Search

Testing for Granger causality relies on estimating the capacity of dynamics in one time series to forecast dynamics in another. The canonical test for such temporal predictive causality is based on fitting multivariate time series models and is cast in the classical null hypothesis testing framework. In this framework, we are limited to rejecting the null hypothesis or failing to reject the null - we can never validly accept the null hypothesis of no Granger causality. This is poorly suited for many common purposes, including evidence integration, feature selection, and other cases where it is useful to express evidence against, rather than for, the existence of an association. Here we derive and implement the Bayes factor for Granger causality in a multilevel modeling framework. This Bayes factor summarizes information in the data in terms of a continuously scaled evidence ratio between the presence of Granger causality and its absence. We also introduce this procedure for the multilevel generalization of Granger causality testing. This facilitates inference when information is scarce or noisy or if we are interested primarily in population-level trends. We illustrate our approach with an application on exploring causal relationships in affect using a daily life study.

Researchers may have at their disposal the raw data of the studies they wish to meta-analyze. The goal of this study is to identify, illustrate, and compare a range of possible analysis options for researchers to whom raw data are available, wanting to fit a structural equation model (SEM) to these data. This study illustrates techniques that directly analyze the raw data, such as multilevel and multigroup SEM, and techniques based on summary statistics, such as correlation-based meta-analytical structural equation modeling (MASEM), discussing differences in procedures, capabilities, and outcomes. This is done by analyzing a previously published collection of datasets using open source software. A path model reflecting the theory of planned behavior is fitted to these datasets using different techniques involving SEM. Apart from differences in handling of missing data, the ability to include study-level moderators, and conceptualization of heterogeneity, results show differences in parameter estimates and standard errors across methods. Further research is needed to properly formulate guidelines for applied researchers looking to conduct individual participant data MASEM., (© 2024 The Author(s). Research Synthesis Methods published by John Wiley & Sons Ltd.)

Background: Causal mediation analysis is widespread in applied medical research, especially in longitudinal settings. However, estimating natural mediational effects in such contexts is often difficult because of the presence of post-treatment confounding. Moreover, many models frequently used in applied research, like multilevel and latent growth models, present an additional difficulty, i.e. the presence of latent variables. In this paper, we propose a causal interpretation of these two classes of models based on a novel type of causal effects called separable, which overcome some of the issues of natural effects., Methods: We formally derive conditions for the identifiability of separable mediational effects and their analytical expressions based on the g-formula. We carry out a simulation study to investigate how moderate and severe model misspecification, as well as violation of the identfiability assumptions, affect estimates. We also present an application to real data., Results: The results show how model misspecification impacts the estimates of mediational effects, particularly in the case of severe misspecification, and that the bias worsens over time. The violation of assumptions affects separable effect estimates in a very different way for the mixed effect and the latent growth models., Conclusion: Our approach allows us to give multilevel and latent growth models an appealing causal interpretation based on separable effects. The simulation study shows that model misspecification can heavily impact effect estimates, highlighting the importance of careful model choice., (© 2024. The Author(s).)

Multilevel structural equation modeling (MSEM) is a statistical framework of major relevance for research concerned with people's intrapersonal dynamics. An application domain that is rapidly gaining relevance is the study of individual differences in the within-person association (WPA) of variables that fluctuate over time. For instance, an individual's social reactivity - their emotional response to social situations - can be represented as the association between repeated measurements of the individual's social interaction quantity and momentary well-being. MSEM allows researchers to investigate the associations between WPAs and person-level outcome variables (e.g., life satisfaction) by specifying the WPAs as random slopes in the structural equation on level 1 and using the latent representations of the slopes to predict outcomes on level 2. Here, we are concerned with the case in which a researcher is interested in nonlinear effects of WPAs on person-level outcomes - a U-shaped effect of a WPA, a moderation effect of two WPAs, or an effect of congruence between two WPAs - such that the corresponding MSEM includes latent interactions between random slopes. We evaluate the nonlinear MSEM approach for the three classes of nonlinear effects (U-shaped, moderation, congruence) and compare it with three simpler approaches: a simple two-step approach, a single-indicator approach, and a plausible values approach. We use a simulation study to compare the approaches on accuracy of parameter estimates and inference. We derive recommendations for practice and provide code templates and an illustrative example to help researchers implement the approaches., (© 2024. The Author(s).)

We present the R package galamm, whose goal is to provide common ground between structural equation modeling and mixed effect models. It supports estimation of models with an arbitrary number of crossed or nested random effects, smoothing splines, mixed response types, factor structures, heteroscedastic residuals, and data missing at random. Implementation using sparse matrix methods and automatic differentiation ensures computational efficiency. We here briefly present the implemented methodology, give an overview of the package and an example demonstrating its use.

In many studies in the social and behavioral sciences, the data have a multilevel structure, with subjects nested within clusters. In the design phase of such a study, the number of clusters to achieve a desired power level has to be calculated. This requires a priori estimates of the effect size and intraclass correlation coefficient. If these estimates are incorrect, the study may be under- or overpowered. This may be overcome by using a group-sequential design, where interim tests are done at various points in time of the study. Based on interim test results, a decision is made to either include additional clusters or to reject the null hypothesis and conclude the study. This contribution introduces Bayesian sequential designs as an alternative to group-sequential designs. This approach compares various hypotheses based on the support in the data for each of them. If neither hypothesis receives a sufficient degree of support, additional clusters are included in the study and the Bayes factor is recalculated. This procedure continues until one of the hypotheses receives sufficient support. This paper explains how the Bayes factor is used as a measure of support for a hypothesis and how a Bayesian sequential design is conducted. A simulation study in the setting of a two-group comparison was conducted to study the effects of the minimum and maximum number of clusters per group and the desired degree of support. It is concluded that Bayesian sequential designs are a flexible alternative to the group sequential design., (© 2023. The Psychonomic Society, Inc.)

Continuous-time modeling using differential equations is a promising technique to model change processes with longitudinal data. Among ways to fit this model, the Latent Differential Structural Equation Modeling (LDSEM) approach defines latent derivative variables within a structural equation modeling (SEM) framework, thereby allowing researchers to leverage advantages of the SEM framework for model building, estimation, inference, and comparison purposes. Still, a few issues remain unresolved, including performance of multilevel variations of the LDSEM under short time lengths (e.g., 14 time points), particularly when coupled multivariate processes and time-varying covariates are involved. Additionally, the possibility of using Bayesian estimation to facilitate the estimation of multilevel LDSEM (M-LDSEM) models with complex and higher-dimensional random effect structures has not been investigated. We present a series of Monte Carlo simulations to evaluate three possible approaches to fitting M-LDSEM, including: frequentist single-level and two-level robust estimators and Bayesian two-level estimator. Our findings suggested that the Bayesian approach outperformed other frequentist approaches. The effects of time-varying covariates are well recovered, and coupling parameters are the least biased especially using higher-order derivative information with the Bayesian estimator. Finally, an empirical example is provided to show the applicability of the approach.

Multilevel modeling (MLM) is commonly used in psychological research to model clustered data. However, data in applied research usually violate one of the essential assumptions of MLM-homogeneity of variance. While the fixed-effect estimates produced by the maximum likelihood method remain unbiased, the standard errors for the fixed effects are misestimated, resulting in inaccurate inferences and inflated or deflated type I error rates. To correct the bias in fixed effects standard errors and provide valid inferences, small-sample corrections such as the Kenward-Roger (KR) adjustment and the adjusted cluster-robust standard errors (CR-SEs) with the Satterthwaite approximation for t tests have been used. The current study compares KR with random slope (RS) models and the adjusted CR-SEs with ordinary least squares (OLS), random intercept (RI) and RS models to analyze small, heteroscedastic, clustered data using a Monte Carlo simulation. Results show the KR procedure with RS models has large biases and inflated type I error rates for between-cluster effects in the presence of level 2 heteroscedasticity. In contrast, the adjusted CR-SEs generally yield results with acceptable biases and maintain type I error rates close to the nominal level for all examined models. Thus, when the interest is only in within-cluster effect, any model with the adjusted CR-SEs could be used. However, when the interest is to make accurate inferences of the between-cluster effect, researchers should use the adjusted CR-SEs with RS to have higher power and guard against unmodeled heterogeneity. We reanalyzed an example in Snijders & Bosker (2012) to demonstrate the use of the adjusted CR-SEs with different models., (© 2024. The Author(s).)

Across different fields of research, the similarities and differences between various longitudinal models are not always eminently clear due to differences in data structure, application area, and terminology. Here we propose a comprehensive model framework that will allow simple comparisons between longitudinal models, to ease their empirical application and interpretation. At the within-individual level, our model framework accounts for various attributes of longitudinal data, such as growth and decline, cyclical trends, and the dynamic interplay between variables over time. At the between-individual level, our framework contains continuous and categorical latent variables to account for between-individual differences. This framework encompasses several well-known longitudinal models, including multilevel regression models, growth curve models, growth mixture models, vector-autoregressive models, and multilevel vector-autoregressive models. The general model framework is specified and its key characteristics are illustrated using famous longitudinal models as concrete examples. Various longitudinal models are reviewed and it is shown that all these models can be united into our comprehensive model framework. Extensions to the model framework are discussed. Recommendations for selecting and specifying longitudinal models are made for empirical researchers who aim to account for between-individual differences. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

In multilevel models, disaggregating predictors into level-specific parts (typically accomplished via centering) benefits parameter estimates and their interpretations. However, the importance of level-specificity has been sparsely addressed in multilevel literature concerning collinearity. In this study, we develop novel insights into the interactivity of centering and collinearity in multilevel models. After integrating the broad literatures on centering and collinearity, we review level-specific and conflated correlations in multilevel data. Next, by deriving formal relationships between predictor collinearity and multilevel model estimates, we demonstrate how the consequences of collinearity change across different centering specifications and identify data characteristics that may exacerbate or mitigate those consequences. We show that when all or some level-1 predictors are uncentered, slope estimates can be greatly biased by collinearity. Disaggregation of all predictors eliminates the possibility that fixed effect estimates will be biased due to collinearity alone; however, under some data conditions, collinearity is associated with biased standard errors and random effect (co)variance estimates. Finally, we illustrate the importance of disaggregation for diagnosing collinearity in multilevel data and provide recommendations for the use of level-specific collinearity diagnostics. Overall, the necessity of disaggregation for identifying and managing collinearity's consequences in multilevel models is clarified in novel ways.

Single-case experimental designs (SCEDs) are used to study the effects of interventions on the behavior of individual cases, by making comparisons between repeated measurements of an outcome under different conditions. In research areas where SCEDs are prevalent, there is a need for methods to synthesize results across multiple studies. One approach to synthesis uses a multilevel meta-analysis (MLMA) model to describe the distribution of effect sizes across studies and across cases within studies. However, MLMA relies on having accurate sampling variances of effect size estimates for each case, which may not be possible due to auto-correlation in the raw data series. One possible solution is to combine MLMA with robust variance estimation (RVE), which provides valid assessments of uncertainty even if the sampling variances of effect size estimates are inaccurate. Another possible solution is to forgo MLMA and use simpler, ordinary least squares (OLS) methods with RVE. This study evaluates the performance of effect size estimators and methods of synthesizing SCEDs in the presence of auto-correlation, for several different effect size metrics, via a Monte Carlo simulation designed to emulate the features of real data series. Results demonstrate that the MLMA model with RVE performs properly in terms of bias, accuracy, and confidence interval coverage for estimating overall average log response ratios. The OLS estimator corrected with RVE performs the best in estimating overall average Tau effect sizes. None of the available methods perform adequately for meta-analysis of within-case standardized mean differences. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Introduction: Childhood stunting is a global public health concern, associated with both short and long-term consequences, including high child morbidity and mortality, poor development and learning capacity, increased vulnerability for infectious and non-infectious disease. The prevalence of stunting varies significantly throughout Ethiopian regions. Therefore, this study aimed to assess the geographical variation in predictors of stunting among children under the age of five in Ethiopia using 2019 Ethiopian Demographic and Health Survey., Method: The current analysis was based on data from the 2019 mini Ethiopian Demographic and Health Survey (EDHS). A total of 5,490 children under the age of five were included in the weighted sample. Descriptive and inferential analysis was done using STATA 17. For the spatial analysis, ArcGIS 10.7 were used. Spatial regression was used to identify the variables associated with stunting hotspots, and adjusted R2 and Corrected Akaike Information Criteria (AICc) were used to compare the models. As the prevalence of stunting was over 10%, a multilevel robust Poisson regression was conducted. In the bivariable analysis, variables having a p-value < 0.2 were considered for the multivariable analysis. In the multivariable multilevel robust Poisson regression analysis, the adjusted prevalence ratio with the 95% confidence interval is presented to show the statistical significance and strength of the association., Result: The prevalence of stunting was 33.58% (95%CI: 32.34%, 34.84%) with a clustered geographic pattern (Moran's I = 0.40, p<0.001). significant hotspot areas of stunting were identified in the west and south Afar, Tigray, Amhara and east SNNPR regions. In the local model, no maternal education, poverty, child age 6-23 months and male headed household were predictors associated with spatial variation of stunting among under five children in Ethiopia. In the multivariable multilevel robust Poisson regression the prevalence of stunting among children whose mother's age is >40 (APR = 0.74, 95%CI: 0.55, 0.99). Children whose mother had secondary (APR = 0.74, 95%CI: 0.60, 0.91) and higher (APR = 0.61, 95%CI: 0.44, 0.84) educational status, household wealth status (APR = 0.87, 95%CI: 0.76, 0.99), child aged 6-23 months (APR = 1.87, 95%CI: 1.53, 2.28) were all significantly associated with stunting., Conclusion: In Ethiopia, under-five children suffering from stunting have been found to exhibit a spatially clustered pattern. Maternal education, wealth index, birth interval and child age were determining factors of spatial variation of stunting. As a result, a detailed map of stunting hotspots and determinants among children under the age of five aid program planners and decision-makers in designing targeted public health measures., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Seifu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Propensity score (PS) analyses are increasingly popular in behavioral sciences. Two issues often add complexities to PS analyses, including missing data in observed covariates and clustered data structure. In previous research, methods for conducting PS analyses with considering either issue alone were examined. In practice, the two issues often co-occur; but the performance of methods for PS analyses in the presence of both issues has not been evaluated previously. In this study, we consider PS weighting analysis when data are clustered and observed covariates have missing values. A simulation study is conducted to evaluate the performance of different missing data handling methods (complete-case, single-level imputation, or multilevel imputation) combined with different multilevel PS weighting methods (fixed- or random-effects PS models, inverse-propensity-weighting or the clustered weighting, weighted single-level or multilevel outcome models). The results suggest that the bias in average treatment effect estimation can be reduced, by better accounting for clustering in both the missing data handling stage (such as with the multilevel imputation) and the PS analysis stage (such as with the fixed-effects PS model, clustered weighting, and weighted multilevel outcome model). A real-data example is provided for illustration.

Accelerated longitudinal designs allow researchers to efficiently collect longitudinal data covering a time span much longer than the study duration. One important assumption of these designs is that each cohort (a group defined by their age of entry into the study) shares the same longitudinal trajectory. Although previous research has examined the impact of violating this assumption when each cohort is defined by a single age of entry, it is possible that each cohort is instead defined by a range of ages, such as groups that experience a particular historical event. In this paper we examined how including cohort membership in linear and quadratic multilevel models performed in detecting and controlling for cohort effects in this scenario. Using a Monte Carlo simulation study, we assessed the performance of this approach under conditions related to the number of cohorts, the overlap between cohorts, the strength of the cohort effect, the number of affected parameters, and the sample size. Our results indicate that models including a proxy variable for cohort membership based on age at study entry performed comparably to using true cohort membership in detecting cohort effects accurately and returning unbiased parameter estimates. This indicates that researchers can control for cohort effects even when true cohort membership is unknown.

Count outcomes are frequently encountered in single-case experimental designs (SCEDs). Generalized linear mixed models (GLMMs) have shown promise in handling overdispersed count data. However, the presence of excessive zeros in the baseline phase of SCEDs introduces a more complex issue known as zero-inflation, often overlooked by researchers. This study aimed to deal with zero-inflated and overdispersed count data within a multiple-baseline design (MBD) in single-case studies. It examined the performance of various GLMMs (Poisson, negative binomial [NB], zero-inflated Poisson [ZIP], and zero-inflated negative binomial [ZINB] models) in estimating treatment effects and generating inferential statistics. Additionally, a real example was used to demonstrate the analysis of zero-inflated and overdispersed count data. The simulation results indicated that the ZINB model provided accurate estimates for treatment effects, while the other three models yielded biased estimates. The inferential statistics obtained from the ZINB model were reliable when the baseline rate was low. However, when the data were overdispersed but not zero-inflated, both the ZINB and ZIP models exhibited poor performance in accurately estimating treatment effects. These findings contribute to our understanding of using GLMMs to handle zero-inflated and overdispersed count data in SCEDs. The implications, limitations, and future research directions are also discussed., (© 2024. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply.)

In Bayesian statistics, the most widely used criteria of Bayesian model assessment and comparison are Deviance Information Criterion (DIC) and Watanabe-Akaike Information Criterion (WAIC). We use a multilevel mediation model as an illustrative example to compare different types of DIC and WAIC. More specifically, we aim to compare the performance of conditional and marginal DICs and WAICs, and investigate their performance with missing data. We focus on two versions of DIC ( D I C 1 and D I C 2 ) and one version of WAIC. In addition, we explore whether it is necessary to include the nuisance models of incomplete exogenous variables in likelihood. Based on the simulation results, whether D I C 2 is better than D I C 1 and WAIC and whether we should include the nuisance models of exogenous variables in likelihood functions depend on whether we use marginal or conditional likelihoods. Overall, we find that the marginal likelihood based- D I C 2 that excludes the likelihood of covariate models generally had the highest true model selection rates., (© 2023. The Psychonomic Society, Inc.)

In psychology and other fields, data often have a cross-classified structure, whereby observations are nested within multiple types of non-hierarchical clusters (e.g., repeated measures cross-classified by persons and stimuli). This paper discusses ways that, in cross-classified multilevel models, slopes of lower-level predictors can implicitly reflect an ambiguous blend of multiple effects (for instance, a purely observation-level effect as well as a unique between-cluster effect for each type of cluster). The possibility of conflating multiple effects of lower-level predictors is well recognized for non-cross-classified multilevel models, but has not been fully discussed or clarified for cross-classified contexts. Consequently, in published cross-classified modeling applications, this possibility is almost always ignored, and researchers routinely specify models that conflate multiple effects. In this paper, we show why this common practice can be problematic, and show how to disaggregate level-specific effects in cross-classified models. We provide a novel suite of options that include fully cluster-mean-centered, partially cluster-mean-centered, and contextual effect models, each of which provides a unique interpretation of model parameters. We further clarify how to avoid both fixed and random conflation, the latter of which is widely misunderstood even in non-cross-classified models. We provide simulation results showing the possible deleterious impact of such conflation in cross-classified models, and walk through pedagogical examples to illustrate the disaggregation of level-specific effects. We conclude by considering additional model complexities that can arise with cross-classification, providing guidance for researchers in choosing among model specifications, and describing newly available software to aid researchers who wish to disaggregate effects in practice., (© 2023. The Psychonomic Society, Inc.)

Introduction: Coronavirus 2019 (COVID- 19) is an acute respiratory viral infectious disease in human being caused by RNA virus that belonged to the family of corona virus. The incidence of this disease was growing exponentially and affects millions of the world population that leads to expose thousands of peoples for death. Thus, this study was targeted to assess the practice of people on COVID-19 infections prevention strategies in the region., Methods: A community based cross sectional study design was conducted in Benishangul Gumuz Region from May 25 -June 15, 2020. Multistage sampling technique was carried out to select 21 kebeles/ketena and 422 study participants. Data were collected by face to face interview using semi-structured questionnaires. The collected data were entered, cleaned and analyzed using STATA software version 14.0. Descriptive, bi-variable and multivariable multilevel models were applied. Variables with p value < 0.25 in bi-variable analysis were selected as candidates for multivariable analysis. Finally, the variables with p-value < 0.5 were considered as statistically significant, then variables with odds ratio, 95% CI were used to interpret the effect of association., Results: The magnitude of good practice on prevention strategies of COVID- 19 infections was 62.1%. The most frequently practiced prevention strategies for COVID-19 infections were hand washing with water and soap (80.7%), alcohol-based hand rub (68.8%), maintaining social/physical distance (74.2%) and avoiding touching eyes. Individual and community level factors that affecting practice of COVID- 19 infection prevention strategies were discovered. Hence, community level factor was place of origin (AOR = 0.1; 95%CI: 0.03, 0.35) whereas individual level factors were able to read and write (AOR = 0.18; 95%CI: 0.04, 0.81) and being merchant (AOR = 2.07; 95%CI: 1.01, 4.28)., Conclusion: The level of practice of community towards COVID-19 infections prevention strategies were low as compared with the expected outcome. Individual and community level factors were identified. This implies that social mobilization and community engagement was not effective. Thus, designing appropriate strategies to improve of practice prevention strategies are strongly recommend., Competing Interests: The authors have declared that no competing interests exist.

Introduction: Anemia among children aged 6-59 months remains a major public health problem in low-and high-income countries including Ethiopia. Anemia is associated with significant consequences on the health of children such as under-five morbidity and mortality, increased risk of infection, and poor academic performance. The prevalence of anemia in Ethiopia has varied across areas. Therefore, this study aimed to investigate the geographic weighted regression analysis of anemia and its associated factors among children aged 6-59 months in Ethiopia., Methods: This study was based on the 2016 Ethiopian Demographic and Health Survey (EDHS) data. A total weighted sample of 8482 children aged 6-59 months was included. For the spatial analysis, Arc-GIS version 10.7 and SaTScan version 9.6 statistical software were used. Spatial regression was done to identify factors associated with the hotspots of anemia and model comparison was based on adjusted R2 and Corrected Akaike Information Criteria (AICc). For the associated factors, the multilevel robust Poisson regression was fitted since the prevalence of anemia was greater than 10%. Variables with a p-value < 0.2 in the bi-variable analysis were considered for the multivariable analysis. In the multivariable multilevel robust Poisson regression analysis, the adjusted prevalence ratio with the 95% confidence interval was reported to declare the statistical significance and strength of association., Results: The prevalence of anemia among children aged 6-59 months was 57.56% (95%CI: 56.50%, 58.61%) with significant spatial variation across regions in Ethiopia. The significant hot spot areas of anemia among children aged 6-59 months were detected in the central, west, and east Afar, Somali, Dire Dawa, Harari, and northwest Gambella regions. Mothers who had anemia, a child aged 23-59 months, mothers aged 15-19 years, and coming from a household with a poorer or poorest household were significant predictors of the spatial variations of anemia among children aged 6-59 months. In the multilevel robust Poisson analysis, born to mothers aged 30-39 (APR = 0.84, 95% CI: 0.76, 0.92) and 40-49 years (APR = 0.73, 95% CI: 0.65, 0.83), mothers who didn't have formal education (APR = 1.10, 95% CI: 1.00, 1.20), Children in the poorest household wealth index (APR = 1.17, 95% CI: 1.06, 1.29), being 4-6 (APR = 1.08, 95% CI: 1.02, 1.13) and above 6 order of birth (APR = 1.15, 95% CI: 1.07, 1.23), children born to anemic mothers (APR = 1.24, 95% CI: 1.19, 1.29), children aged 24-59 months (APR = 0.70, 95% CI: 0.68, 0.73), stunted children (APR = 1.09, 95% CI: 1.04, 1.13) and underweight children (APR = 1.07, 95% CI: 1.03, 1.13) were significantly associated with anemia among children aged 6-59 months., Conclusion and Recommendation: Anemia is still a public health problem for children in Ethiopia. Residing in a geographic area where a high proportion of children born to mothers aged 15-19 years, a child aged 6-23 months, coming from a household with poorer or poorest wealth index, and mothers with anemia increased the risk of experiencing anemia among children aged 6-59 months. Maternal education, maternal age, child age, household wealth, stunting, underweight, birth order, and maternal anemia were significant predictors of anemia among children. The detailed map of anemia hot spots among children aged 6-59 months and its predictors could assist program planners and decision-makers to design targeted public health interventions., Competing Interests: The authors have declared that no competing interests exist.

Background: Knowledge of the ovulatory cycle (KOC) aids women to refrain and engage in sexual intercourse to avoid and to get pregnancy, respectively. The effect of community-level factors on KOC was not yet known in Ethiopia. Therefore, we aimed to investigate the community- and individual-level determinants of KOC among women of childbearing age., Methods: We used the 2016 Ethiopian Demographic and Health Survey, and total weighted samples of 15,683 women were included. Intra-class correlation, median odds ratio, and deviance were executed for model comparison in which a model with the lowest deviance was the best model i.e. model III in this case. A multivariable multilevel logistic regression model was employed to identify community- and individual-level factors of correct KOC. In the ultimate model, an adjusted odds ratio (AOR) with a 95% confidence interval was reported and variables with a p<0.05 were considered as statistically significant., Results: In this study, 3,698 [23.58% (95% CI; 22.92-24.25)] participants had correct KOC. Women's age in years, i.e. 20-24 (AOR = 1.46;1.28-1.68) 25-29 (AOR = 1.72; 1.49-1.99), 30-34 (AOR = 2.21; 1.89-2.58), 35-39 (AOR = 1.78; 1.51-2.09), 40-44 (AOR = 1.97; 1.65-2.37), and 45-49 (AOR = 1.78; 1.44-2.19), knowledge of contraceptive methods (AOR = 3.08; 2.07-4.58), increased women's educational level, i.e. higher (AOR = 4.24; 3.54-5.07), secondary (AOR = 2.89; 2.48-3.36), and primary (AOR = 1.57; 1.39-1.78), higher household's wealth index, i.e. richest (AOR = 1.71; 1.35-2.16), richer (AOR = 1.42; 1.16-1.72), middle (AOR = 1.29; 1.07-1.56), and poorer (AOR = 1.24; 1.03-1.48), current contraceptive use (AOR = 1.26; 1.13-1.39), menstruating in the last six weeks (AOR = 1.13; 1.03-1.24), women's media exposure (AOR = 1.20; 1.07-1.35), and being in the community with a high level of media exposure (AOR = 1.53; 1.24-1.88) were statistically significant with KOC., Conclusions: Knowledge of the ovulatory cycle was low in this study, which demands health education for women of childbearing age. Special attention should be given to teenagers, those with lower educational, and lower economic status. Besides, the strengthening of media campaigns could increase women's KOC, which is crucial for preventing unintended pregnancy., Competing Interests: The authors declared that no competing interests exist.

Background: The National Institute on Aging, in conjunction with the Department of Health and Human Services as part of the National Alzheimer's Project Act, hosted a 2020 Dementia Care, Caregiving, and Services Research Summit Virtual Meeting Series on August 13, 2020. This article reflects three presentations related to Theme 6: Research Resources, Methods, and Data Infrastructure. Dr. Bynum discussed the challenges of identifying people for population- and healthcare-based research, including how definitions of dementia have changed over time, the opportunities and challenges inherent in the use of electronic data sources, and the need to fit data collection strategies to research goals and questions. Dr. Travison provided an overview on the growing use of embedded pragmatic clinical trials (ePCTs) and how to enhance their impact in dementia research. Dr. Wendler presented on the ethical considerations relevant to consent for dementia research, including assessment of decisional capacity and the role of decisional surrogates., Conclusions: The availability of claims data, electronic health records, and other sources of "existing" data has made the use and development of ePCTs both easier and more appealing. Among other things, they offer advantages in terms of lower cost and generalizability to real-world settings. This is turn has necessitated the use of informatic and analytic approaches to account for some of the limitations and complexities of such data, including multilevel clustering and the need to link and jointly analyze data from the person with dementia and those of their care partner. As part of this process, it will be important to broaden the scope of who is assessed for decisional capacity, make those assessments more study specific, and assist surrogates in making decisions based on what the individual would have chosen for themselves if capacitated (i.e., substituted judgment)., (© 2021 The American Geriatrics Society.)

Salmonella enterica serovar Enteritidis is a major cause of foodborne Salmonella infections and outbreaks in humans. Effective surveillance and timely outbreak detection are essential for public health control. Multilevel genome typing (MGT) with multiple levels of resolution has been previously demonstrated as a promising tool for this purpose. In this study, we developed MGT with nine levels for S . Enteritidis and characterised the genomic epidemiology of S . Enteritidis in detail. We examined 26 670 publicly available S . Enteritidis genome sequences from isolates spanning 101 years from 86 countries to reveal their spatial and temporal distributions. Using the lower resolution MGT levels, globally prevalent and regionally restricted sequence types (STs) were identified; avian associated MGT4-STs were found that were common in human cases in the USA; temporal trends were observed in the UK with MGT5-STs from 2014 to 2018 revealing both long lived endemic STs and the rapid expansion of new STs. Using MGT3 to MGT6, we identified multidrug resistance (MDR) associated STs at various MGT levels, which improves precision of detection and global tracking of MDR clones. We also found that the majority of the global S . Enteritidis population fell within two predominant lineages, which had significantly different propensity of causing large scale outbreaks. An online open MGT database has been established for unified international surveillance of S . Enteritidis. We demonstrated that MGT provides a flexible and high-resolution genome typing tool for S . Enteritidis surveillance and outbreak detection.

Background: This study is aimed to analyze the prognostic factors affecting the short-term efficacy of non-surgical treatment of patients in periodontitis from stage II to stage IV by the multilevel modeling analysis., Materials and Methods: A total of 58 patients with chronic periodontitis were included in this study. Patients were clinically explored before and 3 months after the treatment and the difference in probing depth was determined [Reduction of probing depth (Δ PD) = baseline PD - finial probing depth (FPD)] which is considered as the therapeutic evaluation. Three different levels were analyzed: patients, teeth and sites to construct a multi-layer linear model., Results: Probing depth (PD) improved significantly compared with that before treatment (p < 0.05), in which FPD was (3.90 ± 1.39) mm, and the ΔPD was (1.79 ± 0.97) mm. Compared with the mesial sites and distal sites of the multi-rooted teeth, the number of PD ≥ 5 mm or PD < 5 mm after the treatment was significantly different (P < 0.05), and the proportion of PD < 5 mm was higher in mesial sites. The null model showed that Δ PD varied greatly between groups at various levels (P < 0.001), with prediction variable of site level, tooth level, and patient level accounted for 66%, 18%, and 16% of the overall difference, respectively. The complete model showed that the Δ PD of smokers was significantly lower than that of non-smokers (P < 0.001). The Δ PD of the mesial and distal sites was larger than that of the buccolingual central site (P < 0.001). The Δ PD of single-rooted teeth was larger than that of multi-rooted teeth (P < 0.001). The baseline PD, tooth mobility (TM), bleeding index (BI), clinical attachment loss (CAL) were significantly negatively correlated with Δ PD (P < 0.001)., Conclusions: Patients with periodontitis from stage II to stage IV, who were non-smoking, have good compliance, good awareness of oral health, and low percentage sites with PD ≥ 5 mm at baseline, single-rooted teeth with hypomobility, less clinical attachment loss and lower bleeding index and sites of mesial or distal can obtain an ideal short-term efficacy of non-surgical treatment.

The classical twin model can be reparametrized as an equivalent multilevel model. The multilevel parameterization has underexplored advantages, such as the possibility to include higher-level clustering variables in which lower levels are nested. When this higher-level clustering is not modeled, its variance is captured by the common environmental variance component. In this paper we illustrate the application of a 3-level multilevel model to twin data by analyzing the regional clustering of 7-year-old children's height in the Netherlands. Our findings show that 1.8%, of the phenotypic variance in children's height is attributable to regional clustering, which is 7% of the variance explained by between-family or common environmental components. Since regional clustering may represent ancestry, we also investigate the effect of region after correcting for genetic principal components, in a subsample of participants with genome-wide SNP data. After correction, region no longer explained variation in height. Our results suggest that the phenotypic variance explained by region might represent ancestry effects on height.

For more than a decade, it has been known that many common behavior genetics models for a single phenotype can be estimated as multilevel models (e.g., van den Oord 2001; Guo and Wang 2002; McArdle and Prescott 2005; Rabe-Hesketh et al. 2007). This paper extends the current knowledge to (1) multiple phenotypes such that the method is completely general to the variance structure hypothesized, and (2) both higher and lower levels of nesting. The multi-phenotype method also allows extended relationships to be considered (see also, Bard et al. 2012; Hadfield and Nakagawa 2010). The extended relationship model can then be continuously expanded to merge with the case typically seen in the molecular genetics analyses of unrelated individuals (e.g., Yang et al. 2011). We use the multilevel form of behavior genetics models to fit a multivariate three level model that allows for (1) child level variation from unique environments and additive genetics, (2) family level variation from additive genetics and common environments, and (3) neighborhood level variation from broader geographic contexts. Finally, we provide R (R Development Core Team 2020) functions and code for multilevel specification of several common behavior genetics models using OpenMx (Neale et al. 2016).

Brazil has been the benchmark for volleyball performance for at least two decades, providing a unique context to examine expertise development. This study examined the variation in body size, functional capacities, motivation for achievement, competitiveness, and deliberate practice of youth volleyball players associated with differences in biological maturity status, chronological age, and accumulated deliberate volleyball practice, adopting a Bayesian multilevel modeling approach. We considered 68 female and 94 male adolescent players (14.2 years, 90% confidence interval: 12.7 to 16.0). Players were grouped by the onset of deliberate volleyball practice as related to biologic maturation milestones [pre-puberty deliberate practice onset (12% of the sample), mid-puberty deliberate practice onset (51% of the sample), and late-puberty deliberate practice onset (37% of the sample). There was substantial variation in body dimensions and functional performance by gender. There was no variation by gender for motivation for deliberate practice and motivation for achievement and competitiveness. The young volleyball players appeared to be highly motivated and committed to deliberate practice, achievement, and competitiveness. Alignment of chronological age, biological maturation, and accumulated training experience allow more in-depth insights into young volleyball players' development, providing sounder support for coaches´ decisions., Competing Interests: The authors have declared that no competing interests exist.

Objectives: To measure the effects of peer influence and peer selection on drinking behavior in adolescence through a rigorous statistical approach designed to unravel these interrelated processes., Methods: We conducted systematic searches of electronic databases, thesis collections and conference proceedings to identify studies that used longitudinal network design and stochastic actor-oriented modeling to analyze drinking behavior in adolescents. Parameter estimates collected from individual studies were analyzed using multilevel random-effects models., Results: We identified 26 articles eligible for meta-analysis. Meta-analyses for different specifications of the peer influence effect were conducted separately. The peer influence effect was positive for every specification: for average similarity (avSim) mean log odds ratio was 1.27 with 95% confidence interval [0.04; 2.49]; for total similarity (totSim) 0.46 (95% CI = [0.44; 0.48]), and for average alter (avAlt) 0.70 (95% CI = [-0.01; 1.41]). The peer selection effect (simX) was also positive: 0.46 (95% CI = [0.28; 0.63]). Conversion log odds ratio values to Cohen's d gives estimates from 0.25 to 0.70, which is considered as medium to large effect., Conclusions: Advances in methodology for social network analysis have made it possible to accurately estimate peer influence effects free from peer selection effects. More research is necessary to clarify the roles of age, gender, and individual susceptibility on the changing behavior of adolescents under the influence of their peers. Understanding the effects of peer influence should inform practitioners and policy makers to design and deliver more effective prevention programs., Competing Interests: The authors have declared that no competing interests exist.

Introduction: Returning to health facility for postnatal care (PNC) use after giving birth at health facility could reflect the health seeking behavior of mothers. However, such studies are rare though they are critically important to develop vigorous strategies to improve PNC service utilization. Therefore, this study aimed to determine the magnitude and factors associated with returning to health facilities for PNC among mothers who delivered in Ethiopian health facilities after they were discharged., Methods: This cross-sectional study used 2016 Ethiopian Demographic and Health Survey data. A total of 2405mothers who gave birth in a health facility were included in this study. Multilevel mixed-effect logistic regression model was fitted to estimate both independent (fixed) effects of the explanatory variables and community-level (random) effects on return for PNC utilization. Variable with p-value of ≤ 0.25 from unadjusted multilevel logistic regression were selected to develop three models and p-value of ≤0.05 was used to declare significance of the explanatory variables on the outcome variable in the final (adjusted) model. Analysis was done using IBM SPSS statistics version 21., Result: In this analysis, from the total 2405 participants, 14.3% ((95%CI: 12.1-16.8), (n = 344)) of them returned to health facilities for PNC use after they gave birth at a health facility. From the multilevel logistic regression analysis, being employed (AOR = 1.51, 95%CI: 1.04-2.19), receiving eight and above antenatal care visits (AOR = 2.90, 95%CI: 1.05-8.00), caesarean section delivery (AOR = 2.53, 95%CI: 1.40-4.58) and rural residence (AOR = 0.56, 95%CI: 0.36-0.88) were found significantly associated with return to health facilities for PNC use among women who gave birth at health facility., Conclusion: Facility-based PNC utilization among mothers who delivered at health facilities is low in Ethiopia. Both individual and community level variables were determined women to return to health facilities for PNC use. Thus, adopting context-specific strategies/policies could improve PNC utilization and should be paid a due focus., Competing Interests: The authors have declared that no competing interests exist.

Objectives: The aim of this study was to assess patient safety culture (PSC) in maternal and child health (MCH) institutions in China and its individual, organizational, and regional variations., Methods: Using the PSC survey for MCH institutions (PSCS-MCHI), 2021 valid respondents from 25 participating institutions were investigated in three regions (Beijing, Zhejiang, and Jiangxi) of China. Patient safety culture and its subscale scores (1-5) and factors associated with PSC as revealed by multilevel modeling., Results: The respondents had an average PSC score of 3.55 (SD = 0.35), with subscale scores ranging between 2.46 ("staffing and workload") and 4.02 ("work commitment"). There were limited regional differences in PSC: a three-level regression model was only confirmed for the subscale "staff empowerment" (P = 0.006). However, significant organizational variations in PSC were evident: a two-level regression model was assumed for the PSC scale and nine subscales (P < 0.001). The fixed-effect models showed that male respondents, frontline workers, those who were in their mid-career (11-20 y), overloaded (≥9 hours), and had a masters or higher degree reported worse PSC. Frontline workers were less positive than managers in ratings on "managerial response to risks" (-0.11 [-0.20 to -0.02]), "management support" (-0.18 [-0.28 to -0.07]), and "staff empowerment" (-0.23[-0.35 to -0.11])., Conclusions: Patient safety culture in MCH institutions is shaped by organizational and individual characteristics. We observed a gap in perceived PSC between frontline worker, who are less positive, and managers. Actions for improving PSC should consider interventions on organizational management (such as appropriate staffing and workload management) and engagement of frontline workers in the development of management and training activities.

In this study, we examine the concepts of spatial dependence and spatial heterogeneity in the effect of macro-level and micro-level factors on stunting among children aged under five in Uganda. We conducted a cross-sectional analysis of 3624 Ugandan children aged under five, using data from the 2016 Ugandan Demographic and Health Survey. Multilevel mixed-effect analysis, spatial regression methods and multi-scale geographically weight regression (MGWR) analysis were employed to examine the association between our predictors and stunting as well as to analyse spatial dependence and variability in the association. Approximately 28% of children were stunted. In the multilevel analysis, the effect of drought, diurnal temperature and livestock per km 2 on stunting was modified by child, parent and household factors. Likewise, the contextual factors had a modifiable effect on the association between child's sex, mother's education and stunting. The results of the spatial regression models indicate a significant spatial error dependence in the residuals. The MGWR suggests rainfall and diurnal temperature had spatial varying associations with stunting. The spatial heterogeneity of rainfall and diurnal temperature as predictors of stunting suggest some areas in Uganda might be more sensitive to variability in these climatic conditions in relation to stunting than others.

Psychometric models for longitudinal test scores typically estimate quantities associated with single-administration tests, like ability at each time-point. However, models for longitudinal tests have not considered opportunities to estimate new quantities that are unavailable from single-administration tests. Specifically, we discuss dynamic measurement models - which combine aspects of longitudinal IRT, nonlinear growth models, and dynamic assessment - to directly estimate capacity, defined as the expected future score once the construct has fully developed. After discussing the history and connecting these areas into a single framework, we apply the model to verbal test scores from the Intergenerational Studies, which follow 494 people from 3 to 72 years old. The goal is to predict adult verbal scores (Age ≥ 34) from adolescent scores (Age ≤ 20). We held-out the adult data for prediction and compared predictions from traditional longitudinal IRT ability scores and proposed dynamic measurement capacity scores from models fit to the adolescent data. Results showed that the R 2 from capacity scores were 2.5 times larger than the R 2 from longitudinal IRT ability scores (43% vs. 16%), providing some evidence that exploring new quantities available from longitudinal testing could be worthwhile when an interest in testing is forecasting future performance.

The study investigated adult outpatient Health Psychology Services appointment attendance, cancellation, and missed appointments (A/C/M). The first objective was to determine which demographic and process factors predicted the probability of A/C/M. The second objective was to determine whether there remained residual significant differences in A/C/M between therapists (i.e., a "therapist effect"), after controlling for explanatory variables. We conducted a practice-based retrospective 2-year cohort study. Three-level multilevel models were constructed and tested to analyze the probability of A/C/M at (1) assessment appointments ( N = 1,175) and (2) follow-up appointments ( N = 5,441). After controlling for predictor variables, significant therapist effects were found for attendance (10.0% to 13.0%) and cancellation (4.4%) at follow-up appointments (but not assessments), indicating significantly different attendance rates at follow-up between therapists. Predictors of attendance at follow-up included patient age, pretherapy symptom severity scores (including risk and symptom scores), and completion of intake questionnaires. Early morning follow-up appointments were least likely to be canceled, followed by late afternoon and finally midday appointments. Treatment intensity predicted attendance, but among qualified therapists, qualification type and pay level were nonsignificant. No significant predictors of attendance at assessment were detected. Attendance at Health Psychology Services outpatient appointments varies significantly according to patient, therapist, and appointment factors. Key routinely collected variables are predictive of attendance at follow-up. Clinical implications include the potential to identify patients at risk of nonattendance and target engagement interventions to these patients. Research directions include closer examination of variability in follow-up attendance between therapists. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Objectives: Patient safety events offer opportunities to improve patient care, but, unfortunately, events often go unreported. Although some barriers to event reporting can be reduced with electronic reporting systems, insight on organizational and cultural factors that influence reporting frequency may help hospitals increase reporting rates and improve patient safety. The purpose of this study was to evaluate the associations between dimensions of patient safety culture and perceived reporting practices of safety events of varying severity., Methods: We conducted a cross-sectional survey study using previously collected data from The Agency for Healthcare Research and Quality Hospital Survey of Patient Safety Culture as predictors and outcome variables. The dataset included health-care professionals in U.S. hospitals, and data were analyzed using multilevel modeling techniques., Results: Data from 223,412 individuals, 7816 work areas/units, and 967 hospitals were analyzed. Whether examining near miss, no harm, or potential for harm safety events, the dimension feedback about error accounted for the most unique predictive variance in the outcome frequency of events reported. Other significantly associated variables included organizational learning, nonpunitive response to error, and teamwork within units (all P < 0.001). As the perceived severity of the safety event increased, more culture dimensions became significantly associated with voluntary reporting., Conclusions: To increase the likelihood that a patient safety event will be voluntarily reported, our study suggests placing priority on improving event feedback mechanisms and communication of event-related improvements. Focusing efforts on these aspects may be more efficient than other forms of culture change.

In intervention studies having multiple outcomes, researchers often use a series of univariate tests (e.g., ANOVAs) to assess group mean differences. Previous research found that this approach properly controls Type I error and generally provides greater power compared to MANOVA, especially under realistic effect size and correlation combinations. However, when group differences are assessed for a specific outcome, these procedures are strictly univariate and do not consider the outcome correlations, which may be problematic with missing outcome data. Linear mixed or multivariate multilevel models (MVMMs), implemented with maximum likelihood estimation, present an alternative analysis option where outcome correlations are taken into account when specific group mean differences are estimated. In this study, we use simulation methods to compare the performance of separate independent samples t tests estimated with ordinary least squares and analogous t tests from MVMMs to assess two-group mean differences with multiple outcomes under small sample and missingness conditions. Study results indicated that a MVMM implemented with restricted maximum likelihood estimation combined with the Kenward-Roger correction had the best performance. Therefore, for intervention studies with small N and normally distributed multivariate outcomes, the Kenward-Roger procedure is recommended over traditional methods and conventional MVMM analyses, particularly with incomplete data.

The child mortality rate is an essential measurement of socioeconomic growth and the quality of life in Ethiopia which is one among the six countries that account for half of the global under-five deaths. Therefore, this study aimed to identify the potential risk factors for child mortality in Ethiopia. Data for the study was drawn from the Ethiopian Demographic and Health Survey data conducted in 2016. A two-part random effects regression model was employed to identify the associated predictors of child mortality. The study found that 53.3% of mothers did not face any child death, while 46.7% lost at least one. Vaccinated child (IRR = 0.735, 95%CI: 0.647, 0.834), were currently using contraceptive (IRR = 0.885, 95%CI: 0.814, 0.962), who had antenatal care visit four or more times visit (IRR = 0.841, 95%CI: 0.737,0.960), fathers whose level of education is secondary or above(IRR = 0.695, 95%CI: 0.594, 0.814), mothers who completed their primary school(IRR = 0.785, 95%CI: 0.713, 0.864), mothers who have birth interval greater than 36 months (IRR = 0.728, 95%CI: 0.676, 0.783), where the age of the mother at first birth is greater than 16 years(IRR = 0.711, 95%CI: 0.674, 0.750) associated with the small number of child death. While multiple births (IRR = 1.355, 95%CI: 1.249, 1.471, four and above birth order (IRR = 1.487, 95%CI: 1.373, 1.612) and had working father (IRR = 1.125, 95%CI: 1.049, 1.206) associated with a higher number of child death. The variance components for the random effects showed significant variation of child mortality between enumeration areas. Policies and programs aimed at addressing enumeration area variations in child mortality need to be formulated and their implementation must be strongly pursued. Efforts are also needed to extend educational programmers aimed at educating mothers on the benefits of the antenatal checkup before first birth, spacing their birth interval, having their child vaccinated, and selecting a safe place of delivery to reduce child mortality., Competing Interests: The authors have declared that no competing interests exist.

Introduction: Despite the desire of adolescent girls and young women (AGYW) in sub-Saharan Africa (SSA) to use contraceptives, the majority of them have challenges with access to contraceptive services. This is more evident in high fertility countries in SSA. The purpose of this study was to examine the predictors of unmet need for contraception among AGYW in selected high fertility countries in SSA., Materials and Methods: Data from current Demographic and Health Surveys (DHS) carried out between 2010 and 2018 in 10 countries in SSA were analysed. A sample size of 24,898 AGYW who were either married or cohabiting was used. Unmet need for contraception was the outcome variable in this study. The explanatory variables were age, marital status, occupation, educational level, frequency of reading newspaper/magazine, frequency of listening to radio, frequency of watching television and parity (individual level variables) and wealth quintile, sex of household head, place of residence and decision-maker in healthcare (household/community level variables). Descriptive and multilevel logistic regression analyses were carried out. The results of the multilevel logistic regression analyses were reported using adjusted odds ratios at 95% confidence interval., Results: The prevalence of unmet need for contraception in all the countries considered in this study was 24.9%, with Angola, recording the highest prevalence of 42.6% while Niger had the lowest prevalence of 17.8%. In terms of the individual level predictors, the likelihood of unmet need for contraception was low among AGYW aged 20-24 [aOR = 0.82; 95% CI = 0.76-0.88], those with primary [aOR = 1.22; 95% CI = 1.13-1.31] and secondary/higher levels of formal education [aOR = 1.18; 95% CI = 1.08-1.28, p < 0.001], cohabiting AGYW [aOR = 1.52; 95% CI = 1.42-1.63] and AGYW with three or more births [aOR = 3.41; 95% CI = 3.02-3.85]. At the household/community level, the odds of unmet need for contraception was highest among poorer AGYW [aOR = 1.36; 95% CI = 1.21-1.53], AGYW in female-headed households [aOR = 1.22; 95% CI = 1.13-1.33], urban AGYW [aOR = 1.21; 95% CI = 1.11-1.32] and AGYW who took healthcare decisions alone [aOR = 1.10; 95% CI = 1.01-1.21]., Conclusion: This study has identified disparities in unmet need for contraception among AGYW in high fertility countries in SSA, with AGYW in Angola having the highest prevalence. Both individual and household/community level factors predicted unmet need for contraception among AGYW in this study. However, based on the ICC values, household/community level factors prevailed the individual level factors. Enhancing access to contraception among poorer AGYW, those in female-headed households, those in urban areas and those who take healthcare decisions alone by both governmental and non-governmental organisations in high fertility countries is recommended., Competing Interests: The author has declared that no competing interests exist.

When comparing multilevel models (MLMs) differing in fixed and/or random effects, researchers have had continuing interest in using R-squared differences to communicate effect size and importance of included terms. However, there has been longstanding confusion regarding which R-squared difference measures should be used for which kind of MLM comparisons. Furthermore, several limitations of recent studies on R-squared differences in MLM have led to misleading or incomplete recommendations for practice. These limitations include computing measures that are by definition incapable of detecting a particular type of added term, considering only a subset of the broader class of available R-squared difference measures, and incorrectly defining what a given R-squared difference measure quantifies. The purpose of this paper is to elucidate and resolve these issues. To do so, we define a more general set of total, within-cluster, and between-cluster R-squared difference measures than previously considered in MLM comparisons and give researchers concrete step-by-step procedures for identifying which measure is relevant to which model comparison. We supply simulated and analytic demonstrations of limitations of previous MLM studies on R-squared differences and show how application of our step-by-step procedures and general set of measures overcomes each. Additionally, we provide and illustrate graphical tools and software allowing researchers to automatically compute and visualize our set of measures in an integrated manner. We conclude with recommendations, as well as extensions involving (a) how our framework relates to and can be used to obtain pseudo-R-squareds, and (b) how our framework can accommodate both simultaneous and hierarchical model-building approaches.

Multilevel SEM is an increasingly popular technique to analyze data that are both hierarchical and contain latent variables. The parameters are usually jointly estimated using a maximum likelihood estimator (MLE). This has the disadvantage that a large sample size is needed and misspecifications in one part of the model may influence the whole model. We propose an alternative stepwise estimation method, which is an extension of the Croon method for factor score regression. In this article, we extend this method to the multilevel setting. A simulation study was used to compare this new estimation method to the standard MLE. The Croon method outperformed MLE with regard to convergence rate, bias, MSE, and coverage, in particular when models contained a structural misspecification. In conclusion, the Croon method seems to be a promising alternative to MLE.

Background: Ordinal health longitudinal response variables have distributions that make them unsuitable for many popular statistical models that assume normality. We present a multilevel growth model that may be more suitable for medical ordinal longitudinal outcomes than are statistical models that assume normality and continuous measurements., Methods: The data is from an ongoing prospective cohort study conducted amongst adult women who are HIV-infected patients in Kwazulu-Natal, South Africa. Participants were enrolled into the acute infection, then into early infection subsequently into established infection and afterward on cART. Generalized linear multilevel models were applied., Results: Multilevel ordinal non-proportional and proportional-odds growth models were presented and compared. We observed that the effects of covariates can't be assumed identical across the three cumulative logits. Our analyses also revealed that the rate of change of immune recovery of patients increased as the follow-up time increases. Patients with stable sexual partners, middle-aged, cART initiation, and higher educational levels were more likely to have better immunological stages with time. Similarly, patients having high electrolytes component scores, higher red blood cell indices scores, higher physical health scores, higher psychological well-being scores, a higher level of independence scores, and lower viral load more likely to have better immunological stages through the follow-up time., Conclusion: It can be concluded that the multilevel non-proportional-odds method provides a flexible modeling alternative when the proportional-odds assumption of equal effects of the predictor variables at every stage of the response variable is violated. Having higher clinical parameter scores, higher QoL scores, higher educational levels, and stable sexual partners were found to be the significant factors for trends of CD4 count recovery.

Objective: Overweight/obesity among women is associated with an increased risk of gestational diabetes, pre-eclampsia, postpartum haemorrhage, low birth weight, congenital malformation and neonatal deaths. Although the magnitude of overweight and obesity among the reproductive age group women is a common problem in Ethiopia, there are limited studies that determine the associated factors of overweight and obesity at the national level. Therefore, this study aimed to identify the determinant factors of overweight/obesity among reproductive age group women in Ethiopia., Design: Cross-sectional study design., Setting: Ethiopia., Participants: Non-pregnant women aged 15-49 years., Primary Outcome: Overweight/obesity., Methods: The present study used the Ethiopia Demographic Health Survey (EDHS) data for 2016. A total of 10 938 non - pregnant reproductive age group women were included in the analysis. Both bivariable and multivariable multilevel logistic regression were performed to determine the determinants of overweight and obesity among women in Ethiopia. The OR with a 95% CI was estimated for potential determinants included in the final model., Results: Those women with secondary education (adjusted OR (AOR)=1.48, 1.01, 2.18), higher education (AOR=1.78, 1.13, 2.81), richer (AOR=1.85, 1.15, 2.98) and richest wealth index (AOR=3.23, 1.98, 5.29), urban residence (AOR=4.46, 2.89, 6.87), married (AOR=1.79, 1.21, 2.64), widowed (AOR=2.42, 1.41, 4.15), divorced (AOR=1.84, 1.13, 3.00), aged 25-34 years (AOR=2.04, 1.43, 2.89), 35-44 years (AOR=2.79, 1.99, 3.93) and 45-49 years (AOR=2.62, 1.54, 4.45) had higher odds of developing overweight and obesity., Conclusion: Women with higher education level, high wealth status, older age, formerly married and those urban dwellers had higher odds of overweight and obesity. Therefore, regular physical activity, reducing consumption of fat/energy-dense food as well as modifying the mode of transportation is recommended., Competing Interests: Competing interests: None declared., (© Author(s) (or their employer(s)) 2020. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

Ansari et al. (Psychometrika 67:49-77, 2002) applied a multilevel heterogeneous model for confirmatory factor analysis to repeated measurements on individuals. While the mean and factor loadings in this model vary across individuals, its factor structure is invariant. Allowing the individual-level residuals to be correlated is an important means to alleviate the restriction imposed by configural invariance. We relax the diagonality assumption of residual covariance matrix and estimate it using a formal Bayesian Lasso method. The approach improves goodness of fit and avoids ad hoc one-at-a-time manipulation of entries in the covariance matrix via modification indexes. We illustrate the approach using simulation studies and real data from an ecological momentary assessment.

Purpose : Previous studies have demonstrated positive associations between myopia and environmental risk factors such as urbanization. However, these have failed to account for the clustering of individuals within geographical areas, opening analyses to theoretical and statistical limitations. We demonstrate how a multilevel modelling approach can provide a more nuanced understanding of the relationship between geography and myopia. We examined longitudinal associations between onset of myopia and urban/rural status or population density. Methods : Data were collected over 5 visits during an 8-year period for a UK cohort of 3,512 children. Associations between incident myopia (spherical equivalent ≤ -1.00 diopters) and both urban/rural status and population density were examined using discrete time multilevel hazard models which allow the partitioning of variance into different neighborhood and school areas. Results : There was evidence for an association between myopia and higher population density (Hazard Ratio = 1.14; 95% CI = 1.032 to 1.26) after adjustment for a range of risk factors. There was no strong evidence that urban/rural status was associated with incident myopia. Only a minor amount of variation in myopia was attributable to geographical areas (<2.2%), and this was not explained by rurality or population density. Conclusion : Our findings contrast with previous studies and raise the possibility that some of the results reported may have been driven by confounding bias whereby geographical differences in myopia are driven by lifestyle factors that are correlated with geographical setting.

Rationale: Intersectionality has been increasingly adopted as a theoretical framework within quantitative research, raising questions about the congruence between theory and statistical methodology. Which methods best map onto intersectionality theory, with regard to their assumptions and the results they produce? Which methods are best positioned to provide information on health inequalities and direction for their remediation? One method, multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA), has been argued to provide statistical efficiency for high-dimensional intersectional analysis along with valid intersection-specific predictions and tests of interactions. However, the method has not been thoroughly tested in scenarios where ground truth is known., Method: We perform a simulation analysis using plausible data generating scenarios where intersectional effects are present. We apply variants of MAIHDA and ordinary least squares regression to each, and we observe how the effects are reflected in the estimates that the methods produce., Results: The first-order fixed effects estimated by MAIHDA can be interpreted neither as effects on mean outcome when interacting variables are set to zero (as in a correctly-specified linear regression model), nor as effects on mean outcome averaged over the individuals in the population (as in a misspecified linear regression model), but rather as effects on mean outcome averaged over an artificial population where all intersections are of equal size. Furthermore, the values of the random effects do not reflect advantage or disadvantage of different intersectional groups., Conclusions: Because first-order fixed effects estimates are the reference point for interpreting random effects as intersectional effects in MAIHDA analyses, the random effects alone do not provide meaningful estimates of intersectional advantage or disadvantage. Rather, the fixed and random parts of the model must be combined for their estimates to be meaningful. We therefore advise caution when interpreting the results of MAIHDA in quantitative intersectional analyses., (Copyright © 2019 Elsevier Ltd. All rights reserved.)

Educational researchers, psychologists, social, epidemiological and medical scientists are often dealing with multilevel data. Sometimes, the response variable in multilevel data is categorical in nature and needs to be analyzed through Multilevel Logistic Regression Models. The main theme of this paper is to provide guidelines for the analysts to select an appropriate sample size while fitting multilevel logistic regression models for different threshold parameters and different estimation methods. Simulation studies have been performed to obtain optimum sample size for Penalized Quasi-likelihood (PQL) and Maximum Likelihood (ML) Methods of estimation. Our results suggest that Maximum Likelihood Method performs better than Penalized Quasi-likelihood Method and requires relatively small sample under chosen conditions. To achieve sufficient accuracy of fixed and random effects under ML method, we established ''50/50" and ''120/50" rule respectively. On the basis our findings, a ''50/60" and ''120/70" rules under PQL method of estimation have also been recommended., Competing Interests: The authors have declared that no competing interests exist.

This retrospective study evaluated the influence of known risk factors on nonsurgical periodontal treatment (NSPT) response using a pocket depth fine-tuning multilevel linear model (MLM). Overall, 37 patients (24 males and 13 females) with moderate-to-severe chronic periodontitis underwent NSPT. Follow-up visits at 3, 6, and 12 months included measurements of several clinical periodontal parameters. Data were sourced from a previously reported database. In a total of 1416 initially affected sites (baseline PD ≥ 4 mm) on 536 teeth, probing depth (PD) and clinical attachment loss (CAL) reductions after NSPT were evaluated against known risk factors at 3 hierarchical levels (patient, tooth, and site). For each post-treatment follow-up, the variance component models fitted to evaluate the 3-level variance of PD and CAL decrease revealed that all levels contributed significantly to the overall variance (p < 0.001). Patients who underwent NSPT and were continually monitored had curative results. All 3 hierarchical levels included risk factors influencing the degree of PD and CAL reduction. Specifically, the type of tooth, surfaces involved, and tooth mobility site-level risk factors had the strongest impact on these reductions and were highly relevant for the success of NSPT.

Background: Irrational use of antibiotics is a major driver of development of antibiotic resistance, which heavily threatens population health worldwide. Understanding the mechanism of physician's antibiotic prescribing decisions is increasingly highlighted to promote prudent use of antibiotics. Therefore, the current study aimed to fill the gap, modelling physician's antibiotic prescribing and identifying the potential intrinsic and external determinants of antibiotic prescribing in primary care., Methods: A total of 428,475 prescriptions from 499 prescribers in 65 primary care facilities in Hubei of China were audited. Teixeira Antibiotic Prescribing Behavioral Model (TAPBM) was referred as theoretical basis to identify intrinsic and external predictors of antibiotic prescriptions. A questionnaire survey was conducted, covering potential physician's intrinsic determinants (knowledge, attitudes and individual characteristics) and external factors both in individual level (patient pressure, time pressure and financial incentives) and institutional level (setting and local socio-economic development). A two-level path analysis was performed linking potential determinants of antibiotic use with physician's actual practices., Results: About 44.28% of the prescriptions contained antibiotics, with 9.28% containing two or more antibiotics. The multi-level path analysis revealed that knowledge was a significant predictor of attitudes (β = 0.154, p  < 0.05), but higher knowledge and attitudes failed to translate into antibiotic prescribing practices ((β = - 0.076 - 0.039, p  > 0.05). Instead, external factors played a more important role and physician's antibiotic use was significantly associated with patient pressure (β = 0.102, p  = 0.022), time pressure (β = - 0.164, p  = 0.002), financial incentives (β = - 0.133- - 0.155, p  = 0.027) and institutional environments (rural area, β = 0.408, p  = 0.002; and high socioeconomic setting, β = - 0.641 - -0.578, p  < 0.001 ). The prescribers who were male (β = - 0.168, p  = 0.007) or had lower qualification (β = - 0.114, p  = 0.028) were also more likely to prescribe antibiotics than others., Conclusion: Antibiotic prescribing practices are complex process and associated with both intrinsic (prescriber) and external (patients and institutional environment) factors. A systematic approach is required to curb over-prescription of antibiotics. Apart from educating prescribers, it is equally important, if not more, to educate patients, break incentives and nurture professional culture within organization to reduce the overuse of antibiotics., Competing Interests: Competing interestsThe authors declare no conflict of interest.

Background: Immunization services providers play a crucial role in the successful implementation of immunization, particularly for new vaccines. Several childhood vaccinations that are important for public health are not included in the National Immunization Programme in China, although they are available as optional and self-paid vaccines. Their coverage remains low., Objective: To examine the association between providers' knowledge and recommendations of optional vaccines, as well as other supply- and demand-side factors, and their uptake among children., Methods: A cross-sectional study, that included an in-person questionnaire survey for parents of children under-3 years and a self-administrative questionnaire survey for their vaccination services providers, was conducted in 36 townships or sub-districts in three provinces of China in 2013. Using a sample of 1791 household from 30 townships or sub-districts, we applied multilevel logistic analyses to examine the factors associated with the uptake of optional vaccines based on a hierarchal framework that combined demand-side and supply-side factors., Results: Coverage of optional childhood vaccinations varied across small areas. Supply- and demand-side factors were both associated with the uptake of these vaccines. Immunization services providers' recommendations and their knowledge about optional vaccination were positively and significantly associated with uptake. Children were more likely to receive the vaccines if they lived in communities with higher immunization worker density or larger immunization clinics. Several demand-side psychological factors about childhood vaccination were also associated with optional vaccinations., Conclusions: Promoting immunization services providers to conduct evidence-based recommendations about some important childhood optional vaccinations and enhancing their knowledge regarding optional vaccinations and communication skills are useful strategies to increase the coverage of these vaccinations., (Copyright © 2019 Elsevier Ltd. All rights reserved.)

This paper provides a tutorial companion for the methodological approach implemented in Huh et al. (2015) that overcame two major challenges for individual participant data (IPD) meta-analysis. Specifically, we show how to validly combine data from heterogeneous studies with varying numbers of treatment arms, and how to analyze highly-skewed count outcomes with many zeroes (e.g., alcohol and substance use outcomes) to estimate overall effect sizes. These issues have important implications for the feasibility, applicability, and interpretation of IPD meta-analysis but have received little attention thus far in the applied research literature. We present a Bayesian multilevel modeling approach for combining multi-arm trials (i.e., those with two or more treatment groups) in a distribution-appropriate IPD analysis. Illustrative data come from Project INTEGRATE, an IPD meta-analysis study of brief motivational interventions to reduce excessive alcohol use and related harm among college students. Our approach preserves the original random allocation within studies, combines within-study estimates across all studies, overcomes between-study heterogeneity in trial design (i.e., number of treatment arms) and/or study-level missing data, and derives two related treatment outcomes in a multivariate IPD meta-analysis. This methodological approach is a favorable alternative to collapsing or excluding intervention groups within multi-arm trials, making it possible to directly compare multiple treatment arms in a one-step IPD meta-analysis. To facilitate application of the method, we provide annotated computer code in R along with the example data used in this tutorial., (Copyright © 2019. Published by Elsevier Ltd.)

Background: Neighbourhood social context might play a role in modifying mortality outcomes in severe mental illness, but has received little attention to date. Therefore, we aimed to assess in an ethnically diverse and urban location the association of neighbourhood-level characteristics and individual-level factors for all-cause, natural-cause, and unnatural-cause mortality in those with severe mental illness., Methods: We did a retrospective cohort study using a case-registry from a large secondary mental health-care Trust in an ethnically diverse and urban location in south London, UK. Linked data for deaths and areas of residence were identified from the case-registry. We included all individuals aged 15 years or more at the time of diagnosis for a severe mental illness from Jan 1, 2007, to Dec 31, 2014. We used individual-level information in our analyses, such as gender, marital status, and the presence of current or previous substance use disorders. We assessed neighbourhood or area-level indicators at the Lower Super Output Area level. Association of neighbourhood-level characteristics, which included the interaction between ethnicity and own ethnic density, deprivation, urbanicity, and social fragmentation, alongside individual-level factors for all-cause, natural-cause, and unnatural-cause mortality in those with severe mental illness was assessed., Findings: A total of 18 201 individuals were included in this cohort for analyses, with a median follow-up of 6·36 years. There were 1767 (9·7%) deaths from all causes, 1417 (7·8%) from natural causes, and 192 (1·1%) from unnatural causes. In the least ethnically dense areas, the adjusted rate ratio (aRR) for all-cause mortality in ethnic minority groups with severe mental illness compared with white British people with severe mental illness were similar (aRR 0·96, 95% CI 0·71-1·29); however in the highest ethnic density areas, ethnic minority groups with severe mental illness had a lower risk of death (aRR 0·52, 95% CI 0·38-0·71; p<0·0001), with similar trends for natural-cause mortality (p=0·071 for statistical interaction). In the cohort with severe mental illness, residency in deprived, urban, and socially fragmented neighbourhoods was not associated with higher mortality rates. Compared with the general population, age-standardised and gender-standardised mortality ratios were elevated in the cohort with severe mental illness across all neighbourhood-level characteristics assessed., Interpretation: For ethnic minority groups with severe mental illness, residency in areas of higher own-group ethnic density is associated with lower mortality compared to white British groups with severe mental illness., Funding: Health Foundation, National Institute for Health Research, EU Seventh Framework, and National Institute of Mental Health., (Copyright © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

Psychological investigations into the structure of well-being have been largely cross-sectional. However, longitudinal models are needed as Western societies work to improve individual well-being. The current multilevel-modeling study examined within-person dynamics of well-being over 8 years. I asked two questions: (a) How do life satisfaction and psychological well-being (measures drawn from two well-being research traditions) relate over time? and (b) do these relationships vary on the basis of individuals' extraversion or neuroticism? Measures of life satisfaction and psychological well-being were collected in 8 consecutive years from 159 American adults in late midlife. A dispositional-life-satisfaction set point and yearly variation in life satisfaction both related to higher yearly psychological well-being. Neuroticism, but not extraversion, predicted a stronger within-person relationship between life satisfaction and psychological well-being. For participants with very low neuroticism, life satisfaction and psychological well-being varied independently. In sum, experiences of life satisfaction and psychological well-being converged for more neurotic individuals and diverged for more emotionally stable participants.