Latent Class Growth Mixture Modeling Research Articles

Association of longitudinal trajectories of fasting plasma glucose with all-cause and cardiovascular mortality among a Chinese older population: a retrospective cohort study

The association between fasting plasma glucose (FPG), an important indicator of overall glycemic status, and the risk of cardiovascular mortality has been well investigated. The longitudinal study can repeatedly collect measured results for the variables to be studied and then consider the potential effects of intraindividual changes in measurement. This study aimed to identify long-term FPG trajectories and investigate the association between trajectory groups and cardiovascular and all-cause mortality. A latent class growth mixture modeling (LCGMM) was used to identify FPG trajectories. Cox proportional hazard models were used to estimate associations between FPG trajectories and the risk of all-cause and cardiovascular mortality. A U-shaped relationship between FPG and all-cause and cardiovascular mortality was observed in the restricted cubic spline regression models. Two FPG longitudinal trajectories of low-level (mean FPG = 5.12mmol/L) and high-level (mean FPG = 6.74mmol/L) were identified by LCGMM. After being adjusted for potential confounders, compared with the low-level category, the hazard ratios (HRs) for all-cause and cardiovascular mortality were 1.23(1.16–1.30) and 1.25(1.16–1.35), respectively, for the high-level group. Long-term FPG trajectories are significantly associated with and potentially impact the risk of all-cause and cardiovascular mortality.

P062 Trust performance in managing inflammatory arthritis over time in England and Wales: a latent class analysis approach

Abstract Background/Aims To evaluate trust-level performance in time to initiation of disease-modifying antirheumatic drug (DMARD) therapy in patients with early inflammatory arthritis (EIA), with identification of change in performance trajectories over time and investigation of trust characteristics associated with this change. Methods We included 130 trusts from the UK contributing to the National Early Inflammatory Arthritis Audit (NEIAA) from 2018 to 2020. The primary outcome was days from referral to initiation of DMARD therapy in patients with EIA. Latent class growth mixture models were applied to identify distinct groups of trusts with similar trajectories of performance change over time. We used mixed-effects linear and multinomial logistic regression models to evaluate the association between delay in treatment and trust-level characteristics. Results The mean (SD) time to DMARD initiation was 53 (18) days, with an average 0.3-day decrease with each month over time. Four latent trajectories were identified in our cohort, with over 77% of individual trusts showing ongoing improvements in decreasing treatment waiting times. Prior to separating by latent class, time to DMARD initiation was shorter in trusts with higher rheumatology staffing, a local EIA treatment pathway and those with access to musculoskeletal ultrasound. Trusts with more nurses in the rheumatology department were less likely to be in the worse performance group (OR = 0.69 [95%CI 0.49, 0.93]). Conclusion In this cohort study, we observed a reduction in treatment waiting time over time. Trusts with better staffed and improved EIA clinical structure are likely to initiate definitive treatment earlier in patients with EIA. Disclosure Z. Yang: None. N. Arumalla: None. M.D. Russell: Honoraria; Lilly, Menarini, Galapagos. Grants/research support; NIHR. Other; Pfizer, Janssen, Biogen, and UCB. K. Bechman: Honoraria; Menarini, Galapagos. Grants/research support; NIHR,Versus Arthritis and Pfizer Global Medical Grants for Quality Improvement in Rheumatology Practice. Other; UCB. E. Alveyn: Other; UCB. S. Gallagher: None. E. Price: None. S. Norton: None. J. Galloway: Honoraria; AbbVie, Biovitrum, BMS, Celgene, Chugai, Galapagos, Gilead, Janssen, Lilly, Novartis, Pfizer, Roche, Sanofi, Sobi and UCB.

Trajectory patterns and cumulative burden of CEA during follow-up with non-small cell lung cancer outcomes: A retrospective longitudinal cohort study.

Previous studies of non-small cell lung cancer (NSCLC) focused on CEA measured at a single time point, ignoring serial CEA measurements. This retrospective cohort included 2959 patients underwent surgery for stage I-III NSCLC. CEA trajectory patterns and long-term cumulative CEA burden were evaluated using the latent class growth mixture model. Four CEA trajectory groups were identified, named as low-stable, decreasing, early-rising and later-rising. Compared with the low-stable group, the adjusted hazard ratios associated with death were 1.27, 4.50, and 3.68 for the other groups. Cumulative CEA burden were positively associated with the risk of death in patients not belonging to the low-stable group. The 5-year overall survival (OS) rates decreased from 62.3% to 33.0% for the first and fourth quantile groups of cumulative CEA burden. Jointly, patients with decreasing CEA trajectory could be further divided into the decreasing & low and decreasing & high group, with 5-year OS rates to be 77.9% and 47.1%. Patients with rising CEA trajectory and high cumulative CEA were found to be more likely to develop bone metastasis. Longitudinal trajectory patterns and long-term cumulative burden of CEA were independent prognostic factors of NSCLC. We recommend CEA in postoperative surveillance of NSCLC.

Association of healthy lifestyle with life expectancy free of five major disabilities in Chinese older adults.

Whether and to what extent multiple healthy lifestyles affect the longevity of people with disabilities, including those in basic activities of daily living, mobility, vision, hearing and cognition, is crucial to policymakers. We aimed to determine the impact of combined lifestyles on life expectancy (LE) lived with and without five disabilities. We recruited participants (n = 15 121 from the China Longitudinal Healthy Longevity Survey between 2008 and 2018. Healthy lifestyle levels were estimated from six factors: smoking, drinking, physical exercise, diet, cognitive activity, and sleep, which we categorised as favourable and unfavourable using the latent class growth mixture model throughout the follow-up period. We used Multi-state Markov models to assess the different disability stages of LE. Of the total LE at age 65, older adults with a favourable lifestyle spent 59.60% (disability-free LE (DFLE) = 10.24 years) without five disabilities in combination, whereas those with unfavourable lifestyle spent 56.74% (DFLE = 7.28 years). Furthermore, the percentage of DFLE was 64.98 (7.71 years) and 68.38 (9.91 years) in males with unfavourable and favourable lifestyle levels, respectively, and 47.92 (6.62 years) and 55.12 (10.30 years) for females. Compared to older adults with low socioeconomic status (SES) and unfavourable lifestyle level, those with lower SES and favourable lifestyle level had more 3.77 years of DFLE, those with higher SES and unfavourable lifestyle level had more 1.94 years, as well as those with higher SES and favourable lifestyle level had more 5.10 years at age 65. Corresponding associations were found separately for each of the five individual disabilities. A favourable lifestyle level was associated with longer total LE along with a higher proportion of DFLE and may contribute to narrowing socioeconomic health inequalities. Policymakers should develop lifestyle interventions and scale up rehabilitation services in primary care, thereby delaying disabilities to later ages, especially in low- and middle-income countries.

Trust performance in managing inflammatory arthritis over time in England and Wales: a latent class analysis approach.

To evaluate trust-level performance in time to initiation of DMARD therapy in patients with early inflammatory arthritis (EIA), with identification of the change in performance trajectories over time and investigation of trust characteristics associated with this change. We included 130 trusts from the UK contributing to the National Early Inflammatory Arthritis Audit (NEIAA) from 2018 to 2020. The primary outcome was days from referral to initiation of DMARD therapy in patients with EIA. Latent class growth mixture models were applied to identify distinct groups of trusts with similar trajectories of performance change over time. We used mixed effects linear and multinomial logistic regression models to evaluate the association between delay in treatment and trust-level characteristics. The mean time to DMARD initiation was 53days (s.d. 18), with an average 0.3-day decrease with each month over time. Four latent trajectories were identified in our cohort, with >77% of individual trusts showing ongoing improvements in decreasing treatment waiting times. Prior to separating by latent class, time to DMARD initiation was shorter in trusts with higher rheumatology staffing, a local EIA treatment pathway and those with access to musculoskeletal ultrasound. Trusts with more nurses in the rheumatology department were less likely to be in the worst performance group [odds ratio 0.69 (95% CI 0.49, 0.93)]. In this cohort study, we observed a reduction in treatment waiting time over time. Trusts with better staffed and improved EIA clinical structure are likely to initiate definitive treatment earlier in patients with EIA.

Anxiety trajectories from pregnancy to one-year postpartum and their contributors during the COVID-19 pandemic

ABSTRACT Background Anxiety in the peripartum period (pregnancy through one-year postpartum) has negative impacts on mothers and infants. During the COVID-19 pandemic, high rates of anxiety were reported worldwide, but trajectories for how these symptoms change longitudinally in the peripartum period remain unknown. Methods A total of 1,411 women who were pregnant during the second U.S. COVID-19 surge completed four study questionnaires between December 2020 and March 2022, including assessments of anxiety symptoms (GAD-7) and individual and community-level contributors. Latent Class Growth Mixture Models (LCGMM) were used to identify anxiety trajectories, and multinomial logistic regression was used to identify contributors to class membership. Results Three classes of anxiety symptom trajectories were identified. Class 1 (n = 956, 67.8%) experienced minimal/mild anxiety throughout the peripartum period. Class 2 (n = 362, 25.7%) experienced initial moderate anxiety that remained stable. Class 3 (n = 93, 6.6%) experienced initial moderate anxiety that increased to severe in the postpartum period. Stress and history of mental health diagnosis contributed to higher initial levels of anxiety and ongoing social support contributed to lower levels of anxiety. In addition, financial insecurity increased the risk of belonging to Class 3. Conclusions Screening for anxiety symptoms and psychosocial vulnerabilities in the peripartum period is vital, as is early intervention. This is especially the case among women with limited financial and social resources who may experience chronic or worsening anxiety as they transition to parenthood.

Trajectories of Blood Lipids Profile in Midlife Women: Does Menopause Matter?

Background It remains controversial whether changes of lipids over menopause transition (MT) are more age-related or more menopause-related. We aimed to classify women into different trajectory groups based on pattern and level of total cholesterol, low-density lipoprotein cholesterol, apolipoprotein B (ApoB), high-density lipoprotein cholesterol (HDL-C), triglyceride, and apolipoprotein A-I over the MT, as well as examine the effect of MT-related factors on lipid trajectory groups and levels. Methods and Results The cohort included 2582 subjects from the Study of Women's Health Across the Nation. Different trajectory patterns of lipids during the MT were determined using the latent class growth mixture model. The predictors of distinct blood lipids trajectory groups were determined by multiple linear regression models and multinomial logistic regression models. Women were categorized into either inverse U-shape or progressing trajectory group in each blood lipids measurement. The inverse U-shape total cholesterol, low-density lipoprotein cholesterol, apolipoprotein B, high-density lipoprotein cholesterol, log(TG), and apolipoprotein A-I trajectories showed an increasing trend before menopause but a decreasing trend after menopause. The U-shape total cholesterol, low-density lipoprotein cholesterol, and apolipoprotein B trajectories started to rise 5 years before menopause. Age at menopause, follicle-stimulating hormone, vasomotor symptoms, and estradiol predicted the shape and level of the women's lipids over the MT. Conclusions Distinct lipid trajectories were identified during the MT, and the existence of at least 1 trajectory in each lipid parameters suggested a contribution of menopause. Our study highlights the need for earlier and continuous surveillance of lipids during the MT.

Remnant Cholesterol in Young Adulthood Is Associated With Left Ventricular Remodeling and Dysfunction in Middle Age: The CARDIA Study.

Recent studies have shown that remnant cholesterol (RC) is associated with incident heart failure; however, its association with left ventricular (LV) structure and function is unclear. We aimed to evaluate the association between RC levels in young adulthood and LV structure and function in middle age. We included 3321 participants from the CARDIA study (Coronary Artery Risk Development in Young Adults) at baseline. RC was calculated as total cholesterol minus high-density lipoprotein cholesterol minus calculated low-density lipoprotein cholesterol, and the RC trajectories that followed a similar pattern of change over time were identified using the latent class growth mixture model. LV structure and function were assessed using echocardiography at CARDIA study year 25. Multivariable linear regression models were performed to assess the associations of both baseline and trajectories of RC levels with LV structure and function. Among 3321 participants, the mean age was 24.99±3.62 years: 1450 (43.90%) were male, and 1561 (47.00%) were Black. After multivariate adjustment, higher baseline RC (per SD in log-transformed) was associated with higher LV mass index (β=1.29; P=0.004), worse global longitudinal strain (β=0.19; P<0.001), worse global circumferential strain (β=0.16; P=0.014), lower septal e' (β=-0.26; P<0.001), lower lateral e' (β=-0.18; P=0.003), and higher E/e' (β=0.15; P=0.003). Three RC trajectories were identified during follow-up: low increasing (42.4%), moderate increasing (45.5%), and high increasing (12.1%). Similarly, compared with the low-increasing group, the high-increasing RC trajectory group was related to higher LV mass index, worse global longitudinal strain, lower septal e', lower lateral e', and higher E/e'. Elevated RC levels in young adulthood were related to adverse LV structural and functional alterations in midlife. Long-term trajectories of RC levels during young adulthood help identify individuals at a higher risk for adverse LV remodeling and dysfunction. URL: https://www.clinicaltrials.gov; Unique identifier: NCT00005130.

20-year trajectories of positive and negative symptoms after the first psychotic episode in patients with schizophrenia spectrum disorder: results from the OPUS study.

This study aimed to identify the 20-year trajectories of positive and negative symptoms after the first psychotic episode in a sample of patients with an ICD-10 diagnosis of schizophrenia spectrum disorder, and to investigate the baseline characteristics and long-term outcomes associated with these trajectories. A total of 373 participants in the OPUS trial were included in the study. Symptoms were assessed at baseline and after 1, 2, 5, 10 and 20 years using the Scales for the Assessment of Positive and Negative Symptoms. We used latent class growth mixture modelling to identify trajectories, and multinominal regression analyses to investigate predictors of membership to identified trajectories. Five trajectories of positive symptoms were identified: early continuous remission (50.9% of the sample), stable improvement (18.0%), intermittent symptoms (10.2%), relapse with moderate symptoms (11.9%), and continuous severe symptoms (9.1%). Substance use disorder (odds ratio, OR: 2.83, 95% CI: 1.09-7.38, p=0.033), longer duration of untreated psychosis (OR: 1.02, 95% CI: 1.00-1.03, p=0.007) and higher level of negative symptoms (OR: 1.60, 95% CI: 1.07-2.39, p=0.021) were predictors of the relapse with moderate symptoms trajectory, while only longer duration of untreated psychosis (OR: 1.01, 95% CI: 1.00-1.02, p=0.030) predicted membership to the continuous severe symptoms trajectory. Two trajectories of negative symptoms were identified: symptom remission (51.0%) and continuous symptoms (49.0%). Predictors of the continuous symptoms trajectory were male sex (OR: 3.03, 95% CI: 1.48-6.02, p=0.002) and longer duration of untreated psychosis (OR: 1.01, 95% CI: 1.00-1.02, p=0.034). Trajectories displaying continuous positive and negative symptoms were linked to lower neurocognition, as measured by the Brief Assessment of Cognition in Schizophrenia (BACS) (z-score: -0.78, CI: -1.39 to -0.17, for continuous positive symptoms; z-score: -0.33, CI: -0.53 to -0.13, for continuous negative symptoms). The same trajectories were also linked to higher use of antipsychotic medication at 20-year follow-up (continuous positive symptoms: 78%; continuous negative symptoms: 67%). These findings suggest that the majority of patients with first-episode schizophrenia spectrum disorder have a trajectory with early stable remission of positive symptoms. Long duration of untreated psychosis and comorbid substance abuse are modifiable predictors of poor trajectories for positive symptoms in these patients. In about half of patients, negative symptoms do not improve over time. These symptoms, in addition to being associated with poor social and neurocognitive functioning, may prevent patients from seeking help.

Association of blood pressure trajectories with coronary heart disease among the disabled population in Shanghai, China: a cohort study of 7 years following up

BackgroundMuch less is known about the importance of blood pressure (BP) trajectories concerning the incidence of coronary heart disease (CHD) in people with disabilities. Our aim was to evaluate this association.MethodsThis cohort study surveyed 5711 adults from the Shanghai Disability Health Survey from June 2012 to June 2019. The latent class growth mixture model was used to examine distinct BP trajectories. We evaluated the association of BP trajectories with the risk of CHD by Cox proportional hazard models. The model for CHD risk fitted to BP trajectories was compared with models fitted to other BP-related indicators by goodness-of-fit, discrimination, and calibration.ResultsDuring a median follow-up of 71.74 months, 686 cases (median age was 49.03 (54.49, 58.55) years, 51.90% female) with CHD were identified, with a cumulative incidence of 12.01%. Systolic BP (SBP) and diastolic BP (DBP) were categorized into three classes, respectively. A statistically significant association was only observed between SBP trajectories and CHD. Compared with the normotensive stable SBP group (n = 1956), the prehypertension-stable group (n = 3268) had a higher risk (adjust hazards ratio (aHR) = 1.266, 95% confidence interval (CI) 1.014–1.581), and the stage 1 hypertension-decreasing group (n = 487) had the highest risk (aHR = 1.609, 95%CI 1.157–2.238). Among the BP-related indicators, the SBP trajectory was the strongest predictor of new-onset CHD. Findings were similar when sensitivity analyses were conducted.ConclusionsSBP trajectory was a more important risk factor for CHD than other BP-related indicators and stringent BP control strategies may be effective for primary CHD prevention in the disabled population.

Trajectories of Traumatic Stress Symptoms Among Siblings of Children With Cancer: The First Two Years Post-Diagnosis.

Identify and describe trajectories of cancer-related posttraumatic stress symptoms (PTSS) among siblings of children with cancer within two years of diagnosis. Siblings (aged 8-18; M = 11.2 years) across the United States, and for each, one caregiver, were recruited for a cohort sequential longitudinal study with three data collection points six months apart beginning at 6- or 12-months after cancer diagnosis. Siblings (N = 229; 42% of eligible/approached; 53% identifying as female; 68% identifying as non-Hispanic White) completed the Child Posttraumatic Stress Disorder Symptom Scale. Caregivers completed the Strengths and Difficulties Questionnaire (SDQ). Latent class growth analysis (LCGA) and growth mixture modeling (GMM) identified PTSS patterns across time. Fit statistics supported models with three to five PTSS trajectories. The three-class LCGA model included a large mild PTSS group (61%), a moderate PTSS group (35%), and a small (4%) stable severe PTSS group. The four-class LCGA and three- and four-class GMM included groups improving from moderate to mild PTSS (7-21%) and worsening to moderate PTSS across time (12-17%). Across models, siblings with mild PTSS had fewer caregiver-reported emotional and behavioral difficulties on the SDQ. A large group of siblings of children with cancer demonstrate resiliency, however, substantial subsets experience patterns of PTSS that include levels in the moderate-to-severe range during the first two years post-diagnosis. Future research should examine these patterns in more diverse/representative samples and identify factors associated with increasing and sustained severe PTSS to inform intervention targets and reduce cancer-related burden on families.

Identifying patient subgroups with different trends of patient-reported outcomes (PROMs) after elective knee arthroplasty

BackgroundPatient-reported outcome measures (PROMs) are increasingly being used to assess the effectiveness of elective total knee arthroplasty (TKA). However, little is known about how PROMs scores change over time in these patients. The aim of this study was to identify the trajectories of quality of life and joint functioning, and their associated demographic and clinical features in patients undergoing elective TKA.MethodsA prospective, cohort study was conducted, in which PROMs questionnaires (Euro Quality 5 Dimensions 3L, EQ-5D-3L, and Knee injury and Osteoarthritis Outcome Score Patient Satisfaction, KOOS-PS) were administered to patients at a single center undergoing elective TKA before surgery, and at 6 and 12 months after surgery. Latent class growth mixture models were used to analyze the patterns of change in PROMs scores over time. Multinomial logistic regression was used to investigate the association between patient characteristics and PROMs trajectories.ResultsA total of 564 patients were included in the study. The analysis highlighted differential patterns of improvement after TKA. Three distinct PROMs trajectories were identified for each PROMs questionnaire, with one trajectory indicating the most favorable outcome. Female gender appears to be associated with a presentation to surgery with worse perceived quality of life and joint function than males, but also more rapid improvement after surgery. Having an ASA score greater than 3 is instead associated with a worse functional recovery after TKA.ConclusionThe results suggest three main PROMs trajectories in patients undergoing elective TKA. Most patients reported improved quality of life and joint functioning at 6 months, which then stabilized. However, other subgroups showed more varied trajectories. Further research is needed to confirm these findings and to explore the potential clinical implications of these results.

Distinct patterns of adult transport-related physical activity (TRPA) behaviour exist independent of the TRPA behaviours of childhood: the childhood determinants of adult health study

BackgroundTransport-related physical activity (TRPA) is recognised as a potential means of increasing total physical activity participation that may yield substantial health benefits. Public health campaigns focusing on promoting TRPA from a young age aim to develop life-long healthy habits. However, few studies have examined how TRPA changes across the lifecourse and whether childhood TRPA levels influence those observed later in life.MethodsUsing the Australian Childhood Determinants of Adult Health study (baseline, 1985), latent class growth mixture modelling with adjustment for time-varying covariates was performed using four timepoints (ranging from 7 to 49 years) to assess behavioural patterns and retention of TRPA across the lifecourse. As child and adult adjusted TRPA measures could not be harmonised, trajectories of adult TRPA (n = 702) were instead identified, and log-binomial regression analysis was performed to determine whether childhood levels of TRPA (high/medium/low) influenced these trajectories.ResultsTwo stable groups of adult TRPA trajectories were identified: persistently low (n = 520; 74.2%), and increasingly high TRPA (n = 181; 25.8%). There was no significant relationship between childhood TRPA levels and patterns in adulthood (relative risk of high childhood TRPA yielding high adult TRPA trajectory membership = 1.06; 95% confidence interval = 0.95–1.09).ConclusionThis study found childhood TRPA levels were not associated with TRPA patterns in adulthood. These findings suggest that while TRPA in childhood may have health, social, and environmental benefits, it does not appear to impact adult TRPA directly. Therefore, further intervention is required beyond childhood to promote the implementation of healthy TRPA behaviours into adulthood.

Baseline level and change trajectory of the triglyceride-glucose index in relation to the development of NAFLD: a large population-based cohort study.

Insulin resistance (IR) and nonalcoholic fatty liver disease (NAFLD) are closely related. The triglyceride-glucose index (TyG index) has been proposed as a new indicator of IR. It remains unclear whether the triglyceride-glucose (TyG) index is prospectively associated with incident nonalcoholic fatty liver disease (NAFLD). This large-scale study comprised 1 prospective cohort totaling 22,758 subjects without NAFLD at baseline who underwent repeated health examinations and 1 subcohort totaling 7,722 subjects with more than three visits. The TyG index was ascertained mathematically by ln (fasting triglycerides [mg/dL] × fasting glucose [mg/dL]/2). NAFLD was diagnosed by ultrasound without other concomitant liver diseases. A combinatorial Cox proportional hazard model and latent class growth mixture modeling method were used to identify the association of the TyG index and its transition trajectories with NAFLD risk. During 53,481 person-years of follow-up, there were 5319 incident cases with NAFLD. Compared with those in the lowest quartile of the baseline TyG index, participants in the highest quartile had 2.52-fold (95% confidence interval, 2.21-2.86) higher odds of incident NAFLD. Similarly, restricted cubic spline analysis showed a dose-response relationship (p nonlinearity<0.001). Subgroup analyses showed a more significant association in the female and normal body size populations (p for interaction<0.001). Three distinct trajectories of changes in the TyG index were identified. Compared with the continued low group, the moderately increasing and highly increasing groups conferred 1.91-fold (1.65-2.21) and 2.19-fold (1.73-2.77) higher NAFLD risk, respectively. Participants with a higher baseline TyG index or a higher excessive TyG exposure were associated with an increased NAFLD risk. The findings imply that lifestyle interventions and modulation of IR might be considered to both reduce TyG index levels and prevent NAFLD development.

Patient-reported mental health and well-being trajectories in oncology patients during radiation therapy: an exploratory retrospective cohort analysis using the Ontario Cancer Registry.

Mental health and well-being trajectories are not expected to be homogeneous in diverse clinical populations. This exploratory study aims to identify subgroups of patients with cancer receiving radiation therapy who have different mental health and well-being trajectories, and examine which socio-demographic, physical symptoms, and clinical variables are associated with such trajectories. Retrospective analysis of radiation therapy patients diagnosed with cancer in 2017 was conducted using data from the Ontario Cancer Registry (Canada) and linked with administrative health data. Mental health and well-being were measured using items from the Edmonton Symptom Assessment System-revised questionnaire. Patients completed up to 6 repeated measurements. We used latent class growth mixture models to identify heterogeneous mental health trajectories of anxiety, depression, and well-being. Bivariate multinomial logistic regressions were conducted to explore variables associated with the latent classes (subgroups). The cohort (N = 3416) with a mean age of 64.5years consisted of 51.7% females. Respiratory cancer was the most common diagnosis (30.4%) with moderate to severe comorbidity burden. Four latent classes with distinct anxiety, depression, and well-being trajectories were identified. Decreasing mental health and well-being trajectories are associated with being female; living in neighborhoods with lower income, greater population density, and higher proportion of foreign-born individuals; and having higher comorbidity burden. The findings highlight the importance of considering social determinants of mental health and well-being, in addition to symptoms and clinical variables, when providing care for patients undergoing radiation therapy.

Changes in UK parental mental health symptoms over 10months of the COVID-19 pandemic.

The threats to health, associated restrictions and economic consequences of the COVID-19 pandemic have been linked to increases in mental health difficulties for many. Parents, in particular, have experienced many challenges such as having to combine work with home-schooling their children and other caring responsibilities. Yet, it remains unclear how parental mental health has changed throughout the pandemic or what factors may have mitigated or compounded the impact of the pandemic on parents' mental health. We examined monthly survey data from two linked UK-based longitudinal studies: COVID-19: Supporting Parents, Adolescents and Children during Epidemics' (Co-SPACE) and COVID-19: Supporting Parents and Young Children during Epidemics' (Co-SPYCE). Data from 5576 parents/carers of 2-17-year-old children collected between April 2020 and January 2021 was analysed using mixed-effect modelling and latent class growth (mixture) modelling. Parental stress and depression, but not anxiety, were higher during the periods of restrictions. This pattern was most pronounced for parents with primary-school-aged children, those that worked at home or had other adults in the household. Being younger, reporting secondary or below education, working out of home, having secondary-school-aged children or children with special education needs (SEN)/neurodevelopmental disorders (ND) further moderated whether, how and when parental mental health symptoms changed. Although around three quarters of parents reported consistently low mental health symptoms, a substantial minority reported consistently high or increasing symptoms of anxiety, stress and depression. The latter were more likely to be parents who were younger than average, were a single adult in the household, had a pre-existing mental health diagnosis or had a child with special educational needs or a ND. These findings emphasise how different personal circumstances and pre-existing inequalities shaped how parents were affected by this unprecedented global pandemic and highlight the need for support and consideration to meet the needs of families in the future.

Long-term trajectories of depressive symptoms and machine learning techniques for fall prediction in older adults: Evidence from the China Health and Retirement Longitudinal Study (CHARLS)

BackgroundFalls are the most common adverse outcome of depression in older adults, yet a accurate risk prediction model for falls stratified by distinct long-term trajectories of depressive symptoms is still lacking. MethodsWe collected the data of 1617 participants from the China Health and Retirement Longitudinal Study register, spanning between 2011 and 2018. The 36 input variables included in the baseline survey were regarded as candidate features. The trajectories of depressive symptoms were classified by the latent class growth model and growth mixture model. Three data balancing technologies and four machine learning algorithms were utilized to develop predictive models for fall classification of depressive prognosis. ResultsDepressive symptom trajectories were divided into four categories, i.e., non-symptoms, new-onset increasing symptoms, slowly decreasing symptoms, and persistent high symptoms. The random forest-TomekLinks model achieved the best performance among the case and incident models with an AUC-ROC of 0.844 and 0.731, respectively. In the chronic model, the gradient boosting decision tree-synthetic minority oversampling technique obtained an AUC-ROC of 0.783. In the three models, the depressive symptom score was the most crucial component. The lung function was a common and significant feature in both the case and the chronic models. ConclusionsThis study suggests that the ideal model has a good chance of identifying older persons with a high risk of falling stratified by long-term trajectories of depressive symptoms. Baseline depressive symptom score, lung function, income, and injury experience are influential factors associated with falls of depression evolution.

Association of body roundness index and its trajectories with all-cause and cardiovascular mortality among a Chinese middle-aged and older population: A retrospective cohort study.

The body roundness index (BRI) is a novel anthropometric index that is a better indicator for predicting fat distribution than the body mass index (BMI). The longitudinal study can repeatedly collect measured results for the variables to be studied and then consider the potential effects of intraindividual changes in measurement. However, few population-based, longitudinal studies of BRI have been conducted, especially among the Chinese population. The study aimed to investigate the association of BRI and its longitudinal trajectories with all-cause and cardiovascular mortality. A total of 71,166 participants with four times BRI measurements between January 2010 and December 2019 were included in this longitudinal study, with a median follow-up was 7.93 years, and 11,538 deaths were recorded, of which 5,892 deaths were due to cardiovascular disease (CVD). A latent class growth mixture modeling (LCGMM) was used to identify BRI trajectories. Cox proportional hazard models were used to estimate associations between BRI trajectories and the risk of all-cause and cardiovascular mortality. In the restricted cubic spline regression models, a U-shaped relationship between BRI and all-cause and cardiovascular mortality was observed. Three BRI longitudinal trajectories of low-stable (mean BRI = 2.59), moderate-stable (mean BRI = 3.30), and high-stable (mean BRI = 3.65) were identified by LCGMM. After being adjusted for potential confounders, the HRs for all-cause mortality were 1.18 (1.13-1.24) for the moderate-stable group and 1.74 (1.66-1.82) for the high-stable group compared to the low-stable group. The HRs for cardiovascular mortality were 1.12 (1.05-1.18) for the moderate-stable group and 1.64 (1.53-1.75) for the high-stable group compared to the low-stable group. A nonlinear association of BRI with all-cause and cardiovascular mortality was observed, and participants in the higher BRI longitudinal trajectory group were significantly associated with an increased risk of all-cause and cardiovascular mortality.

Patient-Reported Symptom Severity, Health-Related Quality of Life, and Emotional Distress Trajectories During and After Radiation Therapy for Human Papillomavirus–Associated Oropharyngeal Cancer: A TROG 12.01 Secondary Analysis

This secondary analysis of clinical trial TROG 12.01, involving patients with human papillomavirus-associated oropharyngeal squamous cell carcinoma, aimed to identify patient-reported outcome (PRO) trajectories before, during, and after chemoradiotherapy. Head and neck cancer symptom severity (HNSS) and interference (HNSI), generic health-related quality of life (HRQL), and emotional distress were assessed with the MD Anderson Symptom Inventory-Head and Neck, Functional Assessment of Cancer Therapy-General, and Hospital Anxiety and Depression Scale questionnaires, respectively. Latent class growth mixture modeling (LCGMM) was used to identify distinct underlying trajectories. Baseline and treatment variables were compared between trajectory groups. The LCGMM identified latent trajectories for all PROs: HNSS, HNSI, HRQL, anxiety, and depression. Four HNSS trajectories (HNSS1-4) were identified, distinguished by differences in HNSS at baseline, during the peak of treatment symptoms, and during early and intermediate recovery. All trajectories were stable beyond 12 months. The reference trajectory (HNSS4, n=74) score was 0.1 (95% confidence interval [CI], 0.1-0.2) at baseline, peaking at 4.6 (95% CI, 4.2-5.0), with rapid early recovery (1.1; 95% CI, 0.8-2.2) and gradual improvement to 12 months (0.6; 95% CI, 0.5-0.8). Patients in HNSS2 (high baseline, n=30) reported higher baseline scores (1.4; 95% CI, 0.8-2.0) but were otherwise similar to HNSS4. Patients in HNSS3 (low acute, n=53) reported reduced acute symptoms (2.5; 95% CI, 2.2-2.9) with stable scores beyond 9 weeks after chemoradiotherapy (1.1; 95% CI, 0.9-1.4). Patients in HNSS1 (slow recovery, n=25) had slower recovery from an acute peak of 4.9 (95% CI, 4.3-5.6) to 0.9 (95% CI, 0.6-1.3) at 12 months. Age, performance status, education, receipt of cetuximab, and baseline anxiety varied between trajectories. The other PRO models demonstrated clinically relevant trajectories, with distinct associations with baseline factors. LCGMM identified distinct PRO trajectories during and after chemoradiotherapy. These and their associations with variations in the characteristics and treatment factors of patients with human papillomavirus-associated oropharyngeal squamous cell carcinoma provide clinically relevant insights into identifying patients who may require increased support before, during, or after chemoradiotherapy.

Abstract PD6-02: PD6-02 Health-Related Quality of Life Trajectories and Predictors among Older Breast Cancer Survivors: A SEER-MHOS analysis

Abstract Health-related quality of life (HRQOL) is an important issue for breast cancer patients and clinicians in the treatment decision process. In addition, older women are more vulnerable due to pre-existing comorbidities and socioeconomic status. This study explored 10-year trajectories of HRQoL in older breast cancer survivors and their predictors using patient-reported outcomes data queried from the Surveillance, Epidemiology and End Results - Medicare Health Outcomes Survey (SEER-MHOS) data resources among Medicare beneficiaries. Older women diagnosed with breast cancer in 1998-2012 and who participated in the surveys before and at least once after diagnosis were included in the analysis. HRQOL was measured using SF-36/VR-12 questionnaire and summarized as Physical Component Summary (PCS) Score and Mental Component Summary (MCS) Score. Latent Class Growth Mixture Modeling was conducted to identify distinct groups of women with a similar trajectory of HRQOL. A total of 1089 women with breast cancer completed surveys an average of 2.6 times (range 2 - 19). The results showed that there were three latent classes of PCS trajectory: high-declining (46% of the sample), mid-declining (37%), and low-improving (17%). Two latent classes of MCS trajectory were identified: high-stable (76%) and low-declining (24%). The univariate analysis showed that age at diagnosis, body mass index, level of education, geographic region, tumor grade, tumor size, and the number of comorbidities were related to PCS and MCS scores. Multivariable multinomial logistic regression analysis identified the number of comorbidities as the most significant predictor for PCS score and level of education as the most significant predictor for MCS score. Future research needs to identify the most common comorbidities that influence HRQOL deterioration in older breast cancer survivors to develop interventions that better the physical HRQOL in patients. Also, interventions that target less educated, underserved patients to improve mental HRQOL. Citation Format: Eunkyung Lee, Sushantti Rupesh, Katia Ferdowsi, Robert Hines, Victoria Loerzel. PD6-02 Health-Related Quality of Life Trajectories and Predictors among Older Breast Cancer Survivors: A SEER-MHOS analysis [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD6-02.