Marginal structural models for repeated measures where intercept and slope are correlated: An application exploring the benefit of nutritional supplements on weight gain in HIV-infected children initiating antiretroviral therapy

Estimating causal effects requires important prior subject-matter knowledge and, sometimes, sophisticated statistical tools. The latter is especially true when targeting the causal effect of a time-varying exposure in a longitudinal study. Marginal structural models are a relatively new class of causal models that effectively deal with the estimation of the effects of time-varying exposures. Marginal structural models have traditionally been embedded in the counterfactual framework to causal inference. In this paper, we use the causal graph framework to enhance the implementation of marginal structural models. We illustrate our approach using data from a prospective cohort study, the Honolulu Heart Program. These data consist of 8006 men at baseline. To illustrate our approach, we focused on the estimation of the causal effect of physical activity on blood pressure, which were measured at three time points. First, a causal graph is built to encompass prior knowledge. This graph is then validated and improved utilizing structural equation models. We estimated the aforementioned causal effect using marginal structural models for repeated measures and guided the implementation of the models with the causal graph. By employing the causal graph framework, we also show the validity of fitting conditional marginal structural models for repeated measures in the context implied by our data.

Download Full-text

Faculty Opinions recommendation of Marginal structural models to control for time-varying confounding in occupational and environmental epidemiology.

Faculty Opinions – Post-Publication Peer Review of the Biomedical Literature ◽

10.3410/f.718095221.793482998 ◽

2013 ◽

Author(s):

Sherwood Burge

Keyword(s):

Structural Models ◽

Environmental Epidemiology ◽

Marginal Structural Models ◽

Time Varying

Download Full-text

Annual rhythms in adults’ lifestyle and health (ARIA): protocol for a 12-month longitudinal study examining temporal patterns in weight, activity, diet, and wellbeing in Australian adults

BMC Public Health ◽

10.1186/s12889-020-10054-3 ◽

2021 ◽

Vol 21 (1) ◽

Author(s):

Rachel G. Curtis ◽

Timothy Olds ◽

François Fraysse ◽

Dorothea Dumuid ◽

Gilly A. Hendrie ◽

...

Keyword(s):

Physical Activity ◽

Body Weight ◽

Weight Gain ◽

Repeated Measures ◽

Vigorous Physical Activity ◽

Activity Patterns ◽

Intervention Strategy ◽

Temporal Patterns ◽

Trial Registration ◽

Multi Level

Abstract Background Almost one in three Australian adults are now obese, and the rate continues to rise. The causes of obesity are multifaceted and include environmental, cultural and lifestyle factors. Emerging evidence suggests there may be temporal patterns in weight gain related, for example, to season and major festivals such as Christmas, potentially due to changes in diet, daily activity patterns or both. The aim of this study is to track the annual rhythm in body weight, 24 h activity patterns, dietary patterns, and wellbeing in a cohort of Australian adults. In addition, through data linkage with a concurrent children’s cohort study, we aim to examine whether changes in children’s body mass index, activity and diet are related to those of their parents. Methods A community-based sample of 375 parents aged 18 to 65 years old, residing in or near Adelaide, Australia, and who have access to a Bluetooth-enabled mobile device or a computer and home internet, will be recruited. Across a full year, daily activities (minutes of moderate to vigorous physical activity, light physical activity, sedentary behaviour and sleep) will be measured using wrist-worn accelerometry (Fitbit Charge 3). Body weight will be measured daily using Fitbit wifi scales. Self-reported dietary intake (Dietary Questionnaire for Epidemiological Studies V3.2), and psychological wellbeing (WHOQOL-BREF and DASS-21) will be assessed eight times throughout the 12-month period. Annual patterns in weight will be examined using Lowess curves. Associations between changes in weight and changes in activity and diet compositions will be examined using repeated measures multi-level models. The associations between parent’s and children’s weight, activity and diet will be investigated using multi-level models. Discussion Temporal factors, such as day type (weekday or weekend day), cultural celebrations and season, may play a key role in weight gain. The aim is to identify critical opportunities for intervention to assist the prevention of weight gain. Family-based interventions may be an important intervention strategy. Trial registration Australia New Zealand Clinical Trials Registry, identifier ACTRN12619001430123. Prospectively registered on 16 October 2019.

Download Full-text