AMR Spread

Welcome to the AMR Spread user interface!

Using a mechanistic deterministic multi-country model of the transmission of antibiotic-resistant bacteria, simulate the spread of an emergent carbapenem-resistant E. coli worldwide.

Select the emergent resistant E. coli characteristics, national antibiotic consumption levels, international travel flows and index country where emergence starts. Then, simulate for 20 years the emergence corresponding to the selected scenario.

After simulation, visualize two outputs:

a map of antibiotic resistance proportion for healthy carriage

a timeline to visualize at what year surveillance system detects 5 antibiotic resistant infections by country

Run simulation

Emergent resistant bacteria characteristics Choose the characteristics of the emergent resistant bacteria for simulations.

National anitobiotics consumption Choose national antibiotics consumption: decrease, keep or increase consumption for the 20 years of simulation compared to true national antibiotic sales from IQVIA data.

INTERNATIONAL TRAVEL FLOW Choose proportion of travelers from a country traveling to any country worldwide: decrease, keep or increase travel flows for the 20 years of simulation compared to estimated travel matrix from KCMD data.

INDEX COUNTRY Choose one index country where emergence of carbapenem-resistant E. coli will start and let it spread to the whole world.
Optional: Choose a second index country with the possibility to compare with simulations of first index country.

AMR Model

MODEL USED FOR SIMULATIONS

We build a mechanistic deterministic between-host model of the transmission of antibiotic-resistant Escherichia coli. The model is calibrated using empirical data of endemic ESBL-producing E. coli. Then, the model is used to simulate the spread of an emergent resistant bacteria worldwide, such as carbapenem-resistant E. coli.

The model is initialized by defining the characteristics of an emerging resistant bacterium, which is introduced at the start of the simulation in a given index country. National antibiotic consumption levels in every country and travel fluxes between every country can be varied. The model simulates, over 20 years, the spread of the emerging resistant bacteria between individuals within a country, and between different countries connected by travel flows. A "simulation" thus corresponds to the spread of an emerging resistant bacterium from an index country to other countries over time.

Schematic of the between-host model used for simulations

EQUATIONS FOR THE MODEL

INPUT PARAMETERS


	NORMAL BUG SCENARIO	SUPER BUG SCENARIO	FITNESS COST BUG SCENARIO	MULTI-DRUG RESISTANT BUG SCENARIO
β_i	Estimated per country	β_i x 1.1	β_i x 0.9
σ_{S i}	Estimated per country
a_i	Estimated per country
σ_{R i}	σ_{R i} = a_i x σ_{S i}
α	8 days
α'	60 days
γ	1/120 days^-1
φ_ant	9.83
k	0.80
τ_y,i	Carbapenems treatment rate per year y and country i Data source: IQVIA			Carbapenems and β-lactams treatment rate per year y and country i Data source: IQVIA
Δ_y,i,j	travel matrix (% of inhabitants of country i traveling to country j per year y) Data source: KCMD

INITIAL CONDITIONS

CS = N_i x (1 - τ_1,i)
CSr = n
CsR = 0
CS^A = N_i x (τ_1,i)
CsR^A = 0

Ni is population size of country i (data: World Bank)
n is initial number of resistant bacteria carriers at the beginning of the simulation (n = 1000)

Data

DATA USED FOR SIMULATIONS

Antibiotic resistance data: Model calibrated using antibiotic resistance data from the Pfizer ATLAS database. https://atlas-surveillance.com/#/login

Antibiotic consumption data: Synthetic data representing trends of antibiotic consumption by country and by year.

Travel data: Data obtained from the Knowledge Center for Migration and Demography (KCMD) data hub, from the Global Transnational Mobility dataset. Data represent the estimated number of travelers country by country, for year periods between 2011-2016.
https://migration-demography-tools.jrc.ec.europa.eu/data-hub/