Performance of a Mixed Effects Logistic Regression Model for Binary Outcomes With Unequal Cluster Size

2005 ◽  
Vol 15 (3) ◽  
pp. 513-526 ◽  
Author(s):  
Moonseong Heo ◽  
Andrew C. Leon
Keyword(s):  
Logistic Regression ◽  
Regression Model ◽  
Cluster Size ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Binary Outcomes ◽  
Unequal Cluster

Relationship between childhood obesity and socioeconomic status among primary school children in Costa Rica

2021 ◽  
pp. 1-24
Author(s):  
Tatiana Gamboa-Gamboa ◽  
Romain Fantin ◽  
Jeancarlo Cordoba ◽  
Ivannia Caravaca ◽  
Ingrid Gómez-Duarte
Keyword(s):  
Socioeconomic Status ◽  
Logistic Regression ◽  
Childhood Obesity ◽  
Costa Rica ◽  
Regression Model ◽  
Primary Schools ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Overweight And Obesity ◽  
Public And Private

Abstract Objective: This article analyzes the relationship between socioeconomic status and the prevalence of overweight and obesity in the primary school population in Costa Rica. Design: A National School Weight/Height Census was disseminated across Costa Rica in 2016. The percentage of children who were overweight or obese was calculated by sex, age, and socioeconomic indicators (type of institution: private, public, mix; type of geographic location: rural, urban; and the level of development of the district of residence: quartiles). A mixed effects multinomial logistic regression model and mixed effects logistic regression model were used to analyze the association between the prevalence of being overweight or obese and district socioeconomic status. Setting: The survey was carried out in public and private primary schools across Costa Rica in 2016. Participants: 347,366 students from 6 to 12 years old, enrolled in public and private primary schools. Results: The prevalence of overweight and obesity among children was 34.0%. Children in private schools were more likely to be overweight or obese than students in public schools (OR=1.10 [1.07, 1.13]). Additionally, children were less likely to be overweight or obese if attending a school in a district of the lowest socioeconomic quartile compared to the highest socioeconomic quartile (OR=0.79 [0.75, 0.83]), and in a rural area compared to the urban area (OR=0.92 [0.87, 0.97]). Conclusions: Childhood obesity in Costa Rica continues to be a public health problem. Prevalence of overweight and obesity in children was associated with indicators of higher socioeconomic status.

scholarly journals Quantifying the Robustness of the English Sibilant Fricative Contrast in Children

2015 ◽  
Vol 58 (3) ◽  
pp. 622-637 ◽  
Author(s):  
Jeffrey J. Holliday ◽  
Patrick F. Reidy ◽  
Mary E. Beckman ◽  
Jan Edwards
Keyword(s):  
Logistic Regression ◽  
Regression Model ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Vocabulary Size ◽  
Phonological Contrast ◽  
Different Types ◽  
Centroid Frequency ◽  
The Relationship ◽  
Children's Speech

Purpose Four measures of children's developing robustness of phonological contrast were compared to see how they correlated with age, vocabulary size, and adult listeners' correctness ratings. Method Word-initial sibilant fricative productions from eighty-one 2- to 5-year-old children and 20 adults were phonetically transcribed and acoustically analyzed. Four measures of robustness of contrast were calculated for each speaker on the basis of the centroid frequency measured from each fricative token. Productions that were transcribed as correct from different children were then used as stimuli in a perception experiment in which adult listeners rated the goodness of each production. Results Results showed that the degree of category overlap, quantified as the percentage of a child's productions whose category could be correctly predicted from the output of a mixed-effects logistic regression model, was the measure that correlated best with listeners' goodness judgments. Conclusions Even when children's productions have been transcribed as correct, adult listeners are sensitive to within-category variation quantified by the child's degree of category overlap. Further research is needed to explore the relationship between the age of a child and adults' sensitivity to different types of within-category variation in children's speech.

scholarly journals Impact of Transcatheter Aortic Valve Replacement on Hospitalization Rates: Insights From Nationwide Readmission Database

2021 ◽  
Author(s):  
Ahmed Elkaryoni ◽  
Adnan K. Chhatriwalla ◽  
Kevin F. Kennedy ◽  
John T. Saxon ◽  
John J. Lopez ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Aortic Valve ◽  
Aortic Valve Replacement ◽  
Regression Model ◽  
Transcatheter Aortic Valve Replacement ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Transcatheter Aortic Valve ◽  
Hospitalization Rates ◽  
Before And After

Background Hospitalization rates after transcatheter aortic valve replacement (TAVR) remain high, given the age and comorbidities of patients undergoing TAVR. To better understand the impact of TAVR on hospitalization, we sought to compare hospitalization rates before and after TAVR and to examine if underlying patient comorbidities are associated with a differential effect of TAVR on hospitalizations. Methods and Results We used the Nationwide Readmissions Database to identify patients who underwent TAVR. As Nationwide Readmissions Database data do not cross over calendar years, we limited our index admission to hospitalizations during April to September of each calendar year to allow 90 days of observation before and after TAVRs. We calculated the daily risk of all‐cause hospitalization and used a mixed‐effects logistic regression model to explore interactions between patient characteristics, TAVR, and hospitalization risk. Among 39 249 patients who underwent TAVR in 2014 to 2017 (median age, 82 years [interquartile range, 76–87 years]; 45.7% women), 32.0% had at least one hospitalization in the 90 days before TAVR compared with 23.2% in the 90 days post‐TAVR (relative reduction, 27.5%; P <0.001). In the mixed‐effects logistic regression model, TAVR was associated with decreased all‐cause hospitalization rate after TAVR in all comorbidity subgroups. However, younger patients and those with heart failure and reduced ejection fraction appeared to have more robust reduction in hospitalizations. Conclusions Although patients who are treated with TAVR have high rates of rehospitalization, TAVR is associated with an overall reduction in all‐cause hospitalizations regardless of underlying patient comorbidities.

scholarly journals A simulation study for evaluating the performance of clustering measures in multilevel logistic regression

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Nicholas Siame Adam ◽  
Halima S. Twabi ◽  
Samuel O.M. Manda
Keyword(s):  
Logistic Regression ◽  
Regression Model ◽  
Correlation Coefficient ◽  
Odds Ratio ◽  
Cluster Size ◽  
Logistic Regression Model ◽  
Multilevel Logistic Regression ◽  
Logistic Regression Models ◽  
Childhood Anemia ◽  
Cluster Level

Abstract Background Multilevel logistic regression models are widely used in health sciences research to account for clustering in multilevel data when estimating effects on subject binary outcomes of individual-level and cluster-level covariates. Several measures for quantifying between-cluster heterogeneity have been proposed. This study compared the performance of between-cluster variance based heterogeneity measures (the Intra-class Correlation Coefficient (ICC) and the Median Odds Ratio (MOR)), and cluster-level covariate based heterogeneity measures (the 80% Interval Odds Ratio (IOR-80) and the Sorting Out Index (SOI)). Methods We used several simulation datasets of a two-level logistic regression model to assess the performance of the four clustering measures for a multilevel logistic regression model. We also empirically compared the four measures of cluster variation with an analysis of childhood anemia to investigate the importance of unexplained heterogeneity between communities and community geographic type (rural vs urban) effect in Malawi. Results Our findings showed that the estimates of SOI and ICC were generally unbiased with at least 10 clusters and a cluster size of at least 20. On the other hand, estimates of MOR and IOR-80 were less accurate with 50 or fewer clusters regardless of the cluster size. The performance of the four clustering measures improved with increased clusters and cluster size at all cluster variances. In the analysis of childhood anemia, the estimate of the between-community variance was 0.455, and the effect of community geographic type (rural vs urban) had an odds ratio (OR)=1.21 (95% CI: 0.97, 1.52). The resulting estimates of ICC, MOR, IOR-80 and SOI were 0.122 (indicative of low homogeneity of childhood anemia in the same community); 1.898 (indicative of large unexplained heterogeneity); 0.345-3.978 and 56.7% (implying that the between community heterogeneity was more significant in explaining the variations in childhood anemia than the estimated effect of community geographic type (rural vs urban)), respectively. Conclusion At least 300 clusters with sizes of at least 50 would be adequate to estimate the strength of clustering in multilevel logistic regression with negligible bias. We recommend using the SOI to assess unexplained heterogeneity between clusters when the interest also involves the effect of cluster-level covariates, otherwise, the usual intra-cluster correlation coefficient would suffice in multilevel logistic regression analyses.

The morphosyntactic characteristics of null subjects in Old English

2019 ◽  
pp. 93-122
Author(s):  
Kristian A. Rusten
Keyword(s):  
Logistic Regression ◽  
Regression Model ◽  
Old English ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Linguistic Variables ◽  
Null Subjects ◽  
Null Subject ◽  
Structural Variables ◽  
The Individual

Chapter 4 provides an in-depth quantitative investigation of the morpohsyntactic characteristics of null subjects in Old English. The distribution of overt and null subjects is presented according to the structural variables of clause type, the position of the finite verb, person, and number. Attempts are made to fit an explanatory generalized mixed-effects logistic regression model incorporating these linguistic variables, as well as those non-linguistic variables which emerged as significant in Chapter 3. It is demonstrated that correlations between the occurrence of null subjects and such variables are extremely weak when both genre and the individual text are taken into account: only minuscule influence on the odds of having a null subject instead of an overt one is exercised by these variables. It is argued that this strengthens the view of null subjects in Old English as linguistic ‘residue’.

scholarly journals Spatial distribution and factors associated with low birth weight in Ethiopia using data from Ethiopian demographic and health survey 2016: spatial and multilevel analysis

2020 ◽  
Author(s):  
Alemneh Mekuriaw Liyew ◽  
Malede Mequanent Sisay ◽  
Achenef Asmamaw Muche
Keyword(s):  
Logistic Regression ◽  
Spatial Distribution ◽  
Birth Weight ◽  
Low Birth Weight ◽  
Regression Model ◽  
Logistic Regression Model ◽  
Mixed Effects ◽  
Sub Saharan Africa ◽  
Factors Associated ◽  
Risk Areas

AbstractBackgroundLow birth weight (LBW) was a leading cause of neonatal mortality. It showed an increasing trend in Sub-Saharan Africa for the last one and half decade. Moreover, it was a public health problem in Ethiopia. Even though different studies were conducted to identify its predictors, contextual factors were insufficiently addressed in Ethiopia. There was also limited evidence on the spatial distribution of low birth weight. Therefore, this study aimed to explore spatial distribution and factors associated with low birth weight in Ethiopia.MethodSecondary data analysis was conducted using the 2016 EDHS data. A total of 1502 (weighted sample) mothers whose neonates were weighed at birth five years preceding the survey were included. GIS 10.1, SaTscan, stata, and Excel were used for data cleaning and analysis. A multi-level mixed-effects logistic regression model was fitted to identify factors associated with low birth weight. Finally, hotspot areas from GIS results, log-likelihood ratio (LLR) and relative risk with p-value of spatial scan statistics, AOR with 95% CI and random effects for mixed-effects logistic regression model were reported.ResultsLow birth weight was spatially clustered in Ethiopia. Primary (LLR=11.57; P=0.002) clusters were detected in the Amhara region. Whereas secondary (LLR=11.4; P=0.003;LLR=10.14,P=0.0075) clusters were identified at Southwest Oromia, north Oromia, south Afar, and Southeast Amhara regions. Being severely anemic (AOR=1.47;95%CI1.04,2.01), having no education (AOR=1.82;95%CI1.12,2.96), Prematurity (AOR=5.91;95%CI3.21,10.10) female neonate (AOR=1.38;95%CI1.04,1.84)were significantly associated with LBWConclusionLBW was spatially clustered in Ethiopia with high-risk areas in Amhara,Oromia, and Afar regions and it was affected by socio demographic factors. Therefore, focusing the policy intervention in those geogrsphically low birth weight risk areas and improving maternal education and nutrtion could be vital to reduce the low birth weight disparity in Ethiopia.

Sign in / Sign up

Export Citation Format

Share Document