Spatiotemporal dynamics of the Ebola epidemic in Guinea and implications for vaccination and disease elimination: a computational modeling analysis

Abstract

Background: Among the three countries most affected by the Ebola virus disease outbreak in 2014-2015, Guinea presents an unusual spatiotemporal epidemic pattern, with several waves and a long tail in the decay of the epidemic incidence.

Methods: Here, we develop a stochastic agent-based model at the level of a single household that integrates detailed data on Guinean demography, hospitals, Ebola treatment units, contact tracing, and safe burial interventions. The microsimulation-based model is used to assess the effect of each control strategy and the probability of elimination of the epidemic according to different intervention scenarios, including ring vaccination with the recombinant vesicular stomatitis virus-vectored vaccine.

Results: The numerical results indicate that the dynamics of the Ebola epidemic in Guinea can be quantitatively explained by the timeline of the implemented interventions. In particular, the early availability of Ebola treatment units and the associated isolation of cases and safe burials helped to limit the number of Ebola cases experienced by Guinea. We provide quantitative evidence of a strong negative correlation between the time series of cases and the number of traced contacts. This result is confirmed by the computational model that suggests that contact tracing effort is a key determinant in the control and elimination of the disease. In data-driven microsimulations, we find that tracing at least 5-10 contacts per case is crucial in preventing epidemic resurgence during the epidemic elimination phase. The computational model is used to provide an analysis of the ring vaccination trial highlighting its potential effect on disease elimination.

Conclusions: We identify contact tracing as one of the key determinants of the epidemic's behavior in Guinea, and we show that the early availability of Ebola treatment unit beds helped to limit the number of Ebola cases in Guinea.

Keywords: Computational modeling; Ebola epidemiology; Intervention strategies.

Fig. 1

Model validation and estimates. a Weekly number of cases over the period August 2014 – May 2015 according to the WHO Ebola situation report, patient database and Guinean Ministry of Health (bars) and predicted by the model (the blue line is the average, and the shaded blue region is the 95 % confidence interval of simulated epidemics). The red line, set on February 25, 2015, marks the end of the calibration period. b Predicted and observed cumulative number of cases by prefecture and region as of February 25, 2015. c Boxplot of the proportions of transmission by setting as of August 15, 2014, and February 25, 2015, in order to show the variation of these quantities. August 15, 2015, is chosen in such a way as to allow a comparison with the results for Liberia presented previously [5]; February 25, 2015, corresponds to the end of the calibration period

Fig. 2

Transmission tree and distribution of secondary infections. a Transmission tree of one randomly chosen epidemic obtained by simulating the calibrated model. Different colors represent the setting where the individual was infected. As the whole transmission tree would have been too wide to display on a page (thousands of edges on average), we show only one component. Simulated infections occurring over the period April 2014 to September 2014 are shown. b As a, but showing infections over the period April 2014 to November 2014. c As a, but showing infections over the period April 2014 to July 2015 (i.e., the entire simulated epidemic). d Distribution of the number of secondary cases by setting as obtained by the analysis of the whole transmission tree reported in panel c. e Degree distribution of contacts with members of the same household and of the extended family as resulting from the model

Fig. 3

Reproductive number over time. Rt as computed from the time series of cases obtained by simulating the calibrated model (details on the computation of Rt are reported in Additional file 1). Mean and 95 % credible interval are shown

Fig. 4

Correlations between interventions and number of cases. a Probability distribution of number of traced contacts per case over the period August 4, 2014, to February 25, 2015. b Cross-correlation between the average number of contacts included in contact tracing per case at a given time and the incidence of cases observed 0 to 40 days later. The average number of contacts included in contact tracing per case is computed as the sum of followed contacts over 21 days divided by the sum of new cases over the same 21 days (alternative definitions of contact tracing are considered in Additional file 1). The highest absolute value of cross-correlation is obtained for a lag of 17 days. c Red line: daily number of cases (as obtained with a moving average of 15 days, i.e., 1 week previous and 1 week following the data point) over time; blue line: number of traced contacts per case (defined as in b) over time; dotted line: linear model best fitting the number of traced contacts. Scale for blue and dotted curves is on the right axis. d Red line: as in c; blue line: probability of unsafe burials over time computed as the fraction of daily community safe burials over the daily total number of community burials (scale on the right axis); the curve is then obtained by computing a moving average of 15 days, i.e., 1 week previous and 1 week following the data point. e Red line: as in c; blue line: number of admissions to ETUs over time (scale on the right axis); the curve is then obtained by computing a moving average of 15 days, i.e., 1 week previous and 1 week following the data point. Dates in panels c–e refer to the period August 2014 to June 2015

Fig. 5

Disentangling the impact of different interventions. Boxplots for the cumulative number of cases from September 2014 through February 2015 assuming different interventions

Fig. 6

Impact of interventions on disease elimination. a Probability of disease elimination over time, assuming different values for the number of traced contacts per case since February 25, 2014. Blue bars represent a situation comparable to what was observed in April 2015. b As a but assuming ring vaccination starting on March 23, 2015, enrollment 3 days after the admission of ring index cases to ETU, 90 % vaccine efficacy, 6 days for vaccinated individuals to develop protective immunity, and vaccine administered to 90 % of adults (≥18 years old). In 50 % of rings, vaccine is administered with a delay of 21 days with respect to immediately vaccinated rings. c As b but vaccine is administered to 90 % of all individuals and all rings are vaccinated at the time of enrollment. The number of traced contacts matches the data (see Methods and Additional file 1) until February 25, 2015; then, it is assumed to be constant over time until the end of the simulation, at the level reported in the legend