living matter lab

Me233announcement.jpg

Contents

fall 20 - me233 - data-driven modeling of covid-19

ME233aa.jpg
ME233b.jpg

me 233 - data-driven modeling of covid-19

ellen kuhl kevin linka
amelie schafer oguz tikenogullari

announcement, syllabus

fall 2020
tue thu 8:30-9:50

 

objectives

understanding the outbreak dynamics of covid-19 through the lens of mathematical models is an elusive but significant goal. within only half a year, the covid-19 pandemic has resulted in more than 20 million reported cases across 188 countries with more than 700,000 deaths worldwide. this has generated an unprecedented volume of data; yet, the precise role of mathematical modeling in providing quantitative insight into the covid-19 pandemic remains a topic of ongoing debate. this course discusses how to design computational tools to understand the covid-19 outbreak. we focus on mathematical epidemiology, infectious disease models, concepts of the effective reproduction number and herd immunity, network modeling, outbreak dynamics and outbreak control, bayesian methods, model calibration and validation, prediction and uncertainty quantification. we highlight the early success of classical models for infectious diseases and show why these models fail to predict the current outbreak dynamics of covid-19. we illustrate how data-driven modeling can integrate classical epidemiology modeling and machine learning to infer critical disease parameters—in real time—from reported case data to make informed predictions and guide political decision making. we critically discuss questions that current models can and cannot answer and showcase controversial decisions around the early outbreak dynamics, outbreak control, and exit strategies.

in the news

a new class explores how to safely reopen a campus during covid-19 stanford report reopening paper
data-driven modeling of covid-19-lessons learned. webinar announcement webinar paper
md testifies about covid-19 transmission risks of lifting NL travel ban cbc cmech paper
nl doctor testifies about covid-19 transmission risks of lifting travel ban atlantic ctv news cmech paper
pandemie: des chercheurs du RU s'interessent au cas de TNL radio canada medRxiv paper
how covid-19 spread has been contained by travel bans science daily infosurhoy cmbbe paper
is it safe to lift covid-19 travel bans? medical life sciences the newfoundland story
covid-19 travel restrictions have saved millions of people health24 medRxiv Rvalues
how covid-19 spread has been contained by travel bans press release cmbbe paper mobility paper
using math to understand covid-19 outbreak dynamics stanford medicine bmmb paper asymptomatic
predict covid's spread and recovery stanford engineering medRxiv in the us medRxiv in europe

SEIR network model of phased campus reopening

Stanford.jpg

grading

  • 15 % homework 01 - mathematical epidemiology of covid-19
  • 15 % homework 02 - early outbreak dynamics of covid-19
  • 15 % homework 03 - outbreak control of covid-19
  • 25 % final project - presentation graded by class
  • 30 % final project - report graded by instructors

final project

for the final project, you will make informed decisions about stanford campus re-opening. this project combines tools of solving partial differential equations, SEIR compartment modeling, network modeling, uncertainty quantification, and python + PyMC3 towards creating a dynamic network SEIR model. guided by the example of lifting travel restrictions in the canadian province of newfoundland, you will create an influx network of incoming students by country or state and estimate their influx dynamics using an SIR or SEIR model. based on the current regional disease dynamics at the end of the fall quarter you will estimate the number of infectious incoming frosh, sophomores, juniors, and seniors. this will allow you to forecast the SEIR populations on campus and discuss their confidence intervals. with these results, you will make informed recommendations of gradual campus reopening and inviting different subgroups of students back to campus.

syllabus

day date topic slides read hw
tue sept 15 introduction to covid-19 s01
thu sept 17 I. infectious disesaes - a brief history s02 r02
tue sept 22 II. mathematical epidemiology - 1. intro to compartment modeling s03 r02
thu setp 24 II. mathematical epidemiology - 2. compartment models s04 r04 h01
tue sept 29 II. mathematical epidemiology - 3. endemic disease modeling s05 r05
thu oct 01 III. data-driven modeling - 1. compartment modeling of covid-19 s06 r06
tue oct 06 III. data-driven modeling - 2. early outbreak dynamics of covid-19 s07 r07
tue oct 08 III. data-driven modeling - 3. asymptomatic transmission of covid-19 s08 r08 h02
tue oct 13 III. data-driven modeling - 4. inferring outbreak dynamics of covid-19 s09 r09
thu oct 15 IV. modeling outbreak control - 1. managing infectious diseases s10 r10
tue oct 20 IV. modeling outbreak control - 2. change-point modeling of covid-19 s11 r11
thu oct 22 IV. modeling outbreak control - 3. dynamic modeling of covid-19 s12 r12 h03
tue oct 27 V. network modeling - 1. network modeling of epidemiology s13 r13
thu oct 29 V. network modeling - 2. network modeling of covid-19 s14 r14
tue nov 03 V. network modeling - 3. dynamic network modeling of covid-19 s15 r15
thu nov 05 VI. informing political decision making - exit strategies s16 r12 p01
tue nov 10 VI. informing political decision making - vaccination strategies s17 r17
thu nov 12 VI. informing political decision making - the second wave s18 r18
tue nov 17 data-driven modeling of covid-19 - lessons learned s19 r19
thu nov 19 data-driven modeling of covid-19 - final project presentations s20

python files

we'll share the python files for the projects here

homework 01

learn mathematical epidemiology of covid-19
tools solving ordinary differential equations, time integration, python
apps classical SEIR model, alpha, beta, gamma, Ro = beta/gamma
read soper a, the lessons of the pandemic, science xlix (1919) 501-505

  • identify three similarities between the pandemic in 1918 and the covid-19 today
  • identify three differences between the pandemic in 1918 and the covid-19 today
  • discuss, in two sentences for each, why
  • implement SEIR model in python
  • time integration: plot SEIR over t for alpha = 2.5 /days, gamma = 6.5 /days, beta = 13 /days
  • vary the time step size delta t; discuss the results
  • sensitivity analysis: plot SEIR over t for alpha = 2.5 /days; gamma = 6.5 /days
  • vary beta such that Ro = beta / gamma = [ 2.0, 3.0, 4.0, 5.0 ], discuss the results

homework 02

learn early outbreak dynamics of covid-19
tools data-driven modeling, parameter identification, python
apps basic reproduction number Ro = beta/gamma
read liu y et al., the reproduction number of covid-19. j travel med (2020)

  • identify the range of the basic reproduction number for covid-19
  • compare it against classical infectious diseases and other coronavirus diseases
  • download covid-19 data for your home country, state, or city from an online database
  • plot the time evolution of I and R
  • fit the static SEIR model from homework 01 to the data
  • advise: you may fix alpha and gamma, and only fit beta and eta=N*/N
  • plot the recorded I and R and the modeled SEIR over t
  • discuss your results, how good is the fit, when is it best and why
  • calculate the basic reproduction number Ro = beta / gamma
  • compare your Ro against other reported values for covid-19 and other infectious diseases
  • for your Ro, calculate the herd immunity threshold

homework 03

learn outbreak control of covid-19
tools combining machine learning and mechanistic modeling, python + PyMC3
apps effective reproduction number R(t) = beta(t) / gamma
read oden tj, adaptive multiscale predictive modeling. acta numerica (2018) 353-450

  • explain bayes rule, eq (4.1), in your own words
  • interpret its four terms in relation to this homework
  • download covid-19 data for your home country, state, or city from an online database
  • fit the dynamic SEIR model to the data
  • combine dynamic SEIR model with bayesian methods to learn R(t) for the same data as in homework 02
  • plot the recorded I and R and the modeled SEIR over t
  • discuss your results, how does the fit differ from homework 02
  • plot the effective reproduction number R(t) = beta(t) / gamma
  • plot and discuss confidence intervals
  • comment on trends, e.g., shelter in place, lockdown, travel restrictions

final project

learn informing decision making
tools solving partial differential equations, network modeling, python + PyMC3
apps dynamic network SEIR model
read linka k et al. is it safe to lift covid-19 travel bans. comp mech. doi:10.1007/s00466-020-01899-x

  • guided by the newfoundland example, develop your own model for stanford campus reopening
  • create an influx network, e.g., use origin of incoming students, by country or state
  • estimate influx dynamics, select an epidemiology model, e.g., SIR or SEIR
  • use current dynamics of incoming locations from open source database or model
  • estimate number of incoming infectious frosh, sophomores, juniors, seniors
  • forecast SEIR populations on campus and discuss confidence intervals
  • make informed recommendations of gradual campus reopening

additional reading

(r02) bar-on et al., sars-cov-2 (covid-19) by the numbers, elife 9 (2020) e57309.
(r03) bauer f, compartment models in epidemiology, mathematical epidemiology (2008) 19-79.
(r04) hethcode hw, the mathematics of infectious disease, siam review 42 (2020) 599-653.
(r05) dietz k, the estimation of the basic reproduction number for infectious diseases, stat meth med res 2 (1993) 23-41.
(r06) peirlinck m, et al. outbreak dynamics of covid-19 in china and the united states. medRxiv: doi:10.1101/2020.04.06.20055863.
(r07) park et al., reconciling early-outbreak estimates of the basic reproduction number and its uncertainty. j royal soc interface 17 (2020) 20200144.
(r08) ioannidis j, the invection fatality rate of covid-19 inferred from seroprevalence data, medRxiv, doi:10.1101/2020.05.13.20101253.
(r09) peirlinck m et al., visualizing the invisible: the effect of asymptomatic transmission, medRxiv, doi: 10.1101/2020.05.23.20111419
(r10) wilder-smith a, freedman do. isolation, quarantine, social distancing and community containment, j travel med (2020) 1-4.
(r11) dehning et al., inferring change points in the spread of covid-19 reveals the effectiveness of interventions, science doi:10.1126/science.abb9789.
(r12) linka et al., the reproduction number of covid-19 and its correlation with public health interventions, comp mech, doi:10.1007/s00466-020-01880-8.
(r13) pastor-satorras r et al., epidemic processes in complex networks, rev mod phys 87 (2015) 926-973.
(r14) linka k et al. outbreak dynamics of covid-19 in europe and the effect of travel restrictions. comp meth biomech biomed eng; 2020; 23:710-717.
(r15) linka k et al. global and local mobility as a barometer for covid-19 dynamics. medRxiv doi:2020.06.13.20130658.
(r16) linka k et al. is it safe to lift covid-19 travel bans. the newfoundland story. comp mech doi:10.1007/s00466-020-01899-x.
(r17) anderson rm, may rm. directly transmitted infectious diseases: control by vaccination, science 215 (1982) 1053-1060.
(r18) grassly nc, fraser c, seasonal infectious disease epidemiology, proc royal soc b 273 (2006) 2541-2550.
(r19) kuhl e. data-driven modeling of covid-19 - lessons learned. extr mech lett; doi:10.1016/j.eml.2020.100921.
(h01) soper a. the lessons of the pandemic, science xlix (1919) 501-505.
(h02) liu y et al., the reproduction number of covid-19 is higher compared to SARS coronavirus. j travel med (2020)
(h03) oden tj. adaptive multiscale predictive modeling. acta numerica (2018) 353-450.
(p01) linka k et al. is it safe to lift covid-19 travel bans. the newfoundland story. comp mech doi:10.1007/s00466-020-01899-x.