Multilevel Analysis

BackgroundSample size planning for longitudinal data is crucial when designing mediation studies because sufficient statistical power is not only required in grant applications and peer-reviewed publications, but is essential to reliable research results. However, sample size determination is not straightforward for mediation analysis of longitudinal design.MethodsTo facilitate planning the sample size for longitudinal mediation studies with a multilevel mediation model, this article provides the sample size required to achieve 80% power by simulations under various sizes of the mediation effect, within-subject correlations and numbers of repeated measures. The sample size calculation is based on three commonly used mediation tests: Sobel's method, distribution of product method and the bootstrap method.ResultsAmong the three methods of testing the mediation effects, Sobel's method required the largest sample size to achieve 80% power. Bootstrapping and the distribution of the product method performed similarly and were more powerful than Sobel's method, as reflected by the relatively smaller sample sizes. For all three methods, the sample size required to achieve 80% power depended on the value of the ICC (i.e., within-subject correlation). A larger value of ICC typically required a larger sample size to achieve 80% power. Simulation results also illustrated the advantage of the longitudinal study design. The sample size tables for most encountered scenarios in practice have also been published for convenient use.ConclusionsExtensive simulations study showed that the distribution of the product method and bootstrapping method have superior performance to the Sobel's method, but the product method was recommended to use in practice in terms of less computation time load compared to the bootstrapping method. A R package has been developed for the product method of sample size determination in mediation longitudinal study design.

Understanding modifiable determinants of full immunization of children provide a valuable contribution to immunization programs and help reduce disease, disability, and death. This study is aimed to assess the individual and community-level determinants of full immunization coverage among children in the Democratic Republic of Congo. This study used data from the Demographic and Health Survey 2013-14 from the Democratic Republic of Congo. Data regarding total 3,366 children between 12 and 23 months of age were used in this study. Children who were immunized with one dose of BCG, three doses of polio, three doses of DPT, and a dose of measles vaccine was considered fully immunized. Descriptive statistics were calculated for the prevalence and distribution of full immunization coverage. Two-level multilevel logistic regression analysis, with individual-level (level 1) characteristics nested within community-level (level 2) characteristics, was used to assess the individual- and community-level determinants of full immunization coverage. This study found that about 45.3% [95%CI: 42.02, 48.52] of children aged 12-23 months were fully immunized in the DRC. The results confirmed immunization coverage varied and ranged between 5.8% in Mongala province to 70.6% in Nord-Kivu province. Results from multilevel analysis revealed that, four Antenatal Care (ANC) visits [AOR: 1.64; 95%CI: 1.23, 2.18], institutional delivery [AOR: 2.37; 95%CI: 1.52, 3.72], and Postnatal Care (PNC) service utilization [AOR: 1.43; 95%CI: 1.04, 1.95] were statistically significantly associated with the full immunization coverage. Similarly, children of mothers with secondary or higher education [AOR: 1.32; 95%CI: 1.00, 1.81] and from the richest wealth quintile [AOR: 1.96; 95%CI: 1.18, 3.27] had significantly higher odds of being fully immunized compared to their counterparts whose mothers were relatively poorer and less educated. Among the community-level characteristics, residents of the community with a higher rate of institutional delivery [AOR: 2.36; 95%CI: 1.59, 3.51] were found to be positively associated with the full immunization coverage. Also, the random effect result found about 35% of the variation in immunization coverage among the communities was attributed to community-level factors.The Democratic Republic of Congo has a noteworthy gap in full immunization coverage. Modifiable factors-particularly health service utilization including four ANC visits, institutional delivery, and postnatal visits-had a strong positive effect on full immunization coverage. The study underlines the importance of promoting immunization programs tailored to the poor and women with little education.

BackgroundEarly or timely initiation of antenatal care and regular visits based on the schedule have a tremendous effect on both maternal and fetal health. Despite the paramount benefits of early initiation of ANC within the first 12 weeks of pregnancy, women still do not have adequate and equal access to high-quality early antenatal care.ObjectiveTo determine the prevalence and factors associated with delayed first ANC booking in Ethiopia.MethodA secondary data analysis was conducted using the 2016 Ethiopian demographic and health survey data. All reproductive age women who gave birth in the five years preceding the survey and who had ANC visit for their last child were included in this study. The total weighted sample size analyzed was 4,741. Due to the hierarchical nature of the EDHS data, Multi-level logistic regression model was used to identify the individual and community level factors associated with delayed first ANC booking.ResultIn this study, the prevalence of delayed first ANC booking was 67.31% [95% CI: 65.96% to 68.63%]. Women with secondary and higher education [Adjusted Odd Ratio (AOR) = 0.78; 95%CI: 0.61, 0.99] and [AOR = 0.61; 95%CI: 0.44, 0.83] respectively had lower odds of delayed first ANC booking. But woman who were multiparous and grand multiparous [AOR = 1.21; 95%CI: 1.01, 1.45] and [AOR = 1.50; 95%CI: 1.16, 1.93] respectively, women with the last pregnancy wanted no more [AOR = 1.52; 95%CI: 1.10, 2.09], a woman who was living in the rural area [AOR = 1.66; 95%CI: 1.25, 2.21], and a woman who was living in large central regions and small peripheral regions [AOR = 2.76; 95%CI: 2.20, 3.47] and [AOR = 2.70; 95%CI: 2.12, 3.45] respectively had higher odds of delayed first ANC booking.ConclusionDespite the documented benefits of early antenatal care initiation, late ANC booking is still predominant in Ethiopia as highlighted by this study. Maternal education, parity, wanted the last child, residence and region were significantly associated with delayed first ANC booking. Therefore, taking special attention for these high-risk groups could decrease delayed first ANC booking and this intern decreases maternal and fetal health problems by identifying and intervene early.

Meta-analyses often include multiple dependent effect sizes, yet current methods typically neglect the resulting within-study dependencies or fail to address model uncertainty and publication bias adequately. We extend robust Bayesian meta-analysis (RoBMA) to a multilevel framework, simultaneously handling within-study dependencies, model uncertainty, heterogeneity, moderators, and publication bias. Specifically, the three-level RoBMA integrates approximate Bayesian selection models with PET-PEESE adjustments within a hierarchical Bayesian setting. We illustrate the methodology through empirical examples and demonstrate its performance via simulations. The approach is implemented in the RoBMA R package and JASP.