Honored to join the Strategic Advisory Group of Experts (SAGE) Working Group on pertussis vaccines. Together with a dozen other colleagues, we will support the WHO by reviewing scientific evidence and drafting recommendations on pertussis vaccines. www.who.int/groups/strat...
Thank you!
💡What are the implications? Ans: Social contact structure can affect the impact of PCVs and should be taken into consideration when estimating vaccine impact.
Congratulations to Anabelle Wong for publishing her PhD work and a big thank you to co-authors Sarah Kramer and Dan Weinberger! (6/6)
🔍 We found that varying the social contact matrix alone led to a range of time-to-elimination (3.8-6 years). We further found that such variation was largely explained by the social contact features (total contact rate and assortativity) of children under 5. (5/6)
👩🏻💻 We developed a compartmental transmission model to investigate the effect of social contact structure on the impact of PCVs (i.e., how fast PCVs eliminate vaccine-targeted serotypes). (4/6)
❓Many population factors can influence the dynamics of VT elimination, for example, social behaviours. So we asked, what is the effect of social contact structure on the PCV-induced VT elimination dynamics? (3/6)
🦠Pneumococcus is highly diverse but pneumococcal vaccines (PCVs) only target a fraction of the many serotypes. As PCVs reduced carriage of vaccine-targeted serotypes (VT), more carriage of non-vaccine-targeted serotypes (NVT) were observed — a phenomenon known as serotype replacement. (2/6)
🚨New paper alert: Sharing another latest study from our lab, "Assessing the effect of social contact structure on the impact of pneumococcal conjugate vaccines." nature.com/articles/s41...
Read the 🧵to find out more! (1/6)
A big thank you to co-authors Pej Rohani, Tine Dalby, and Anabelle Wong! (6/6)
💡What is the implication? True infection burden is probably somewhere between reported case number and seropositivity-based estimates. Mathematical models integrating both data streams may get us a better estimate in the future! (5/6)
🧑🏻💻 By fitting a methematical model to data from serosurveys during the whole-cell pertussis vaccine era, we found that the postive predictive value (PPV) of using seropositivity to estimate transmissible infections is low, esp. in young adults (20-39y) where PPV was <50%. (4/6)
📈 While seroprevalence data can aid estimation of the circulating pertussis infection burden, ignoring natural immune boosting would lead to an overestimation of cases and underestimation of vaccine impact. (3/6)
🩸Natural immune boosting occurs when pathogen triggers a detectable immune response in the (vaccinated or previously infected and recovered) host without causing a transmissible infection. (2/6)
Happy to share our @natcomms.nature.com paper 📢 “Natural immune boosting biases pertussis infection estimates in seroprevalence studies.” nature.com/articles/s41...
Read the🧵 to find out more! (1/6)
Massive congratulations to the new Dr. Laura Barrero Guevara (@labarreroguevara.bsky.social), who successfully defended her PhD on causal inference and infectious diseases yesterday, with summa cum laude!! Check out her work here rdcu.be/ewCNj and here doi.org/10.1093/infd... 🥳🎉
(10/10) Big thank you to our co-authors! Laura Barrero Guevara, Sarah Kramer and Tobias Kurth!
(9/10) Integrating #causalinference concepts with transmission models is necessary for inferring the effect of weather on infectious diseases and subsequently predicting the consequences of climate change on infectious diseases. Check out the paper here: www.nature.com/articles/s41... 🥳
(8/10) Fourth vignette: causal inference concepts can help to interpret the direct and indirect effects of weather on transmission. For example, temperature can affect transmission directly and indirectly (through humidity), and these effects vary by local climate.
(7/10) Third vignette: causal inference helps identify and avoid confounding bias. Gradients in climate across locations can masquerade as spatial spread of disease.
(6/10) Second vignette: causal inference can inform strategic choices of a study location to achieve the set-up of a natural experiment. By comparing temperate and tropical climates, we highlight how local conditions can help isolate the causal weather variable.
(5/10) First vignette: causal inference concepts can guide study design. Considering the complex causal paths between weather, transmission, and incidence, we show that measurement bias is a concern for time-series regression studies linking weather and incidence.
(4/10) Our new paper shows how applying causal inference concepts can help. We illustrate this with four short case studies based on our causal graph #dag ⬇️ linking weather, disease transmission, and reported cases.
(3/10) In practice, this often means using observational data—case counts and weather variables. Yet, interpreting such data can be challenging, as associations do not necessarily imply true causal effects.
(2/10) A key question arising from climate change is how it will impact the transmission of infectious diseases. Predicting these effects demands understanding how weather affects their transmission dynamics.
How does weather affect the transmission of #infectiousdiseases, and how can we predict the effects of #climatechange on them? Our new article in @natureportfolio.bsky.social Ecology & Evolution explores these questions using #causalinference and #transmissionmodels. See the 🧵for more! (1/10)
(7/7) Our study provides one of the first estimates of the strength and duration of the interaction between flu and RSV. We show how #mathematicalmodels can be vital to understanding virus-virus interactions. Check out the full paper here: www.nature.com/articles/s41...!
(6/7) We also used our model to explore the potential for using live influenza #vaccines to control RSV outbreaks. We found that the effectiveness of this strategy is likely to depend on the size and timing of flu and RSV outbreaks in a given location.
(5/7) We found evidence of a moderate to strong, negative interaction between flu and RSV – being infected with either virus may provide protection against infection with the other. Our results also suggest this protection could last for anywhere from 1 to 5 months.
A schematic of a two-pathogen interacting transmission model.
(4/7) Here, we used a mathematical model of #flu and #RSV cocirculation to estimate the strength and duration of the interaction between the two viruses. Specifically, we fitted our model to flu and RSV data from Hong Kong and Canada.
(3/7) However, mathematical models can explicitly account for these complex and random processes.