Upcoming Seminar – 27/01/26
#SSACAB
|
#scienceforafricafoundation
|
#wellcometrust
|
#universityofpretoria
|
#UpcomingSeminar
|
#ResearchSeminar
|
#Statistics
|
#Biostatistics
|
#MetaAnalysis
|
#RandomEffectsModels
|
#Heterogeneity
|
#FiniteMixtureModels
|
#BayesianStatistics
|
#HealthResearch
|
#QuantitativeMethods
|
#AcademicEvents
📢
Upcoming Seminar
We are pleased to invite you to an engaging seminar featuring two complementary talks that explore advanced methods in meta-analysis and evidence synthesis.
🔹
Talk 1
Exploring Heterogeneity in Random Effects Meta-Analysis Using Finite Mixture Models
Presenter: Itesiwaju Babalola
Key focus areas:
- Meta-analysis
- Random-effects modelling
- Between-study heterogeneity
- Finite mixture modelling
- Mixture random-effects models
🔹
Talk 2
A Simulation Study on Bayesian Network Meta-Analysis Comparing Multiple Treatments Simultaneously for Binary Outcomes
Presenter: Mashudu Thagwana
Key focus areas:
- Fixed- and random-effects meta-analysis models
- Network meta-analysis methodology
- Simulation-based evaluation of model performance under varying levels of between-study heterogeneity
- Comparison of frequentist and Bayesian approaches
🗓 Date: 27 January 2026
⏰Time: 14:00 – 15:00 (SAST)
🔗
Join the seminar:
https://bit.ly/45ir990
🔍
Why It Matters
Why exploring heterogeneity with finite mixture models matters
Between-study heterogeneity is a common challenge in meta-analysis and can mask important differences across studies. Finite mixture models provide a flexible framework for identifying latent subgroups of studies, leading to more realistic modelling of variability and more reliable pooled estimates. This approach enhances the interpretability and robustness of evidence synthesis, particularly in complex or diverse research settings.
Why Bayesian network meta-analysis matters
Network meta-analysis enables the simultaneous comparison of multiple treatments, even when direct comparisons are limited or unavailable. By evaluating model performance through simulation and comparing Bayesian and frequentist approaches, this work helps researchers understand when and how different methods perform best—supporting more informed, transparent, and reliable decision-making in clinical and policy contexts.