The Hitchhiker’s Guide to Longitudinal Models: A Primer on Model Selection for Repeated-Measures Methods

Longitudinal data is becoming increasingly available in developmental neuroimaging. To maximize the promise of this wealth of information on how biology, behavior, and cognition change over time, there is a need to incorporate broad and rigorous training in longitudinal methods into the repertoire of developmental neuroscientists. Fortunately, these models have an incredibly rich tradition in the broader developmental sciences that we can draw from. Here, we provide a primer on longitudinal models, written in a beginner-friendly (and slightly irreverent) manner, with a particular focus on selecting among different modeling frameworks (e.g., multilevel versus latent curve models) to build the theoretical model of development a researcher wishes to test. Our aims are three-fold: 1) lay out a heuristic framework for longitudinal model selection, 2) build a repository of references that ground each model in its tradition of methodological development and practical implementation with a focus on connecting researchers to resources outside traditional neuroimaging journals, and 3) provide practical resources in the form of a codebook companion demonstrating how to fit these models. These resources together aim to enhance training for the next generation of developmental neuroscientists by providing a solid foundation for future forays into advanced modeling applications.

Obtaining Interpretable Parameters From Reparameterized Longitudinal Models: Transformation Matrices Between Growth Factors in Two Parameter Spaces

This study proposes transformation functions and matrices between coefficients in the original and reparameterized parameter spaces for an existing linear-linear piecewise model to derive the interpretable coefficients directly related to the underlying change pattern. Additionally, the study extends the existing model to allow individual measurement occasions and investigates predictors for individual differences in change patterns. We present the proposed methods with simulation studies and a real-world data analysis. Our simulation study demonstrates that the method can generally provide an unbiased and accurate point estimate and appropriate confidence interval coverage for each parameter. The empirical analysis shows that the model can estimate the growth factor coefficients and path coefficients directly related to the underlying developmental process, thereby providing meaningful interpretation.

HBV co-infection is associated with persistently elevated liver stiffness measurement in HIV-positive adults: A 6-year single-centre cohort study in Nigeria

Background In Nigeria, the effect of Hepatitis B virus (HBV) on long-term liver outcomes in persons with HIV (PLH) has not been described. We determined changes in liver stiffness measure (LSM) using transient elastography over 6 years in HIV mono-infected and HIV-HBV co-infected Nigerians initiating antiretroviral therapy (ART) and factors associated with LSM decline. Methods This single centre, cohort study enrolled ART-naïve HIV mono- and HIV-HBV co-infected adults (≥18 years) at the APIN Public Health Initiatives–supported HIV Care and Treatment Centre at Jos University Teaching Hospital, Nigeria, from 7/2011 to 2/2012. LSM at baseline, Years 3 and 6 were analysed using longitudinal models to estimate changes over time and their predictors. Results Data from 100 (31%) HIV-HBV co-infected and 225 (69%) HIV mono-infected participants were analysed. Median LSM at baseline was 6.10 (IQR: 4.60–7.90) kPa in co-infected and 5.10 (IQR: 4.40–6.10) kPa in mono-infected participants. In adjusted analyses, average LSM was not significantly different between Year 0 and 3 (β = 0.02, −0.22 to 0.26, p = 0.87 and Year 0 and 6 (β = −0.02, −0.23 to 0.27, p = 0.88) in both groups ( p>0.05), but co-infected participants had significantly higher LSM than mono-infected throughout follow-up (β = 0.018, 0.019–0.28, p < 0.001). Year 3 LSM differed according to ART initiation status by Year 3 (initiators - non-initiators: −0.87, −1.70 to −0.29). Conclusion In this cohort, LSM remained higher among HIV-HBV co-infected versus HIV mono-infected participants throughout follow-up. Our findings emphasize the continuing need for monitoring of liver outcomes in HIV-HBV co-infected populations on ART and the importance of preventing HBV infection among PLH to optimize liver health.

Age varying polygenic effects on alcohol use in African Americans and European Americans from adolescence to adulthood

Genetic effects on alcohol use can vary over time but are often examined using longitudinal models that predict a distal outcome at a single time point. The vast majority of these studies predominately examine effects using White, European American (EA) samples or examine the etiology of genetic variants identified from EA samples in other racial/ethnic populations, leading to inconclusive findings about genetic effects on alcohol use. The current study examined how genetic influences on alcohol use varied by age across a 15 year period within a diverse ethnic/racial sample of adolescents. Using a multi-ethnic approach, polygenic risk scores were created for African American (AA, n = 192) and EA samples (n = 271) based on racially/ethnically aligned genome wide association studies. Age-varying associations between polygenic scores and alcohol use were examined from age 16 to 30 using time-varying effect models separately for AA and EA samples. Polygenic risk for alcohol use was found to be associated with alcohol use from age 22–27 in the AA sample and from age 24.50 to 29 in the EA sample. Results are discussed relative to the intersection of alcohol use and developmental genetic effects in diverse populations.

These are not the effects you are looking for: Causality and the within-/between-person distinction in longitudinal data analysis

In psychological science, researchers often pay particular attention to the distinction between within- and between-person relationships in longitudinal data analysis. Here, we aim to clarify the relationship between the within- and between-person distinction and causal inference, and show that the distinction is informative but does not play a decisive role for causal inference. Our main points are threefold. First, within-person data are not necessary for causal inference; for example, between-person experiments can inform us about (average) causal effects. Second, within-person data are not sufficient for causal inference; for example, time-varying confounders can lead to spurious within-person associations. Finally, despite not being sufficient, within-person data can be tremendously helpful for causal inference. We provide pointers to help readers navigate the more technical literature on longitudinal models, and conclude with a call for more conceptual clarity: Instead of letting statistical models dictate which substantive questions we ask, we should start with well-defined theoretical estimands which in turn determine both study design and data analysis.

Application of Multivariate Generalized Linear Mixed Model to Identify Effect of Dialysate Temperature on Physiologic Indicators among Hemodialysis Patients

Introduction: One of the complications of hemodialysis treatment is hypotension, which can increase morbidity and mortality and compromise dialysis efficacy. Dialysate temperature is an important factor that contributes to hemodynamic stability during hemodialysis. This study investigated the effect of dialysate temperature on the patients' blood pressure and pulse rate. Model-based approaches were used to produce more reliable results compared with traditional methods. Methods: A total of 30 patients were studied during 9 dialysis sessions. Dialysate temperatures were 37° C, 36° C and 35° C. A joint longitudinal model was used to analyze both responses of blood pressure and pulse rate, simultaneously. Results: The results showed that low-dialysate temperature was not significantly associated with higher systolic blood pressure (p>0.05) or a higher pulse rate (p>0.05) either during or after dialysis. Pulse rate and blood pressure were higher for women during dialysate (p<0.001). However, increasing age was associated with higher blood pressure and a lower pulse rate (p<0.001). Conclusion: Using several separate, repeated measure analysis of variances may produce misleading results, when there is more than one response variable measured over time, Multivariate statistical methods (including joint longitudinal models), should be used.

Routinely collected antenatal data for longitudinal prediction of preeclampsia in nulliparous women: a population-based study

The objective was to evaluate the sequentially updated predictive capacity for preeclampsia during pregnancy, using multivariable longitudinal models including data from antenatal care. This population-based cohort study in the Stockholm-Gotland Counties, Sweden, included 58,899 pregnancies of nulliparous women 2008–2013. Prospectively collected data from each antenatal care visit was used, including maternal characteristics, reproductive and medical history, and repeated measurements of blood pressure, weight, symphysis-fundal height, proteinuria, hemoglobin and blood glucose levels. We used a shared-effects joint longitudinal model including all available information up until a given gestational length (week 24, 28, 32, 34 and 36), to update preeclampsia prediction sequentially. Outcome measures were prediction of preeclampsia, preeclampsia with delivery < 37, and preeclampsia with delivery ≥ 37 weeks’ gestation. The area under the curve (AUC) increased with gestational length. AUC for preeclampsia with delivery < 37 weeks’ gestation was 0.73 (95% CI 0.68–0.79) at week 24, and increased to 0.87 (95% CI 0.84–0.90) in week 34. For preeclampsia with delivery ≥ 37 weeks’ gestation, the AUC in week 24 was 0.65 (95% CI 0.63–0.68), but increased to 0.79 (95% CI 0.78–0.80) in week 36. The addition of routinely collected clinical measurements throughout pregnancy improve preeclampsia prediction and may be useful to individualize antenatal care.

Adverse Childhood Experiences, Family Social Capital, and Externalizing Behavior Problems: An Analysis Across Multiple Ecological Levels

Some research on childhood adversity is critiqued for emphasizing the experiences of white, middle/upper-middle-class people and not accounting for adversities faced by more diverse populations. Adversities are also often summed up in ways that are unhelpful for targeting interventions to reduce risk of poor outcomes. I examine adversities across ecological levels—child, parent, family, and neighborhood—to determine the risk of externalizing behavior problems (EBP) using a racially diverse sample from the Longitudinal Studies of Child Abuse and Neglect ( N = 1058). I consider whether family social capital can offset the effects of adversity across ecological levels. Longitudinal models indicate that adversities across multiple levels and those at the child, parent, and neighborhood levels increase risk of EBP throughout childhood. Cross-sectional models yield that early family social capital is associated with fewer EBP for children with multiple levels of adversity and at the child, parent, family, and neighborhood levels.