Childhood factors related to diverging body mass index trajectories from childhood into mid-adulthood: A mixed methods study.

Body mass index (BMI) trajectories that improve over the lifecourse result in better cardiometabolic profiles, but only a small proportion of children of an unhealthy weight show improving BMI trajectories. This study aimed to examine the childhood factors related to diverging BMI trajectories from childhood into adulthood using data from the Childhood Determinants of Adult Health study. A convergent parallel mixed methods design was used. Quantitative data (n = 2206) came from the first (2004-06) and second (2009-11) adult follow-ups of 8498 Australian children (7-15 years) assessed in 1985. Using BMI z-scores, group-based trajectory modelling identified five trajectory groups: Persistently Low, Persistently Average, High Decreasing, Average Increasing and High Increasing. Qualitative data (n = 50) were collected from a sub-group (2016; 38-46 years). Semi-structured interviews with 6-12 participants from each BMI trajectory group focused on individual, social and environmental influences on weight, diet and physical activity across the lifecourse. Log multinomial regression modelling estimated relative risks of trajectory group membership across childhood demographic, behavioural, health, parental and school factors. Qualitative data were thematically analysed using a constant comparative approach. Childhood factors influenced BMI trajectories. Paternal education, main language spoken, alcohol and self-rated health were significant quantitative childhood predictors of BMI trajectory. A distinct 'legacy effect' of parental lifestyle influences during childhood was apparent among interview participants in the Stable and High Decreasing groups, a strong and mostly positive concept discussed by both men and women in these groups and persisting despite phases of unhealthy behaviours. In contrast, the 'legacy effect' was much weaker in the two Increasing BMI groups. This study is the first to simultaneously identify important quantitative and qualitative childhood factors related to divergent BMI trajectories, and to observe a legacy effect of parents' lifestyle behaviours on divergent BMI trajectories. This work provides direction for further exploration of the factors driving divergent BMI trajectories.

Social-ecological predictors of physical activity patterns: A longitudinal study of women from socioeconomically disadvantaged areas.

Limited longitudinal evidence of the predictors of physical activity (PA) patterns over time exists, particularly among high-risk groups such as women living in socioeconomically disadvantaged areas. This study aimed to: 1) describe leisure-time PA (LTPA) and transport-related PA (TRPA) patterns over time; and 2) identify individual, social and physical environmental predictors of LTPA and TRPA patterns over five years. Baseline (2007-08) data were collected and analysed (2016-18) from n = 4349 women (18-46 years) from disadvantaged areas of Victoria, Australia. Three- and five-year follow-up data were collected in 2010-11 (n = 1912) and 2012 (n = 1560). LTPA and TRPA were self-reported using the International Physical Activity Questionnaire, and patterns categorised as consistently low, persistently increasing, persistently decreasing, or inconsistent. Compared to a consistently low LTPA pattern, greater family support predicted both persistent decreases (odds ratio [OR] 1.20, 95% CI 1.05-1.36) and persistent increases (OR 1.17, 95% CI 1.04-1.32) in LTPA, while access to childcare predicted inconsistent LTPA patterns (OR 1.66, 95% CI 1.03-2.65). For both LTPA and TRPA, PA enjoyment predicted persistent increases (LTPA: OR 1.05, 95% CI 1.02-1.10; TRPA: OR 1.03, 95% CI 1.00-1.07), persistent decreases (LTPA: OR 1.04, 95% CI 1.00-1.08; TRPA OR 1.04, 95% CI 0.99-1.08), and inconsistent patterns (LTPA: OR 1.04, 95% CI 1.02-1.07; TRPA: OR 1.03, 95% CI 1.01-1.06). Although directionality was inconsistent, and the magnitude of effects were small, PA enjoyment, family social support for PA and access to childcare warrant further investigation and consideration as potentially key factors impacting PA patterns among women living in socioeconomically disadvantaged areas.

Summary goodness-of-fit statistics for binary generalized linear models with noncanonical link functions.

Generalized linear models (GLM) with a canonical logit link function are the primary modeling technique used to relate a binary outcome to predictor variables. However, noncanonical links can offer more flexibility, producing convenient analytical quantities (e.g., probit GLMs in toxicology) and desired measures of effect (e.g., relative risk from log GLMs). Many summary goodness-of-fit (GOF) statistics exist for logistic GLM. Their properties make the development of GOF statistics relatively straightforward, but it can be more difficult under noncanonical links. Although GOF tests for logistic GLM with continuous covariates (GLMCC) have been applied to GLMCCs with log links, we know of no GOF tests in the literature specifically developed for GLMCCs that can be applied regardless of link function chosen. We generalize the Tsiatis GOF statistic originally developed for logistic GLMCCs, (TG), so that it can be applied under any link function. Further, we show that the algebraically related Hosmer-Lemeshow (HL) and Pigeon-Heyse (J(2) ) statistics can be applied directly. In a simulation study, TG, HL, and J(2) were used to evaluate the fit of probit, log-log, complementary log-log, and log models, all calculated with a common grouping method. The TG statistic consistently maintained Type I error rates, while those of HL and J(2) were often lower than expected if terms with little influence were included. Generally, the statistics had similar power to detect an incorrect model. An exception occurred when a log GLMCC was incorrectly fit to data generated from a logistic GLMCC. In this case, TG had more power than HL or J(2) .