for baseline demographic and socioeconomic characteristics, smoking status, limitations in activities of daily living, a wide range of chronic illnesses, and self-rated health. Further analysis suggested that BMI trajectories were more predictive of mortality risk than was static BMI status. Using attributable risk analysis, we found that approximately 7.2% of deaths after 51 years of age among the 1931-1941 birth cohort were due to class I and class II/III obese upward trajectories. This suggests that trajectories of increasing obesity past 51 years of age pose a substantive threat to future gains in life expectancy.

Trajectories of cognitive decline in Alzheimer's disease

Late-onset Alzheimer disease (LOAD) is a clinically heterogeneous complex disease defined by progressively disabling cognitive impairment. Psychotic symptoms which affect approximately one-half of LOAD subjects have been associated with more rapid cognitive decline. However, the variety of cognitive trajectories in LOAD, and their correlates, have not been well defined. We therefore used latent class modeling to characterize trajectories of cognitive and behavioral decline in a cohort of AD subjects. 201 Caucasian subjects with possible or probable Alzheimer's disease (AD) were evaluated for cognitive and psychotic symptoms at regular intervals for up to 13.5 years. Cognitive symptoms were evaluated serially with the Mini-mental State Examination (MMSE), and psychotic symptoms were rated using the CERAD behavioral rating scale (CBRS). Analyses undertaken were latent class mixture models of quadratic trajectories including a random intercept with initial MMSE score, age, gender, education, and APOE 4 count modeled as concomitant variables. In a secondary analysis, psychosis status was also included. AD subjects showed six trajectories with significantly different courses and rates of cognitive decline. The concomitant variables included in the best latent class trajectory model were initial MMSE and age. Greater burden of psychotic symptoms increased the probability of following a trajectory of more rapid cognitive decline in all age and initial MMSE groups. APOE 4 was not associated with any trajectory. Trajectory modeling of longitudinal cognitive and behavioral data may provide enhanced resolution of phenotypic variation in AD.

Diverse trajectories of cocaine use through early adulthood among rebellious and socially conforming youth

This paper tests predictions of continuity and change in antisocial behavior over time as derived from population heterogeneity and life-course perspectives. These predictions are assessed with respect to a rarely studied form of delinquent/criminal behavior, cocaine use during the late-teenage and young adult years. We first examine the extent to which differential propensities toward antisocial behavior can be detected in a nationally representative sample of youth aged 14–16 in 1979. Based on self-reported delinquent and criminal activities in late adolescence, traditional cross-sectional latent-class analysis identifies three groups of antisocial/rebellious respondents and a group of non-offenders. We then follow these groups into early adulthood, examining age trajectories of cocaine usage between 1984 and 1998. Latent-class trajectory models identify clusters of respondents who show similar age trajectories of cocaine use over time and provide parameter estimates that predict membership in those clusters. In support of the population heterogeneity perspective, we find that antisocial/rebellious youth have higher probabilities of cocaine use throughout early adulthood than non-offending youth. There is, however, much variation in drug use patterns among the groups as they aged. In support of a life-course perspective, we find that social ties to schools, families, religion, and the labor market help differentiate youth who refrain from, maintain, or desist from using cocaine through early adulthood.