Member of medical staff using a mobile phone to report on malaria

Providing free to access and open-source software packages, code and webtools to make the work we do accessible to global policy makers, country decision makers and researchers.

Malaria transmission models


The Plasmodium falciparum malaria transmission model code is now publicly available:

The model can be run through R. The main goals of this software are to make an extensible, maintainable and fast simulation to evaluate and report on malaria parasite intervention strategies and their potential impact for controlling and eliminating the disease.

The model is defined within the malariasimulation package and is executed using the individual package. Vignettes are included to help demonstrate how different interventions could be incorporated into a simulation.

To help users of malariasimulation set up the model to represent a specific location or transmission setting there is a suite of packages to capture important site-characteristics. These include the seasonality of rainfall (, population demography (, and ITN coverage metrics (  associated with a setting.

Deterministic malaria model

In addition to this stochastic, individual-based implementation of the model, model code for the equivalent deterministic model, expressed as a set of differential equations, is also available: This version of the model, again in R, may be more suitable to use as a teaching tool, or for those who wish to familiarise themselves with the model, or who wish to edit it e.g. to create a stand-alone model that does not need the full functionality of the IBM.

Data analysis and decision-making tools


Plasmodium genomic data has proven itself as a source of useful data for malaria surveillance programmes. However, designing a study to obtain genomic data specifically for surveillance purposes can be a challenging task with many potential pitfalls. Questions like "should I obtain more samples or sequence more loci" are common, and unfortunately no out-of-the-box power analysis tools can provide answers to these questions due to the long and complex pipelines required to analyse genomic data. The SIMPLEGEN pipeline aims to make life easier by allowing researchers to test out different sample designs through simulation before committing costly resources. The project – currently in development – will allow a range of simulation models, and will be able to output genomic data in various common formats for use in down-stream tools. It is our hope that this tool will not only assist malaria control programmes in their decision-making, but will also lead to improved understanding of when and in where genomic data can add the most value to existing surveillance.


The novel interventions now available have contrasting public health impact given the characteristics of the local mosquito population, the level of disease and history of control interventions. The MINT tool has capacity to theoretically explore some of the potential outcomes of the deployment of multiple interventions given differences at baseline in different regions. The challenge for decision makers with limited budgets tasked with protecting communities from malaria is addressed by providing a tool within which these theoretical scenarios can be explored.

MINT is coded within a single page Javascript application written in Vue.js. An API that is coded within R is used to retrieve and process pre-run model results that can be stored in a lightning memory-mapped database so that simulations can be rapidly explored by the user.

MINT can be accessed here: (Release Date: Feb 2022).

A user guide is provided in English and French within the webtool interface.

Version 1 of this vector control decisions tool MINT is designed to help National Malaria Control Programs explore the most cost-effective current World Health Organisation (WHO) recommended ITN and IRS products for falciparum malaria control (link). Local human, mosquito and cost data are used to characterise the setting of interest (we refer to this setting as a zone in the tool). The tool then summarising the impact of different vector control interventions used in the zone and calculates the cost effectiveness of each of the intervention packages. A maximum budget can be set to help determine the most appropriate and affordable intervention for that particular zone according to local goals.

No model is as good as high-quality local surveillance data which should be collected and used to inform future policy. Predictions are made using average entomological and epidemiological data from throughout Africa so may not be representative of all settings. For example, IRS efficacy is thought to vary according to the type of wall material which will vary from site to site (1,2). These differences and uncertainties should be considered in any decision-making process. Similarly, predictions from the model are only as good as data used to parameterise them, so the better the local data the better the predictions are likely to be. Simplifications have been made (for example, in the range of endemicity settings that are explored) so individual estimates of impact and cost-effectiveness will be different from reality. Nevertheless, the relative difference between intervention options is likely to be more consistent and predictions of impact have been shown to adequately reflect changes in malaria prevalence observed in the field (see main manuscript).


Uragayala S, Kamaraju R, Tiwari S, Ghosh SK, Valecha N. Small-scale evaluation of the efficacy and residual activity of alpha-cypermethrin WG (250 g AI/kg) for indoor spraying in comparison with alpha-cypermethrin WP (50 g AI/kg) in India. Malar J [Internet]. 2015;14(1):223. Available from:

Etang J, Nwane P, Mbida J, Piameu M, Manga B, Souop D, et al. Variations of insecticide residual bio-efficacy on different types of walls: results from a community-based trial in south Cameroon. Malar J [Internet]. 2011;10(1):333. Available from:

General software tools


drjacoby is a user-friendly and flexible statistical software R package for running Markov chain Monte Carlo (MCMC) for statistical inference. drjacoby allows the user to submit their down, custom likelihood and prior functions written in the R or c++ programming languages. drjacoby is specially tailored for to work well for MCMC problems that mix poorly due to highly correlated and/or multi-modal posteriors.

For more information an introduction to using drjacoby please see:


Charles G, Winskill P, Topazian H, Challenger J, Fitzjohn R (2022). malariasimulation: An individual based model for malaria. R package version 1.3.0.

Sherrard-Smith E, Winskill P, Hamlet A, Ngufor C, N'Guessan R, Guelbeogo MW, Sanou A, Nash RK, Hill A, Russell EL, Woodbridge M, Tungu P, Kont MD, McLean T, Fornadel C, Richardson JH, Donnelly MJ, Staedke SG, Gonahasa S, Protopopoff N, Rowland M, Churcher TS. 2022, Optimising the deployment of vector control tools against malaria: a data-informed modelling study, The Lancet Planetary Health, Vol: 6, Pages: e100-e109, ISSN: 2542-5196