Welcome
Welcome! I am a Principal Biostatistician at Novartis Pharmaceuticals and an Honorary Research Associate at Imperial College London.
At Novartis, I am part of the Advanced Methodology & Data Science (AMDS) team. I am developing original statistical methods for pharmaceutical applications including the integration of imaging data for survival prediction using deep learning.
I completed my PhD in “Modern Statistics and Statistical Machine Learning” at Imperial College London in 2022. My supervisors were Dr. Oliver Ratmann and Prof. Samir Bhatt. I focused my research on Bayesian models and methods to estimate age-specific infectious disease transmission dynamics, integrating disease surveillance time series, mobile phone data, and vaccination data.
Interests
I am interested in various methodologies and models relevant for understanding complex public health and biomedical phenomena. On the public health side, this includes age-specific transmission modelling, high-resolution human contact patterns, reconstruction of transmission networks. On the biomedical side, this includes dose-escalation models, dose-finding models and causal inference. I am also interested in the development of scalable inference techniques especially in the Bayesian context.
To inform those models, complex and large datasets are needed. In my work, I have analysed diverse datasets including mobile phone data, time series and spatio-temporal data, time-to-event data, large survey data, genomics, deep-sequenced data and imaging data.
In more details…
Bayesian modelling for large and complex data
I am interested in Bayesian models and methods to capture complex dynamics, allowing for a deeper understanding of intricate systems. I am also interested in developing efficient implementations using Stan and scalable inference techniques that enables the analysis of large datasets and complex models with greater speed and accuracy.
Deep Learning for medical imaging
Integrating imaging data for survival prediction using deep learning.
Epidemiological Forecasting
My research is driven by a strong commitment to guiding public health decisions and addressing key topics of critical importance. Some of the key areas of my work include:
-
Investigation of the COVID-19 outbreak in Europe and the United States: Within this context, we delved into critical questions such as the demographics of individuals who drove infection, whether the epidemics were under control, the effectiveness of non-pharmaceutical interventions such as lockdowns or school closures in controlling transmission, and the potential outcomes if such interventions had not been implemented. (Flaxman 2020, Monod 2021, Mishra 2021)
-
Characterisation of the gender disparity in HIV infection in Rakai, Uganda: We aimed to comprehensively assess how HIV incidence has evolved over time and identify the specific population groups that played a significant role in driving transmission. We explored the impacts of behavioral interventions-more specifically leading a men’s initiative aimed at closing the viral suppression gap-seems. (Monod, 2023)
Mobile phone data
During the COVID-19 pandemic, my research leveraged mobility data from Google and Foursquare to closely monitor the real-time evolution of the virus spread in the United States. This rich data provided valuable insights into the movement patterns of individuals and their potential impact on transmission dynamics. (Monod, 2021)
Spatial statistics
One area of particular interest is the development of scalable non-parametric spatial methods. Such methods allow to efficiently analyze large-scale spatial datasets and extract meaningful insights. For example, in Monod et al (2022), we focused on advancing the field of spatial modeling by introducing a novel approach. We developed a low-rank two-dimensional Gaussian process (GP) that was projected using regularized B-splines.
Genomics
In my research, I have explored the intricacies of the virus’s genetic evolution using deep-sequenced RNA HIV datasets. In Monod (2023), we used phylogenetic trees reconstructed from patient-specific deep-sequenced RNA using a powerful tool called Phyloscanner (Wymant, 2018). By combinining advanced analytical techniques, including phylogenetic tree analysis, we aimed to contribute to reconstruct transmission trees (MSc thesis) and characterize transmission dynamics over time (Monod, 2023)
Survival analysis
Here are two notable instances where I applied survival analysis techniques:-
At Novartis, I had the opportunity to employ survival analysis techniques to develop a time-to-event model for dose escalation. This involved analyzing the time it takes for specific events, such as adverse reactions or treatment response, to occur in patients.
-
We conducted an incidence analysis to estimate the longitudinal and fine-grained age-specific HIV incidence rates (Monod, 2023). This analysis provides valuable insights into the dynamics of HIV transmission and helps identify vulnerable populations. Moreover, it facilitates the development of targeted interventions and prevention strategies to mitigate the spread of HIV.
Public Health
Estimating fine age social contact patterns: Recognizing the critical role of social contact in pathogen spread, we focused on unraveling the complexities of social interactions across different age groups. (Monod 2021)
Data scraping
During the COVID-19 outbreak, we compiled an unprecedented new dataset of age-specific COVID-19 daily deaths data retrieved from city or state Department of Health (DoH) websites, data repositories or via data requests to DoH (Monod, 2021). The data were extracted daily using multiple python libraries, including fitz, BeautifulSoup and selenium.