Bacterial sexually transmitted infections (STIs) continue to pose major global public health challenges. The authors emphasise Neisseria gonorrhoeae as a pathogen of particular concern because of persistently high prevalence and the progressive emergence of antimicrobial resistance that has reduced treatment options. In settings where prevention is critical, determining whether transmission is homogeneous or dominated by a minority of individuals (that is, superspreading) affects how interventions should be prioritised.
Traditional approaches to quantify transmission heterogeneity commonly depend on dense contact tracing, which is not always feasible. The study explores the use of genomic data and phylodynamic inference as an alternative way to estimate superspreading in bacterial STI outbreaks.
The authors applied Bayesian phylodynamic methods to infer transmission dynamics from bacterial sequence data. They choose a multi-type birth–death model framework, which allows explicit representation of two transmission types (superspreaders and non-superspreaders) and estimation of differential transmission parameters. This class of model has been used extensively for viral phylodynamics and is adapted here for bacterial outbreak data.
Model parameters were set to enable direct estimation of key quantities relevant to transmission heterogeneity: the fraction of cases that are superspreaders, the relative contribution or impact of superspreaders compared with non-superspreaders, and separate reproductive numbers for the two types. These parameter estimates aim to quantify both how many individuals act as superspreaders and how much more they transmit relative to other cases.
To increase statistical power to detect superspreading within individual clusters, the authors used a hierarchical modelling strategy with partial pooling. This approach borrows strength across clusters while still allowing cluster-specific estimates, improving detection capability when individual clusters have limited data.
The model's performance was evaluated across a range of superspreading scenarios. Evaluations used both transmission-informed phylogenies (where the tree reflects known transmission patterns) and sequence-derived data incorporating phylogenetic uncertainty. The authors report successful model performance across the tested scenarios, indicating the method can recover superspreading signals under varying conditions of data and tree uncertainty.
When applied to empirical genomic datasets, the method identified a substantial role for superspreading in N. gonorrhoeae transmission during the COVID-19 pandemic in Australia. The authors present these results as evidence that transmission heterogeneity contributed meaningfully to observed outbreak dynamics in that context. Specific numerical estimates, cluster-level results, and additional empirical details are provided in the full preprint and supplementary materials.
The findings underscore that detecting and quantifying superspreading in bacterial STIs can inform more efficient public health responses. If a minority of cases disproportionately drive transmission, targeted interventions toward high-transmission settings or behaviours could be more effective than uniform measures. The authors highlight the utility of genomic phylodynamic approaches as a complement to traditional epidemiological methods, especially where contact tracing is incomplete or infeasible.
This work is reported as a preprint on medRxiv and has not undergone peer review; the authors explicitly note that it should not be used to guide clinical practice without further validation. Ethical approvals and participant consent statements are reported in the article. Funding sources declared include the Agence Nationale de la Recherche and the Australian Research Council. The preprint provides supplementary materials and states that model evaluation used both idealised transmission-informed trees and sequence data with phylogenetic uncertainty; readers should consult the full text and supplements for detailed methods, parameter settings, and cluster-specific estimates.
Overall, the study adapts a multi-type birth–death model within a Bayesian phylodynamic framework and implements hierarchical partial pooling to sensitively quantify transmission heterogeneity from bacterial genomic data, demonstrating practical application to N. gonorrhoeae during the COVID-19 period in Australia.