Stochastic Modeling of Infectious Diseases The 34 th Quality and - - PowerPoint PPT Presentation
Stochastic Modeling of Infectious Diseases The 34 th Quality and Productivity Research Conference - 2017 Volodymyr Serhiyenko vserhiyenko@metabiota.com June 15 th 2017 Agenda Hi Historical cal Examp xample Metabiota Overview
Volodymyr Serhiyenko vserhiyenko@metabiota.com June 15th 2017
The 34th Quality and Productivity Research Conference - 2017
Historical cal Examp xample
Coronavirus Risk Model
was treating “atypical pneumonia” in Guangdong province (China) arrived in Hong Kong to attend a wedding and stayed in Hot
Hotel l Me Metr tropole
Next da day he felt ill and was admitted to the intensive care unit
mysterious respiratory disease of unk unkno nown n ori rigin n
20 cases ses were associated with the
transmission on 9th floor started from the index patient who spent only on
Source: Christopher R. Braden, Scott F. Dowell, Daniel B. Jernigan, and James M. Hughes - Emerging Infectious Diseases Journal, Volume 19, Number 6—June 2013
9th floor layout of the Ho Hotel l Metropole le in Hong Kong
for consequent large outbreaks in
Ca Canada, Vi Vietnam, Si Sing ngapore,
and Ho
Hong Kong itself
Carlo Urbani, a WHO physician, recognized a new and highly
contagious disease. He later became infected and died, but his ea early warni ning ng started a massive response worldwide
Co Count untry Ca Cases es Fa Fatal
China 5327 349 Hong Kong 1755 299 Taiwan 346 73 Canada 251 43 Singapore 238 33 Vietnam 63 5 USA 27 Philippines 14 2 Other 75 6 TO TOTAL 8096 810
Sev Sever ere e ac acute respirat atory syndrome me (SARS)
16 No November 2002 2002 and ended in Taiwan on 5 5 Jul uly 2003 2003 (spreading to 27 27 count untries es)
Super er-Sp Sprea eading ng Eve vents ts, like one in Metropolitan Hotel, had been recorded
SARS-Co CoV spread to humans from wild pa palm civet ca cats ts that are valued for their meat and are sold in Chinese markets.
bats are the na natur ural re reservoirs of SARS-like coronaviruses.
flex exible to consider different di disease s spe pecific ch charact cteristics cs like Super-Spreading Events, availability of vaccines, vaccination strategies (mass or ring), etc.
newly em emer erging ng infectious diseases
Globa bal co connect ctivi vity and tra travel el patterns play a crucial role in the spread and magnitude of the disease epidemic
Metabiot
Overvie iew
Coronavirus Risk Model
MERS (Korea, 2015)
$2B
Ebola (W. Africa, 2015)
$900M
Dengue Fever (Brazil, 2013)
Foot & Mouth (UK, 2001)
Avian Flu (US Midwest, 2015)
In the last decade, there have been over 470
human disease
SARS (Global, 2003)
modeling
Disease Spr pread d Mode deling
Coronavirus Risk Model
Metabiot
Alessand essandro Ve Vespignani and his colleagues from Northeastern University's Laboratory for the Mod Modeling of
Bi Biological and Socio-tec techni hnical Sy System stems
Global al E Epidemic an and Mo Mobility ty mode model l (GLE LEaM) that stochastically simulates the spread
models to realistically simulate disease spark, spread, and duration of epidemics
around transportation hubs (i.e. airports). The resulting network consists of 3,
3,362 362
geographic subpopulations + Full ai airline transportation data + Sho Short rt-ra rang nge mobility network
S
Susceptible
E
Exposed
I
Infected
R
Removed Underlying Compartmental Model
Balcan, D., Gonçalves, B., Hu, H., Ramasco, J. J., Colizza, V., & Vespignani, A. (2010). Modeling the spatial spread of infectious diseases: The GLobal Epidemic and Mobility computational
in indiv ivid iduals ls through each compartment by stoc stocha hasti stically si simul ulati ting ng values from binomial and multinomial distributions
𝑁𝑣𝑚𝑢𝑗𝑜𝑝𝑛𝑗𝑏𝑚 𝐹
+ 𝑢 , 𝑞./→1/
2, 𝑞./→1/ 3, 𝑞./→1/ 43
𝐶𝑗𝑜𝑝𝑛𝑗𝑏𝑚 𝑇
+ 𝑢 , 𝑞7/→./
Preparedness Index and Co Coronavirus Risk Model
SARS 2003 2003 outbreak: 8096
8096 cases, 810 810 deaths, 27 27 countries effected
Middle le East Respir iratory Syndrome (MERS) 2013 outbreak: 1980
1980 cases, 699 699 deaths, 15 15
countries effected (as of June 6, 2017) – cased by novel MERS-CoV
Source: de Wit et al., SARS and MERS: recent insights into emerging coronaviruses, 2016
2012, still on
going
(80% of total cases)
South Korea 2015 MERS outbreak Caused by on
182 182 cases with 37 37 deaths
R0 - basic reproductive number (number
𝜗9: - incubation period 𝜈9: - infectious period Travel Reduction (%) Transmissibility reduction time, etc.
𝑂𝑣𝑛𝑐𝑓𝑠 𝑝𝑔 𝑡𝑓𝑑𝑝𝑜𝑒𝑏𝑠𝑧 𝑑𝑏𝑡𝑓𝑡 ~ 𝑂𝑓𝐶𝑗𝑜𝑝𝑛𝑗𝑏𝑚(𝑆J, 𝑙)
Hospital beds per capita
Source: World Bank
If outbreak starts in US
USA, is it
going to be different from Ch
Chin ina
Leone outbreak of the
same disease?
How do we capture these differences?
Country-level differences in
measures
(1 (1=most p prepar ared, 4 , 4=leas ast p prepar ared)
PHI: Public Health Infrastructure PI: Physical and Communications Infrastructure IC: Institutional Capacity EF: Economic Factors PHC: Public Health Communications
On average, improving country’s Epidemic Preparedness by one unit is decreasing odds of dying by 28%
Pr Proprietary Da Data Set
continuously updated Disease Model Li Library
weekly resolution informed the event catalog
subpopulations
hours to date
Data Layers Bioclimatic Data Number of shared human-bat viruses Zoonotic mammal species Proximity to large cities Human Density Bat (Taphozous sp.) Dromedary Camel abundance PREDICT (Metabiota) data