Estimating the relative importance of epidemiological and behavioural parameters for epidemic mpox transmission: a modelling study

Abstract

Background: Many European countries experienced outbreaks of mpox in 2022, and there was an mpox outbreak in 2023 in the Democratic Republic of Congo. There were many apparent differences between these outbreaks and previous outbreaks of mpox; the recent outbreaks were observed in men who have sex with men after sexual encounters at common events, whereas earlier outbreaks were observed in a wider population with no identifiable link to sexual contacts. These apparent differences meant that data from previous outbreaks could not reliably be used to parametrise infectious disease models during the 2022 and 2023 mpox outbreaks, and modelling efforts were hampered by uncertainty around key transmission and immunity parameters.

Methods: We developed a stochastic, discrete-time metapopulation model for mpox that allowed for sexual and non-sexual transmission and the implementation of non-pharmaceutical interventions, specifically contact tracing and pre- and post-exposure vaccinations. We calibrated the model to case data from Berlin and used Sobol sensitivity analysis to identify parameters that mpox transmission is especially sensitive to. We also briefly analysed the sensitivity of the effectiveness of non-pharmaceutical interventions to various efficacy parameters.

Results: We found that variance in the transmission probabilities due to both sexual and non-sexual transmission had a large effect on mpox transmission in the model, as did the level of immunity to mpox conferred by a previous smallpox vaccination. Furthermore, variance in the number of pre-exposure vaccinations offered was the dominant contributor to variance in mpox dynamics in men who have sex with men. If pre-exposure vaccinations were not available, both the accuracy and timeliness of contact tracing had a large impact on mpox transmission in the model.

Conclusions: Our results are valuable for guiding epidemiological studies for parameter ascertainment and identifying key factors for success of non-pharmaceutical interventions.

Keywords: Global sensitivity analysis; Infectious disease modelling; Mpox.

Fig. 1

Cumulative infections in kindergarten-going children plotted against transmission probabilities. The clear vertical trend in the plot against transmission per non-sexual contact (left) indicates that cumulative infections in kindergarten-going children are very sensitive to this parameter, whereas the horizontal strata in the plot against transmission per sexual contact (right) on the right show that variance in this parameter is not a major contributor to variance in the number of cumulative infections in kindergarten-going children

Fig. 2

Cumulative infections in the MSM community plotted against contact tracing parameters. The plot shows cumulative infections in the MSM community plotted against (left) delay in tracing with colour indicating base probability of being traced and (right) base probability of being traced with colour indicating delay in tracing in analysis 2.2. The clear vertical trends indicate the sensitivity of the infection dynamics in the MSM community to both tracing parameters of cumulative infections in the MSM community to both these parameters in analysis 2.2