Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Jun 23;14(1):95.
doi: 10.1186/s12916-016-0637-z.

Modeling the spread of polio in an IPV-vaccinated population: lessons learned from the 2013 silent outbreak in southern Israel

Rami Yaari  1   2 , Ehud Kaliner  3 , Itamar Grotto  3   4 , Guy Katriel  5 , Jacob Moran-Gilad  3   4 , Danit Sofer  6 , Ella Mendelson  6   7 , Elizabeth Miller  7   8 , Amit Huppert  9   7 ; POG group; E Anis, E Kopel, Y Manor, O Mor, L Shulman, R Singer, M Weil
Affiliations

Modeling the spread of polio in an IPV-vaccinated population: lessons learned from the 2013 silent outbreak in southern Israel

Rami Yaari et al. BMC Med. .

Erratum in

Abstract

Background: Polio eradication is an extraordinary globally coordinated health program in terms of its magnitude and reach, leading to the elimination of wild poliovirus (WPV) in most parts of the world. In 2013, a silent outbreak of WPV was detected in Israel, a country using an inactivated polio vaccine (IPV) exclusively since 2005. The outbreak was detected using environmental surveillance (ES) of sewage reservoirs. Stool surveys indicated the outbreak to be restricted mainly to children under the age of 10 in the Bedouin population of southern Israel. In order to curtail the outbreak, a nationwide vaccination campaign using oral polio vaccine (OPV) was conducted, targeting all children under 10.

Methods: A transmission model, fitted to the results of the stool surveys, with additional conditions set by the ES measurements, was used to evaluate the prevalence of WPV in Bedouin children and the effectiveness of the vaccination campaign. Employing the parameter estimates of the model fitting, the model was used to investigate the effect of alternative timings, coverages and dosages of the OPV campaign on the outcome of the outbreak.

Results: The mean estimate for the mean reproductive number was 1.77 (95 % credible interval, 1.46-2.30). With seasonal variation, the reproductive number maximum range was between zero and six. The mean estimate for the mean infectious periods was 16.8 (8.6-24.9) days. The modeling indicates the OPV campaign was effective in curtailing the outbreak. The mean estimate for the attack rate in Bedouin children under 10 at the end of 2014 was 42 % (22-65 %), whereas without the campaign the mean projected attack rate was 57 % (35-74 %). The campaign also likely shortened the duration of the outbreak by a mean estimate of 309 (2-846) days. A faster initiation of the OPV campaign could have reduced the incidence of WPV even if a lower coverage was reached, at the risk of prolonging the outbreak.

Conclusions: OPV campaigns are essential for interrupting WPV transmission, even in a developed country setting with a high coverage of IPV. In this setting, establishing ES of WPV circulation is particularly crucial for early detection and containment of an outbreak.

Keywords: Inactivated polio vaccine; Model fitting; Oral polio vaccine; Polio; Transmission model; Vaccination strategies.

PubMed Disclaimer

Figures

Fig. 1
Fig. 1
Diagram describing the transmission model compartments. The solid arrows denote the transitions related to infection with wild polio virus. The dashed arrows denote transitions related to vaccination with oral polio vaccine (OPV). The dotted arrow entering the group S1 denotes births. Individuals in the population can be in one of four general states: susceptible (S), exposed (E), infectious (I) or recovered (R). In addition, susceptible individuals are divided into three groups: S1 includes individuals that did not yet receive OPV, S2 includes individuals that received one dose of OPV but remained susceptible, and S3 includes individuals that received two doses of OPV but remained susceptible
Fig. 2
Fig. 2
Posterior distributions obtained for the model parameters using MCMC: a Posterior distribution for the mean reproductive number. b Posterior distribution for the mean infectious period. c Posterior distribution for the amplitude of seasonal variation in transmission. The red curve shows the prior distribution based on the variation in the amplitude in 10 southern US states (Table 1 and Additional file 1). The posterior distribution is shifted left from the prior distribution with a mean of 0.57 compared to a mean of 1 in the prior distribution. d Posterior distribution for the peak time of seasonal variation in transmission. The red curve shows the prior distribution based on the variation in the peak time in 10 southern US states (Table 1 and Additional file 1). The posterior distribution is shifted left from the prior distribution, with a mean peak day of 138 (May 18) compared to a mean of 156 (June 5) in the prior distribution. e Posterior distribution for the per-dose efficacy of OPV. The red curve shows the prior distribution based on [23] (Table 1). The posterior distribution is shifted right from the prior distribution, with a mean efficacy of 0.63 compared to a mean of 0.56 in the prior distribution. f Posterior distribution for the start time of the outbreak
Fig. 3
Fig. 3
Top panel: Results of the ES. Red line indicates a positive finding of WPV1 in one of the four relevant sites (see ‘Methods’). Green line indicates no positive findings. Data shown here is up to the end of April 2014, after which there were no positive findings of WPV1 in any sewage sample. Bottom panel: The fit of the model to the stool samples data. The grey area marks the estimated 95 % credible interval of WPV1 prevalence in the modeled population of Bedouin children. The yellow area within the grey area presents a more restricted estimated range of WPV1 prevalence using parameter values whose log-likelihood is within 2 log-likelihood units of the best fit (a commonly used threshold for selection of the more probable fits to the data [34]). The blue x marks the proportion of stool samples positive for WPV1 in each of the days that samples were collected. The magenta dots present a weekly smoothing of the sampled data. For a description of the smoothing and the confidence intervals related to the stochasticity of the observation of the stool data see Figure S7 in Additional file 1. While the likelihood was calculated using the non-smoothed stool sample data, the smoothed data captures the trend of the estimated prevalence better, as it blends in the effect of days with zero positive samples, of which there were many in the second stool survey due to the low number of samples taken each day. The red lines show the cumulative vaccine coverage (right y-axis) of the first (solid line) and second (dashed line) OPV doses in the modeled population
Fig. 4
Fig. 4
Top panel: 1000 plots of the value of the reproductive number (R) in time, calculated using Eq. S2 and S4 in Additional file 1 with 1000 values of R¯, δ and ϕ, randomly sampled out of the values obtained by the MCMC. The range includes plots with no or weak seasonal variation in which R=R¯~1.8 (blue curves showing results for δ ≤ 0.1), plots with strong seasonal variation in which R varies from a minimum of close to zero during winter to a maximum of around six during late spring – early summer (red curves showing results for δ ≥ 1) and everything in between (green curves). Middle panel: 95 % CI of WPV1 prevalence with the oral polio vaccine (OPV) campaign (dark grey) and without the OPV campaign (light grey). Bottom Panel: The outcome without the OPV campaign (light grey area in middle panel) depends on the estimated strength of the seasonality. The dashed blue lines depict a subset range of the prevalence without the OPV campaign obtained using weak or no seasonality (δ ≤ 0.1), while the red dotted-dashed lines show a subset range of prevalence without the OPV campaign obtained using strong seasonality (δ ≥ 1.0). The range obtained using weak seasonality consists of a single long wave, with a tail possibly extending into the first half of 2014, whereas the range obtained using strong seasonality consists of a shorter wave in 2013, with the possibility of a second wave during the second half of 2014. a The posterior distribution of the overall attack rate at the end of 2014 with (dark grey bars) and without (light grey bars) the OPV campaign. b The posterior distribution for the end time of the outbreak showing the probability of the outbreak ending on a particular month with (dark grey bars) and without (light grey bars) the OPV campaign. With the campaign the model estimates the outbreak ended sometime between January 2014 and October 2014. Without the OPV campaign the model projects the outbreak could have lasted until November 2016 (Table 2)
Fig. 5
Fig. 5
The effect of alternative OPV campaign scenarios on the cumulative incidence of the outbreak (a) and the outbreak duration (b). The outbreak duration was defined as the time when the incidence in the model drops below 10 infected individuals in order to take into account the probability of a stochastic fade-out of the outbreak during the periods of low transmissibility. Color bars show the mean and error bars show the 95 % CI obtained for each scenario by running the model using the 95 % best fitting parameter values given by the MCMC simulation. The simulations were run for up to five years
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. IMB. Independent Monitoring Board of the Global Polio Eradication Initiative - Twelfth Report. 2015.
    1. Anis E, Kopel E, Singer S, Kaliner E, Moerman L, et al. Insidious reintroduction of wild poliovirus into Israel, 2013. Eurosurveillance. 2013;18:20586. doi: 10.2807/1560-7917.ES2013.18.38.20586. - DOI - PubMed
    1. Kaliner E, Moran-Gilad J, Grotto I, Somekh E, Kopel E, Gdalevich M, et al. Silent reintroduction of wild-type poliovirus to Israel, 2013–risk communication challenges in an argumentative atmosphere. Eurosurveillance. 2014;19:20703. doi: 10.2807/1560-7917.ES2014.19.7.20703. - DOI - PubMed
    1. Manor Y, Shulman L, Kaliner E. Intensified environmental surveillance supporting the response to wild poliovirus type 1 silent circulation in Israel, 2013. Eurosurveillance. 2014;19:20708. doi: 10.2807/1560-7917.ES2014.19.7.20708. - DOI - PubMed
    1. Shulman L, Gavrilin E, Jorba J, Martin J, Burns C, et al. Molecular epidemiology of silent introduction and sustained transmission of wild poliovirus type 1, Israel, 2013. Eurosurveillance. 2014;19:20709. doi: 10.2807/1560-7917.ES2014.19.7.20709. - DOI - PubMed

MeSH terms

Substances