Developing survey weights to ensure representativeness in a national, matched cohort study: results from the children and young people with Long Covid (CLoCk) study.

BACKGROUND: Findings from studies assessing Long Covid in children and young people (CYP) need to be assessed in light of their methodological limitations. For example, if non-response and/or attrition over time systematically differ by sub-groups of CYP, findings could be biased and any generalisation limited. The present study aimed to (i) construct survey weights for the Children and young people with Long Covid (CLoCk) study, and (ii) apply them to published CLoCk findings showing the prevalence of shortness of breath and tiredness increased over time from baseline to 12-months post-baseline in both SARS-CoV-2 Positive and Negative CYP. METHODS: Logistic regression models were fitted to compute the probability of (i) Responding given envisioned to take part, (ii) Responding timely given responded, and (iii) (Re)infection given timely response. Response, timely response and (re)infection weights were generated as the reciprocal of the corresponding probability, with an overall 'envisioned population' survey weight derived as the product of these weights. Survey weights were trimmed, and an interactive tool developed to re-calibrate target population survey weights to the general population using data from the 2021 UK Census. RESULTS: Flexible survey weights for the CLoCk study were successfully developed. In the illustrative example, re-weighted results (when accounting for selection in response, attrition, and (re)infection) were consistent with published findings. CONCLUSIONS: Flexible survey weights to address potential bias and selection issues were created for and used in the CLoCk study. Previously reported prospective findings from CLoCk are generalisable to the wider population of CYP in England. This study highlights the importance of considering selection into a sample and attrition over time when considering generalisability of findings.

Target Trial Emulation and Bias Through Missing Eligibility Data: An Application to a Study of Palivizumab for the Prevention of Hospitalization Due to Infant Respiratory Illness.

Target trial emulation (TTE) applies the principles of randomized controlled trials to the causal analysis of observational data sets. One challenge that is rarely considered in TTE is the sources of bias that may arise if the variables involved in the definition of eligibility for the trial are missing. We highlight patterns of bias that might arise when estimating the causal effect of a point exposure when restricting the target trial to individuals with complete eligibility data. Simulations consider realistic scenarios where the variables affecting eligibility modify the causal effect of the exposure and are missing at random or missing not at random. We discuss means to address these patterns of bias, namely: 1) controlling for the collider bias induced by the missing data on eligibility, and 2) imputing the missing values of the eligibility variables prior to selection into the target trial. Results are compared with the results when TTE is performed ignoring the impact of missing eligibility. A study of palivizumab, a monoclonal antibody recommended for the prevention of respiratory hospital admissions due to respiratory syncytial virus in high-risk infants, is used for illustration.

Using linked administrative data to aid the handling of non-response and restore sample representativeness in cohort studies: the 1958 national child development study and hospital episode statistics data.

BACKGROUND: There is growing interest in whether linked administrative data have the potential to aid analyses subject to missing data in cohort studies. METHODS: Using linked 1958 National Child Development Study (NCDS; British cohort born in 1958, n = 18,558) and Hospital Episode Statistics (HES) data, we applied a LASSO variable selection approach to identify HES variables which are predictive of non-response at the age 55 sweep of NCDS. We then included these variables as auxiliary variables in multiple imputation (MI) analyses to explore the extent to which they helped restore sample representativeness of the respondents together with the imputed non-respondents in terms of early life variables (father's social class at birth, cognitive ability at age 7) and relative to external population benchmarks (educational qualifications and marital status at age 55). RESULTS: We identified 10 HES variables that were predictive of non-response at age 55 in NCDS. For example, cohort members who had been treated for adult mental illness had more than 70% greater odds of bring non-respondents (odds ratio 1.73; 95% confidence interval 1.17, 2.51). Inclusion of these HES variables in MI analyses only helped to restore sample representativeness to a limited extent. Furthermore, there was essentially no additional gain in sample representativeness relative to analyses using only previously identified survey predictors of non-response (i.e. NCDS rather than HES variables). CONCLUSIONS: Inclusion of HES variables only aided missing data handling in NCDS to a limited extent. However, these findings may not generalise to other analyses, cohorts or linked administrative datasets. This work provides a demonstration of the use of linked administrative data for the handling of missing cohort data which we hope will act as template for others.

Congenital diaphragmatic hernia subtypes: Comparing birth prevalence, occurrence by maternal age, and mortality in a national birth cohort.

BACKGROUND: Population-based administrative data have rarely been used to compare the birth prevalence, risk factors for occurrence, and mortality of congenital diaphragmatic hernia (CDH) subtypes. OBJECTIVES: We used a national birth cohort to identify CDH subtypes and compared their birth prevalence, relationship with maternal age after accounting for sociodemographic factors, and 1-year mortality rates. METHODS: Linked hospital admission and death records were used to identify isolated and complex CDH cases (involving additional anomalies) among singleton livebirths in England between 2002 and 2018. The prevalence of each CDH subtype per 10,000 livebirths was estimated overall and by infant, birth and maternal characteristics. The relationship between maternal age and each subtype relative to no CDH was examined using multivariable log-binomial regression to estimate risk ratios (RRs). One-year mortality rates were examined using Kaplan-Meier curves and the hazard ratio (HR) of complex versus isolated CDH was calculated using Cox regression. RESULTS: Among 9.5 million livebirths, we identified 1285 with isolated CDH and 1150 with complex CDH. The overall prevalence of isolated and complex CDH was 1.4 (95% confidence interval [CI] 1.3, 1.4) and 1.2 (95% CI 1.1, 1.3) per 10,000 livebirths, respectively. Only complex CDH was associated with maternal age. Compared with maternal age 25-34 years, complex CDH risk was elevated for maternal age < 20 years (RR 1.31, 95% CI 1.00, 1.72). Risk was highest for maternal age ≥ 40 years (RR 1.61, 95% CI 1.21, 2.15) although accounting for chromosomal anomalies attenuated the risk (RR 1.39, 95% CI 1.00, 1.92). The 1-year mortality rate for complex CDH (33.1%, 95% CI 30.5, 35.9) was slightly higher than for isolated CDH (29.7%, 95% CI 27.3, 32.3) (HR 1.10, 95% CI 0.96, 1.27). CONCLUSIONS: Mechanisms of occurrence differed between and within CDH subtypes and 1-year mortality of complex CDH was slightly higher than for isolated CDH.

Could interventions on physical activity mitigate genomic liability for obesity? Applying the health disparity framework in genetically informed studies.

Polygenic scores (PGS) are now commonly available in longitudinal cohort studies, leading to their integration into epidemiological research. In this work, our aim is to explore how polygenic scores can be used as exposures in causal inference-based methods, specifically mediation analyses. We propose to estimate the extent to which the association of a polygenic score indexing genetic liability to an outcome could be mitigated by a potential intervention on a mediator. To do this this, we use the interventional disparity measure approach, which allows us to compare the adjusted total effect of an exposure on an outcome, with the association that would remain had we intervened on a potentially modifiable mediator. As an example, we analyse data from two UK cohorts, the Millennium Cohort Study (MCS, N = 2575) and the Avon Longitudinal Study of Parents and Children (ALSPAC, N = 3347). In both, the exposure is genetic liability for obesity (indicated by a PGS for BMI), the outcome is late childhood/early adolescent BMI, and the mediator and potential intervention target is physical activity, measured between exposure and outcome. Our results suggest that a potential intervention on child physical activity can mitigate some of the genetic liability for childhood obesity. We propose that including PGSs in a health disparity measure approach, and causal inference-based methods more broadly, is a valuable addition to the study of gene-environment interplay in complex health outcomes.

Factors associated with excess all-cause mortality in the first wave of the COVID-19 pandemic in the UK: A time series analysis using the Clinical Practice Research Datalink.

BACKGROUND: Excess mortality captures the total effect of the Coronavirus Disease 2019 (COVID-19) pandemic on mortality and is not affected by misspecification of cause of death. We aimed to describe how health and demographic factors were associated with excess mortality during, compared to before, the pandemic. METHODS AND FINDINGS: We analysed a time series dataset including 9,635,613 adults (≥40 years old) registered at United Kingdom general practices contributing to the Clinical Practice Research Datalink. We extracted weekly numbers of deaths and numbers at risk between March 2015 and July 2020, stratified by individual-level factors. Excess mortality during Wave 1 of the UK pandemic (5 March to 27 May 2020) compared to the prepandemic period was estimated using seasonally adjusted negative binomial regression models. Relative rates (RRs) of death for a range of factors were estimated before and during Wave 1 by including interaction terms. We found that all-cause mortality increased by 43% (95% CI 40% to 47%) during Wave 1 compared with prepandemic. Changes to the RR of death associated with most sociodemographic and clinical characteristics were small during Wave 1 compared with prepandemic. However, the mortality RR associated with dementia markedly increased (RR for dementia versus no dementia prepandemic: 3.5, 95% CI 3.4 to 3.5; RR during Wave 1: 5.1, 4.9 to 5.3); a similar pattern was seen for learning disabilities (RR prepandemic: 3.6, 3.4 to 3.5; during Wave 1: 4.8, 4.4 to 5.3), for black or South Asian ethnicity compared to white, and for London compared to other regions. Relative risks for morbidities were stable in multiple sensitivity analyses. However, a limitation of the study is that we cannot assume that the risks observed during Wave 1 would apply to other waves due to changes in population behaviour, virus transmission, and risk perception. CONCLUSIONS: The first wave of the UK COVID-19 pandemic appeared to amplify baseline mortality risk to approximately the same relative degree for most population subgroups. However, disproportionate increases in mortality were seen for those with dementia, learning disabilities, non-white ethnicity, or living in London.

Predictive model for long COVID in children 3 months after a SARS-CoV-2 PCR test.

BACKGROUND: To update and internally validate a model to predict children and young people (CYP) most likely to experience long COVID (i.e. at least one impairing symptom) 3 months after SARS-CoV-2 PCR testing and to determine whether the impact of predictors differed by SARS-CoV-2 status. METHODS: Data from a nationally matched cohort of SARS-CoV-2 test-positive and test-negative CYP aged 11-17 years was used. The main outcome measure, long COVID, was defined as one or more impairing symptoms 3 months after PCR testing. Potential pre-specified predictors included SARS-CoV-2 status, sex, age, ethnicity, deprivation, quality of life/functioning (five EQ-5D-Y items), physical and mental health and loneliness (prior to testing) and number of symptoms at testing. The model was developed using logistic regression; performance was assessed using calibration and discrimination measures; internal validation was performed via bootstrapping and the final model was adjusted for overfitting. RESULTS: A total of 7139 (3246 test-positives, 3893 test-negatives) completing a questionnaire 3 months post-test were included. 25.2% (817/3246) of SARS-CoV-2 PCR-positives and 18.5% (719/3893) of SARS-CoV-2 PCR-negatives had one or more impairing symptoms 3 months post-test. The final model contained SARS-CoV-2 status, number of symptoms at testing, sex, age, ethnicity, physical and mental health, loneliness and four EQ-5D-Y items before testing. Internal validation showed minimal overfitting with excellent calibration and discrimination measures (optimism-adjusted calibration slope: 0.96575; C-statistic: 0.83130). CONCLUSIONS: We updated a risk prediction equation to identify those most at risk of long COVID 3 months after a SARS-CoV-2 PCR test which could serve as a useful triage and management tool for CYP during the ongoing pandemic. External validation is required before large-scale implementation.

Parental feeding and childhood genetic risk for obesity: exploring hypothetical interventions with causal inference methods.

BACKGROUND: Parental-feeding behaviors are common intervention targets for childhood obesity, but often only deliver small changes. Childhood BMI is partly driven by genetic effects, and the extent to which parental-feeding interventions can mediate child genetic liability is not known. Here we aim to examine how potential interventions on parental-feeding behaviors can mitigate some of the association between child genetic liability and BMI in early adolescence, using causal inference methods. METHODS: Data from the Avon Longitudinal Study of Parents and Children were used to estimate an interventional disparity measure for a child polygenic score for BMI (PGS-BMI) on BMI at 12 years. The approach compares counterfactual outcomes for different hypothetical interventions on parental-feeding styles applied when children are 10-11 years (n = 4248). Results are presented as adjusted total association (Adj-Ta) between genetic liability (PGS-BMI) and BMI at 12 years, versus the interventional disparity measure-direct effect (IDM-DE), which represents the association that would remain, had we intervened on parental-feeding under different scenarios. RESULTS: For children in the top quintile of genetic liability, an intervention shifting parental feeding to the levels of children with lowest genetic risk, resulted in a difference of 0.81 kg/m2 in BMI at 12 years (Adj-Ta = 3.27, 95% CI: 3.04, 3.49; versus IDM-DE = 2.46, 95% CI: 2.24, 2.67). CONCLUSIONS: Findings suggest that parental-feeding interventions have the potential to buffer some of the genetic liability for childhood obesity. Further, we highlight a novel way to analyze potential interventions for health conditions only using secondary data analyses, by combining methodology from statistical genetics and social epidemiology.

Access to palivizumab among children at high risk of respiratory syncytial virus complications in English hospitals.

AIMS: Palivizumab is a monoclonal antibody which can prevent infection with respiratory syncytial virus (RSV). Due to its high cost, it is recommended for high-risk infants only. We aimed to determine the proportion of infants eligible for palivizumab treatment in England who receive at least one dose. METHODS: We used the Hospital Treatment Insights database, which contains hospital admission records linked to hospital pharmacy dispensing data for 43 out of 153 hospitals in England. Infants born between 2010 and 2016 were considered eligible for palivizumab if their medical records indicated chronic lung disease (CLD), congenital heart disease (CHD) or severe immunodeficiency (SCID), and they met additional criteria based on gestational age at birth and age at start of the RSV season (beginning of October). We calculated the proportion of infants who received at least one dose of palivizumab in their first RSV season, and modelled the odds of treatment according to multiple child characteristics using logistic regression models. RESULTS: We identified 3712 eligible children, of whom 2479 (67%) had complete information on all risk factors. Palivizumab was prescribed to 832 of eligible children (34%). Being born at <30 weeks' gestation, aged <6 months at the start of RSV season, and having two or more of CLD, CHD or SCID were associated with higher odds of treatment. CONCLUSION: In England, palivizumab is not prescribed to the majority of children who are eligible to receive it. Doctors managing these infants may be unfamiliar with the eligibility criteria or constrained by other considerations, such as cost.

Identification and prediction of novel classes of long-term disease trajectories for patients with juvenile dermatomyositis using growth mixture models.

OBJECTIVES: Uncertainty around clinical heterogeneity and outcomes for patients with JDM represents a major burden of disease and a challenge for clinical management. We sought to identify novel classes of patients having similar temporal patterns in disease activity and relate them to baseline clinical features. METHODS: Data were obtained for n = 519 patients, including baseline demographic and clinical features, baseline and follow-up records of physician's global assessment of disease (PGA), and skin disease activity (modified DAS). Growth mixture models (GMMs) were fitted to identify classes of patients with similar trajectories of these variables. Baseline predictors of class membership were identified using Lasso regression. RESULTS: GMM analysis of PGA identified two classes of patients. Patients in class 1 (89%) tended to improve, while patients in class 2 (11%) had more persistent disease. Lasso regression identified abnormal respiration, lipodystrophy and time since diagnosis as baseline predictors of class 2 membership, with estimated odds ratios, controlling for the other two variables, of 1.91 for presence of abnormal respiration, 1.92 for lipodystrophy and 1.32 for time since diagnosis. GMM analysis of modified DAS identified three classes of patients. Patients in classes 1 (16%) and 2 (12%) had higher levels of modified DAS at diagnosis that improved or remained high, respectively. Patients in class 3 (72%) began with lower DAS levels that improved more quickly. Higher proportions of patients in PGA class 2 were in DAS class 2 (19%, compared with 16 and 10%). CONCLUSION: GMM analysis identified novel JDM phenotypes based on longitudinal PGA and modified DAS.

Comprehensive analysis of the association of seasonal variability with maternal and neonatal nutrition in lowland Nepal.

OBJECTIVE: To provide a comprehensive seasonal analysis of pregnant mothers' eating behaviour and maternal/newborn nutritional status in an undernourished population from lowland rural Nepal, where weather patterns, agricultural labour, food availability and disease prevalence vary seasonally. DESIGN: Secondary analysis of cluster-randomised Low Birth Weight South Asia Trial data, applying cosinor analysis to predict seasonal patterns. OUTCOMES: Maternal mid-upper arm circumference (MUAC), BMI, dietary diversity, meals per day, eating down and food aversion in pregnancy (≥31 weeks' gestation) and neonatal z-scores of length-for-age (LAZ), weight-for-age (WAZ) and head circumference-for-age (HCAZ) and weight-for-length (WLZ). SETTING: Rural areas of Dhanusha and Mahottari districts in plains of Nepal. PARTICIPANTS: 2831 mothers aged 13-50 and 3330 neonates. RESULTS: We found seasonal patterns in newborn anthropometry and pregnant mothers' anthropometry, meal frequency, dietary diversity, food aversion and eating down. Seasonality in intake varied by food group. Offspring anthropometry broadly tracked mothers'. Annual amplitudes in mothers' MUAC and BMI were 0·27 kg/m2 and 0·22 cm, with peaks post-harvest and nadirs in October when food insecurity peaked. Annual LAZ, WAZ and WLZ amplitudes were 0·125, 0·159 and 0·411 z-scores, respectively. Neonates were the shortest but least thin (higher WLZ) in winter (December/January). In the hot season, WLZ was the lowest (May/June) while LAZ was the highest (March and August). HCAZ did not vary significantly. Food aversion and eating down peaked pre-monsoon (April/May). CONCLUSIONS: Our analyses revealed complex seasonal patterns in maternal nutrition and neonatal size. Seasonality should be accounted for when designing and evaluating public heath nutrition interventions.

Geospatial and seasonal variation of bronchiolitis in England: a cohort study using hospital episode statistics.

BACKGROUND: Rates of hospital admissions for bronchiolitis vary seasonally and geographically across England; however, seasonal differences by area remain unexplored. We sought to describe spatial variation in the seasonality of hospital admissions for bronchiolitis and its association with local demographic characteristics. METHODS: Singleton children born in English National Health Service hospitals between 2011 and 2016 (n=3 727 013) were followed up for 1 year. Poisson regression models with harmonic functions to model seasonal variations were used to calculate weekly incidence rates and peak timing of bronchiolitis admissions across English regions and clinical commissioning groups (CCGs). Linear regression was used to estimate the joint association of population density and deprivation with incidence and peak timing of bronchiolitis admissions at the CCG level. RESULTS: Bronchiolitis admission rates ranged from 30.9 per 1000 infant-years (95% CI 30.4 to 31.3) in London to 68.7 per 1000 (95% CI 67.9 to 69.5) in the North West. Across CCGs, there was a 5.3-fold variation in incidence rates and the epidemic peak ranged from week 49.3 to 52.2. Admission rates were positively associated with area-level deprivation. CCGs with earlier peak epidemics had higher population densities, and both high and low levels of deprivation were associated with earlier peak timing. CONCLUSIONS: Approximately one quarter of the variation in admission rates and two-fifths of the variation in peak timing of hospital admissions for bronchiolitis were explained by local demographic characteristics. Implementation of an early warning system could help to prepare hospitals for peak activity and to time public health messages.

Ethnic disparities in psychotic experiences explained by area-level syndemic effects.

BACKGROUND: Ethnic inequalities in health outcomes are often explained by socioeconomic status and concentrated poverty. However, ethnic disparities in psychotic experiences are not completely attenuated by these factors. AIMS: We investigated whether disparities are better explained by interactions between individual risk factors and place-based clustering of disadvantage, termed a syndemic. METHOD: We performed a cross-sectional survey of 3750 UK men, aged 18-34 years, oversampling Black and minority ethnic (BME) men nationally, together with men residing in London Borough of Hackney. Participants completed questionnaires covering psychiatric symptoms, substance misuse, crime and violence, and risky sexual health behaviours. We included five psychotic experiences and a categorical measure of psychosis based on the Psychosis Screening Questionnaire. RESULTS: At national level, more Black men reported psychotic experiences but disparities disappeared following statistical adjustment for social position. However, large disparities for psychotic experiences in Hackney were not attenuated by adjustment for social factors in Black men (adjusted odds ratio, 3.24; 95% CI 2.14-4.91; P < 0.002), but were for South Asian men. A syndemic model of joint effects, adducing a four-component latent variable (psychotic experiences and anxiety, substance dependence, high-risk sexual behaviour and violence and criminality) showed synergy between components and explained persistent disparities in psychotic experiences. A further interaction confirmed area-level effects (Black ethnicity × Hackney residence, 0.834; P < 0.001). CONCLUSIONS: Syndemic effects result in higher rates of non-affective psychosis among BME persons in certain inner-urban settings. Further research should investigate how syndemics raise levels of psychotic experiences and related health conditions in Black men in specific places with multiple deprivations.

Association between prenatal maternal infection and disordered eating behaviours in adolescence: a UK population-based prospective birth cohort study.

BACKGROUND: Prenatal infections have been proposed as a putative risk factor for a number of psychiatric outcomes across a continuum of severity. Evidence on eating disorders is scarce. We investigated whether exposure to prenatal maternal infections is associated with an increased risk of disordered eating and weight and shape concerns in adolescence in a large UK birth cohort. METHODS: We used data from the Avon Longitudinal Study of Parents and Children. The primary exposure was maternal experience of infections at any time in pregnancy. Study outcomes were presence of any, monthly or weekly disordered eating at 14 and 16 years of age, and weight and shape concerns at 14 years. We defined the causal effect of the exposure on these outcomes using a counterfactual framework adjusting our analyses for a number of hypothesised confounders, and imputing missing confounder data using multiple imputation. RESULTS: In total, 4884 children had complete exposure and outcome data at age 14 years, and 4124 at 16 years. Exposed children had a greater risk of reporting weekly disordered eating at both age 14 [risk difference (RD) 0.9%, 95% confidence interval (CI) -0.01 to 1.9, p = 0.08] and 16 (RD 2.3%, 95% CI 0.6-3.9, p < 0.01), though evidence of an association was weak at age 14 years. Exposed children also had greater weight and shape concerns at age 14 years (mean difference 0.15, 95% CI 0.05-0.26, p < 0.01). CONCLUSIONS: Exposure to prenatal maternal infection is associated with greater risk of disordered eating in adolescence. This association could be explained by in utero processes leading to impaired neurodevelopment or altered immunological profiles. Residual confounding cannot be excluded.

Longitudinal associations between circulating interleukin-6 and C-reactive protein in childhood, and eating disorders and disordered eating in adolescence.

OBJECTIVE: Few studies have explored the association between inflammation and eating disorders and none used a longitudinal design. We investigated the association between serum-levels of interleukin 6 (IL-6) and C-reactive protein (CRP) measured in childhood and eating disorders and related behaviours and cognitions in adolescence in a large general population sample. METHODS: We used data from the Avon Longitudinal Study of Parents and Children (ALSPAC). Our exposures were thirds of IL6 and CRP derived from serum measurements taken at age nine years, and outcomes were eating disorder diagnoses and self-reported disordered eating behaviours at ages 14, 16, and 18 years. We used univariable and multivariable multilevel logistic regression models adjusting for a number of potential confounders, including sex, fat mass, and pre-existing mental health difficulties. RESULTS: Our sample included 3480 children. Those in the top third of CRP had lower odds of binge eating (odds ratio(OR):0.62, 95% confidence interval (CI):0.39,1.00,p "equals" 0.05) and fasting (OR:0.63, 95% CI:0.38,1.07,p "equals" 0.09) after adjustment for confounders. We also observed weak associations of comparable magnitude for purging, anorexia nervosa, and bulimia nervosa. We did not find any associations between levels of IL6 and any of the outcomes under study. CONCLUSIONS: There was little evidence of an association between CRP and IL-6 and adolescent eating disorder outcomes. The inverse association observed between CRP and binge eating was unexpected, so caution is needed when interpreting it. One possible explanation is that higher CRP levels could have a protective role for disordered eating by affecting appetitive traits.

Formulating causal questions and principled statistical answers.

Although review papers on causal inference methods are now available, there is a lack of introductory overviews on what they can render and on the guiding criteria for choosing one particular method. This tutorial gives an overview in situations where an exposure of interest is set at a chosen baseline ("point exposure") and the target outcome arises at a later time point. We first phrase relevant causal questions and make a case for being specific about the possible exposure levels involved and the populations for which the question is relevant. Using the potential outcomes framework, we describe principled definitions of causal effects and of estimation approaches classified according to whether they invoke the no unmeasured confounding assumption (including outcome regression and propensity score-based methods) or an instrumental variable with added assumptions. We mainly focus on continuous outcomes and causal average treatment effects. We discuss interpretation, challenges, and potential pitfalls and illustrate application using a "simulation learner," that mimics the effect of various breastfeeding interventions on a child's later development. This involves a typical simulation component with generated exposure, covariate, and outcome data inspired by a randomized intervention study. The simulation learner further generates various (linked) exposure types with a set of possible values per observation unit, from which observed as well as potential outcome data are generated. It thus provides true values of several causal effects. R code for data generation and analysis is available on www.ofcaus.org, where SAS and Stata code for analysis is also provided.

Identifying typical trajectories in longitudinal data: modelling strategies and interpretations.

Individual-level longitudinal data on biological, behavioural, and social dimensions are becoming increasingly available. Typically, these data are analysed using mixed effects models, with the result summarised in terms of an average trajectory plus measures of the individual variations around this average. However, public health investigations would benefit from finer modelling of these individual variations which identify not just one average trajectory, but several typical trajectories. If evidence of heterogeneity in the development of these variables is found, the role played by temporally preceding (explanatory) variables as well as the potential impact of differential trajectories may have on later outcomes is often of interest. A wide choice of methods for uncovering typical trajectories and relating them to precursors and later outcomes exists. However, despite their increasing use, no practical overview of these methods targeted at epidemiological applications exists. Hence we provide: (a) a review of the three most commonly used methods for the identification of latent trajectories (growth mixture models, latent class growth analysis, and longitudinal latent class analysis); and (b) recommendations for the identification and interpretation of these trajectories and of their relationship with other variables. For illustration, we use longitudinal data on childhood body mass index and parental reports of fussy eating, collected in the Avon Longitudinal Study of Parents and Children.

Steady Growth in Early Infancy Is Associated with Greater Anthropometry in Indian Children Born Low Birth Weight at Term.

BACKGROUND: Patterns of early growth are associated with later body composition and risk of adult noncommunicable disease but information from low-income countries is limited. OBJECTIVES: The aim of this study was to investigate early growth trajectories and later anthropometric and bone density outcomes among children born term low birth weight (LBW: 1.8-2.5 kg). METHODS: We used data from 902 children from the Delhi Infant Vitamin D Supplementation study of LBW term infants (which collected monthly anthropometry from birth to 6 mo) and who had height, weight, midupper arm circumference (MUAC), midupper arm muscle circumference (MUAMC), subscapular and triceps skinfold thicknesses, tibia and radius bone density measured at age 4-6 y. We investigated how growth in the first 6 mo of life, modeled using the SuperImposition by Translation and Rotation (SITAR) growth curve model, was related to these outcomes. SITAR summarizes each infant's weight and length trajectory in terms of a population mean curve and child-specific growth parameters: size, timing, and intensity. These were included as explanatory variables in linear regression models for the childhood outcomes. RESULTS: Considering the infant weight and length SITAR parameters jointly, childhood weight was strongly associated with infant length timing [estimated regression coefficient ß = 0.25 (95% CI: 0.10, 0.39)] and with weight size, timing, and intensity [ß = 9.01 (6.75, 11.27), ß = -0.25 (-0.43, -0.07), ß = 5.03 (3.22, 6.84), respectively]. Childhood height was associated only with the length parameters [ß = 0.97 (0.71, 1.23), ß = -0.43 (-0.77, -0.09), ß = 11.68 (8.60, 14.75), respectively]; childhood MUAC, MUAMC, and skinfolds with all parameters; and bone density with none. Overall, delayed and sustained growth in infant weight and length resulted in higher values of all outcomes except bone density, with the period up to 15 wk of age appearing critical for setting childhood anthropometry in this population. CONCLUSIONS: The explanation for the effects of delayed growth and length of the period in which trajectories are set is unclear; however, sustained and delayed growth in early infancy appears to be beneficial for these LBW children at least in the short-term. The trial was registered at clinicaltrials.gov as BT/PR7489/PID/20/285/2006.

Adjusting for BMI in analyses of volumetric mammographic density and breast cancer risk.

BACKGROUND: Fully automated assessment of mammographic density (MD), a biomarker of breast cancer risk, is being increasingly performed in screening settings. However, data on body mass index (BMI), a confounder of the MD-risk association, are not routinely collected at screening. We investigated whether the amount of fat in the breast, as captured by the amount of mammographic non-dense tissue seen on the mammographic image, can be used as a proxy for BMI when data on the latter are unavailable. METHODS: Data from a UK case control study (numbers of cases/controls: 414/685) and a Norwegian cohort study (numbers of cases/non-cases: 657/61059), both with volumetric MD measurements (dense volume (DV), non-dense volume (NDV) and percent density (%MD)) from screening-age women, were analysed. BMI (self-reported) and NDV were taken as measures of adiposity. Correlations between BMI and NDV, %MD and DV were examined after log-transformation and adjustment for age, menopausal status and parity. Logistic regression models were fitted to the UK study, and Cox regression models to the Norwegian study, to assess associations between MD and breast cancer risk, expressed as odds/hazard ratios per adjusted standard deviation (OPERA). Adjustments were first made for standard risk factors except BMI (minimally adjusted models) and then also for BMI or NDV. OPERA pooled relative risks (RRs) were estimated by fixed-effect models, and between-study heterogeneity was assessed by the I2 statistics. RESULTS: BMI was positively correlated with NDV (adjusted r = 0.74 in the UK study and r = 0.72 in the Norwegian study) and with DV (r = 0.33 and r = 0.25, respectively). Both %MD and DV were positively associated with breast cancer risk in minimally adjusted models (pooled OPERA RR (95% confidence interval): 1.34 (1.25, 1.43) and 1.46 (1.36, 1.56), respectively; I2 = 0%, P >0.48 for both). Further adjustment for BMI or NDV strengthened the %MD-risk association (1.51 (1.41, 1.61); I2 = 0%, P = 0.33 and 1.51 (1.41, 1.61); I2 = 0%, P = 0.32, respectively). Adjusting for BMI or NDV marginally affected the magnitude of the DV-risk association (1.44 (1.34, 1.54); I2 = 0%, P = 0.87 and 1.49 (1.40, 1.60); I2 = 0%, P = 0.36, respectively). CONCLUSIONS: When volumetric MD-breast cancer risk associations are investigated, NDV can be used as a measure of adiposity when BMI data are unavailable.

Growth Trajectories, Breast Size, and Breast-Tissue Composition in a British Prebirth Cohort of Young Women.

Mammographic percent density, the proportion of fibroglandular tissue in the breast, is a strong risk factor for breast cancer, but its determinants in young women are unknown. We examined associations of magnetic resonance imaging (MRI) breast-tissue composition at age 21 years with prospectively collected measurements of body size and composition from birth to early adulthood and markers of puberty (all standardized) in a sample of 500 nulliparous women from a prebirth cohort of children born in Avon, United Kingdom, in 1991-1992 and followed up to 2011-2014. Linear models were fitted to estimate relative change in MRI percent water, which is equivalent to mammographic percent density, associated with a 1-standard-deviation increase in the exposure of interest. In mutually adjusted analyses, MRI percent water was positively associated with birth weight (relative change (RC) = 1.03, 95% confidence interval (CI): 1.00, 1.06) and pubertal height growth (RC = 1.07, 95% CI: 1.02, 1.13) but inversely associated with pubertal weight growth (RC = 0.86, 95% CI: 0.84, 0.89) and changes in dual-energy x-ray absorptiometry percent body fat mass (e.g., for change between ages 11 years and 13.5 years, RC = 0.96, 95% CI: 0.93, 0.99). Ages at thelarche and menarche were positively associated with MRI percent water, but these associations did not persist upon adjustment for height and weight growth. These findings support the hypothesis that growth trajectories influence breast-tissue composition in young women, whereas puberty plays no independent role.