The field of phylodynamics hypothesises that the way viruses spread leaves footprints in their genetic information. This opens great potential to provide insights into complex epidemiological dynamics, but in practice, there are many hurdles to be overcome. One of them is that the number of parameters to infer from raw DNA or RNA sequence data is huge, making likelihood-based methods challenging for non-trivial models. A promising alternative is a simulation-based approach called ABC for Approximate Bayesian Computation. Its downside is that it is computationally intensive, since thousands (if not millions) of simulation runs to obtain a good quality model fit. To avoid this, people have proposed all sorts of optimization schemes, such as regression ABC, which relies on state-of-the-art machine learning techniques.
These short courses will take place at SACEMA, Stellenbosch: Introduction to R: Management, Exploration, and Communication of Data (2-6 July 2018) and Advanced Epidemiological Methods (30 July-2 August 2018).
Two months ago Lander Willem and I organized the first edition of the short course “Individual-based modelling in epidemiology: A practical introduction”. The feedback at the end of the course was overwhelmingly positive, which left us feeling empowered and encouraged to not leave it at this first edition. Participants of the next edition should expect an even more hands-on course, with more time to acquire skills in developing, exploring and fitting individual-based models.
In modelling hierarchical data we can take into account spatial and temporal correlations by introducing spatiotemporal random effects in the model. Several other hurdles have to be overcome when modelling hierarchical mortality data, but Bayesian techniques with the aid of the Markov chain Monte Carlo (MCMC) simulation methods have successfully overcome these and fit spatiotemporal random effects for reasonably sized geo-locations. However, as the number of geo-locations increases, MCMC computations become infeasible or extremely slow, which is a norm in Big Data Analytics (BDA). This problem is popularly known as the “big m” or “big N”.
Epidemiological models for describing how a disease spreads through a population have been extremely useful to reduce the number of individuals who get sick or even die from illness. Developing meaningful and useful models is not easy however. In this paper, we first motivate the use of agent-based modelling and secondly, we present common challenges associated with agent-based modelling (of HIV) and our approaches to dealing with them.
T. b. rhodesiense is the acute form of African human trypanosomiasis or sleeping sickness which is common in East and Southern Africa. Trypanosomiasis is caused by the parasite Trypanosoma brucei and transmitted by tsetse flies (genus Glossina spp). Treatment of livestock in sub-Saharan Africa with trypanocidal drugs has been hindered by drug resistance and proves to be too expensive for many farmers. Tsetse control methods include aerial and ground spraying, sterile insect technique, and bait technology, including the use of insecticide-treated cattle (ITC). We compared two techniques of application of insecticides on cattle using a mathematical model: whole-body (WB), where insecticides are applied on the entire animals body and restricted application (RAP), where insecticides are applied on the legs, belly and ears of the animal.
The currently used mathematical models for medical treatment at the individual or population level are largely phenomenological and have limited quantitative predictive power. In 2013 Prof J. Snoep took up the SACEMA SARCHI Chair in mechanistic modelling of health and epidemiology. The task of the chair is to provide a mechanistic modelling approach with more predictive strength to pharmaceutical drug and intervention steps for individual and public health compared to current models. In this contribution an overview of the project is given and some of the work performed in the first year is highlighted.
Reliable mortality data are essential for planning health interventions, yet such data are often not available or reliable in developing countries, especially in sub-Saharan Africa. Health and socio-demographic surveillance sites, such as Agincourt in South Africa, are often the only way to assess and prospectively understand health trends at a population level, and thus have the potential to address this gap. This article summarises the main findings from my PhD in which advanced methods were applied to better understand the dynamics of age-specific mortality both in space and time, to identify age-specific mortality risk factors which have a high “impact” at a population level, and to relate inequalities in risk factor distributions to observed spatial mortality risk patterns.
The importance of concurrency (overlapping sexual partnerships in which sexual intercourse with one partner occurs between two acts of intercourse with another partner) in driving HIV transmission in hyperendemic settings remains controversial. A modelling study concluded that the role of concurrency in accelerating the spread of HIV is dramatically reduced by coital dilution (the reduction in frequency of sex acts per sexual partner, as a result of acquiring additional partners). We recently examined self-reported data on coital frequency and condom use during monogamous and concurrent relationship episodes from a survey in three communities with a high HIV prevalence. A key question in our analysis was if there is evidence for coital dilution and/or increased condom use during episodes of concurrency.
All models are imperfect, but unless the model both accounts for the known biology of the disease and is challenged with data this might not be detected.