Evaluation of outbreak response immunization in the control of pertussis using agent-based modeling

Abstract

BackgroundPertussis control remains a challenge due to recently observed effects of waning immunity to acellular vaccine and suboptimal vaccine coverage. Multiple outbreaks have been reported in different ages worldwide. For certain outbreaks, public health authorities can launch an outbreak response immunization (ORI) campaign to control pertussis spread. We investigated effects of an outbreak response immunization targeting young adolescents in averting pertussis cases.MethodsWe developed an agent-based model for pertussis transmission representing disease mechanism, waning immunity, vaccination schedule and pathogen transmission in a spatially-explicit 500,000-person contact network representing a typical Canadian Public Health district. Parameters were derived from literature and calibration. We used published cumulative incidence and dose-specific vaccine coverage to calibrate the model’s epidemiological curves. We endogenized outbreak response by defining thresholds to trigger simulated immunization campaigns in the 10–14 age group offering 80% coverage. We ran paired simulations with and without outbreak response immunization and included those resulting in a single ORI within a 10-year span. We calculated the number of cases averted attributable to outbreak immunization campaign in all ages, in the 10–14 age group and in infants. The count of cases averted were tested using Mann–Whitney U test to determine statistical significance. Numbers needed to vaccinate during immunization campaign to prevent a single case in respective age groups were derived from the model. We varied adult vaccine coverage, waning immunity parameters, immunization campaign eligibility and tested stronger vaccination boosting effect in sensitivity analyses.Results189 qualified paired-runs were analyzed. On average, ORI was triggered every 26 years. On a per-run basis, there were an average of 124, 243 and 429 pertussis cases averted across all age groups within 1, 3 and 10 years of a campaign, respectively. During the same time periods, 53, 96, and 163 cases were averted in the 10–14 age group, and 6, 11, 20 in infants under 1 (p < 0.001, all groups). Numbers needed to vaccinate ranged from 49 to 221, from 130 to 519 and from 1,031 to 4,903 for all ages, the 10–14 age group and for infants, respectively. Most sensitivity analyses resulted in minimal impact on a number of cases averted.DiscussionOur model generated 30 years of longitudinal data to evaluate effects of outbreak response immunization in a controlled study. Immunization campaign implemented as an outbreak response measure among adolescents may confer benefits across all ages accruing over a 10-year period. Our inference is dependent on having an outbreak of significant magnitude affecting predominantly the selected age and achieving a comprehensive vaccine coverage during the campaign. Economic evaluations and comparisons with other control measures can add to conclusions generated by our work.

Evaluation of the effect of chickenpox vaccination on shingles epidemiology using agent-based modeling

“…Chickenpox vaccination parameters, such as those associated with primary vaccine failure and waning of vaccination immunity were derived from literature ( Gershon, Takahashi & Seward, 2012 ; Bonanni et al, 2013 ; Duncan et al, 2017 ). We built the mechanism whereby vaccine coverage was generated by our model based on distribution of vaccination probabilities and population vaccination attitudes as described by Doroshenko, Qian & Osgood (2016) . We classified all individuals into three groups: those who accept, reject and are hesitant to receive vaccination, and we assigned vaccination probabilities for each of these groups.…”

A multi-method simulation of a high-frequency bus line

“…The combination of discrete-event simulation and agentbased modelling is created in our work using the Anylogic tool. With this tool, several models have been created in different fields, for example for the analysis of power and performance of data centers [11], the outbreak response of immunization [4] and airport checkpoint pedestrian flow [8].…”

Impacts of observation frequency on proximity contact data and modeled transmission dynamics

“…Together with pathogen-specific parameters, high-fidelity representations of such contact networks within transmission models [22] can enable a much higher resolution view of the process of a disease spreading than is possible with the random mixing assumptions required in compartmental models within the traditional susceptibleinfectious-recovered (SIR) family [19,20,24]. Such a view can support real-time identification early of outbreaks and an estimation of the attack rate, as well as retrospective evaluation and assessment of improved effectiveness of altered vaccine schedules, aid in planning of interventions such as outbreak response immunization [25], public health orders and quarantine, and support assessment of the impact of the scope, speed, and breadth of contact tracing [26]. Transmission models structured with a detailed contact network aid inferencing of population-scale effects from individual-level behavior of infections by enabling characterization of the transmission of contagious diseases over the close-proximity contacts shaping outbreak dynamics [22,27].…”