Evaluation of the US COVID-19 Scenario Modeling Hub for informing pandemic response under uncertainty
Nature Communications,
Год журнала:
2023,
Номер
14(1)
Опубликована: Ноя. 20, 2023
Our
ability
to
forecast
epidemics
far
into
the
future
is
constrained
by
many
complexities
of
disease
systems.
Realistic
longer-term
projections
may,
however,
be
possible
under
well-defined
scenarios
that
specify
state
critical
epidemic
drivers.
Since
December
2020,
U.S.
COVID-19
Scenario
Modeling
Hub
(SMH)
has
convened
multiple
modeling
teams
make
months
ahead
SARS-CoV-2
burden,
totaling
nearly
1.8
million
national
and
state-level
projections.
Here,
we
find
SMH
performance
varied
widely
as
a
function
both
scenario
validity
model
calibration.
We
show
remained
close
reality
for
22
weeks
on
average
before
arrival
unanticipated
variants
invalidated
key
assumptions.
An
ensemble
participating
models
preserved
variation
between
(using
linear
opinion
pool
method)
was
consistently
more
reliable
than
any
single
in
periods
valid
assumptions,
while
projection
interval
coverage
near
target
levels.
were
used
guide
pandemic
response,
illustrating
value
collaborative
hubs
Язык: Английский
Characterising information gains and losses when collecting multiple epidemic model outputs
Epidemics,
Год журнала:
2024,
Номер
47, С. 100765 - 100765
Опубликована: Март 27, 2024
Collaborative
comparisons
and
combinations
of
epidemic
models
are
used
as
policy-relevant
evidence
during
outbreaks.
In
the
process
collecting
multiple
model
projections,
such
collaborations
may
gain
or
lose
relevant
information.
Typically,
modellers
contribute
a
probabilistic
summary
at
each
time-step.
We
compared
this
to
directly
simulated
trajectories.
aimed
explore
information
on
key
quantities;
ensemble
uncertainty;
performance
against
data,
investigating
potential
continuously
from
single
cross-sectional
collection
results.
July
2022
projections
European
COVID-19
Scenario
Modelling
Hub.
Five
modelling
teams
projected
incidence
in
Belgium,
Netherlands,
Spain.
by
incidence,
peaks,
cumulative
totals.
created
drawn
all
trajectories,
ensembles
median
across
model's
quantiles,
linear
opinion
pool.
measured
predictive
accuracy
individual
trajectories
observations,
using
weighted
ensemble.
repeated
sequentially
increasing
weeks
observed
data.
evaluated
these
reflect
with
varying
By
modelled
we
showed
characteristics.
Trajectories
contained
right-skewed
distribution
well
represented
an
pool,
but
not
models'
quantile
intervals.
Ensembles
typically
retained
range
plausible
over
time,
some
cases
narrowed
excluding
shapes.
several
gains
rather
than
distributions,
including
for
updated
collection.
The
value
losses
vary
collaborative
effort's
aims,
depending
needs
projection
users.
Understanding
differing
methods
collect
can
support
accuracy,
sustainability,
communication
infectious
disease
efforts.
All
code
data
available
Github:
https://github.com/covid19-forecast-hub-europe/aggregation-info-loss
Язык: Английский
Projecting the future impact of emerging SARS-CoV-2 variants under uncertainty: Modeling the initial Omicron outbreak
Epidemics,
Год журнала:
2024,
Номер
47, С. 100759 - 100759
Опубликована: Март 2, 2024
Over
the
past
several
years,
emergence
of
novel
SARS-CoV-2
variants
has
led
to
multiple
waves
increased
COVID-19
incidence.
When
Omicron
variant
emerged,
there
was
considerable
concern
about
its
potential
impact
in
winter
2021-2022
due
fitness.
However,
also
uncertainty
regarding
likely
questions
relative
transmissibility,
severity,
and
degree
immune
escape.
We
sought
evaluate
ability
an
agent-based
model
forecast
incidence
context
this
emerging
pathogen
variant.
To
project
cases
deaths
Indiana,
we
calibrated
our
hospitalizations,
deaths,
test-positivity
rates
through
November
2021,
then
projected
April
2022
under
four
different
scenarios
that
covered
plausible
ranges
Omicron's
Our
initial
projections
from
December
2021
March
indicated
a
pessimistic
scenario
with
high
disease
peak
weekly
Indiana
would
be
larger
than
previous
2020.
retrospective
analyses
indicate
severity
closer
optimistic
scenario,
even
though
hospitalizations
reached
new
peak,
fewer
occurred
during
peak.
According
results,
rapid
spread
consistent
combination
higher
transmissibility
escape
earlier
variants.
updated
starting
January
accurately
predicted
mid-January
decline
rapidly
over
next
months.
The
performance
shows
following
variant,
models
can
help
quantify
range
outbreak
magnitudes
trajectories.
Agent-based
are
particularly
useful
these
because
they
efficiently
track
individual
vaccination
infection
histories
varying
degrees
cross-protection.
Язык: Английский
Role of heterogeneity: National scale data-driven agent-based modeling for the US COVID-19 Scenario Modeling Hub
Epidemics,
Год журнала:
2024,
Номер
48, С. 100779 - 100779
Опубликована: Июнь 28, 2024
UVA-EpiHiper
is
a
national
scale
agent-based
model
to
support
the
US
COVID-19
Scenario
Modeling
Hub
(SMH).
uses
detailed
representation
of
underlying
social
contact
network
along
with
data
measured
during
course
pandemic
initialize
and
calibrate
model.
In
this
paper,
we
study
role
heterogeneity
on
complexity
resulting
epidemic
dynamics
using
UVA-EpiHiper.
We
discuss
various
sources
that
encounter
in
use
modeling
analysis
under
scenarios.
also
how
affects
computational
corresponding
simulations.
Using
round
13
SMH
as
an
example,
was
initialized
calibrated.
then
output
produced
by
can
be
analyzed
obtain
interesting
insights.
find
despite
model,
software,
computation
incurred
scenario
modeling,
it
capable
capturing
heterogeneities
real-world
systems,
especially
those
networks
behaviors,
enables
analyzing
epidemiological
outcomes
between
different
demographic,
geographic,
cohorts.
applying
disease
are
within
states,
demographic
groups,
which
attributed
population
demographics,
structures,
initial
immunity.
Язык: Английский
Asymmetric limits on timely interventions from noisy epidemic data
medRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2025,
Номер
unknown
Опубликована: Янв. 28, 2025
Abstract
Deciding
on
when
to
initiate
or
relax
an
intervention
in
response
emerging
infectious
disease
is
both
difficult
and
important.
Uncertainties
from
noise
epidemiological
surveillance
data
must
be
hedged
against
the
potentially
unknown
variable
costs
of
false
alarms
delayed
actions.
Here
we
clarify
quantify
how
case
under-reporting
latencies
ascertainment,
which
are
predominant
sources,
can
restrict
timeliness
decision-making.
Decisions
modelled
as
binary
choices
between
responding
not
that
informed
by
reported
curves
transmissibility
estimates
those
curves.
Optimal
responses
triggered
thresholds
numbers
estimate
confidence
levels,
with
set
various
choices.
We
show
that,
for
growing
epidemics,
sources
induce
additive
delays
hitting
any
case-based
multiplicative
reductions
our
estimated
reproduction
growth
rates.
However,
declining
these
have
counteracting
effects
limited
cumulative
impact
estimates.
find
this
asymmetry
persists
even
if
more
sophisticated
feedback
control
algorithms
consider
longer-term
interventions
employed.
Standard
therefore
provide
substantially
weaker
support
deciding
a
action
than
determining
it.
This
information
bottleneck
during
epidemic
may
justify
proactive
Язык: Английский
Evaluation and communication of pandemic scenarios
The Lancet Digital Health,
Год журнала:
2024,
Номер
6(8), С. e543 - e544
Опубликована: Июль 24, 2024
In
recent
years,
publications
in
The
Lancet
Digital
Health
have
presented
research
involving
pandemic
scenarios.1Nixon
K
Jindal
S
Parker
F
et
al.Real-time
COVID-19
forecasting:
challenges
and
opportunities
of
model
performance
translation.Lancet
Health.
2022;
4:
e699-e701Summary
Full
Text
PDF
PubMed
Scopus
(6)
Google
Scholar
However,
during
the
early
stages
pandemic,
terms
prediction,
scenario,
forecast
were
often
used
interchangeably
as
discussed
by
Kristen
Nixon
colleagues,1Nixon
leading
to
confusion.
Although
distinctions
between
these
concepts
been
refined
pandemic,2Howerton
E
Contamin
L
Mullany
LC
al.Evaluation
US
Scenario
Modeling
Hub
for
informing
response
under
uncertainty.Nat
Commun.
2023;
147260Crossref
(16)
we
find
that
clarification
is
needed
use
scenario
projections
narrative
devices.
We
also
encourage
more
discussion
regarding
terminology
describe
scenarios
how
they
are
evaluated.
What
distinction
why
important?
Underlying
all
mathematical
models
that,
together
with
numerical
values
their
parameters
(eg,
rate
transmission),
solved
numerically
generate
outputs
describing
future.
These
descriptions
future
states
a
system
usually
called
predictions.
term
has
denote
an
unconditional
prediction
about
what
will
happen
future.3Schroeder
SA
How
interpret
predictions:
reassessing
IHME's
model.Philosophy
Medicine.
2021;
2:
1-7Google
By
contrast,
projection
conditional
prediction—ie,
on
set
assumptions
(ie,
scenario).
Forecasts
typically
short
(less
than
month)
because
uncertainty
makes
them
functionally
useless
longer
time
scales,
whereas
medium
long
term.
separation
seen
important,
difference
not
clear-cut.4Winsberg
Harvard
Purposes
duties
scientific
modelling.J
Epidemiol
Community
76:
512-517Crossref
(13)
All
contain
some
idealisations
need
be
considered
when
assessing
validity
model.
One
purpose
forecasts
public
health
inform
expected
disease
incidence
support
allocation
health-care
resources.
Such
deemed
useful
many
countries
especially
decision
at
local
or
regional
levels.5Fox
SJ
Lachmann
M
Tec
surveillance
using
hospital
admissions
mobility
data.Proc
Natl
Acad
Sci
USA.
119e2111870119Crossref
(29)
projections,
other
hand,
multitude
purposes
pandemic—eg,
severity
outbreaks,
estimating
effects
different
vaccination
strategies
non-pharmaceutical
interventions,
outlining
worst-case
bed
demand.6Borchering
RK
Viboud
C
Howerton
al.Modeling
cases,
hospitalizations,
deaths,
rates
nonpharmaceutical
intervention
scenarios—United
States,
April–September
2021.MMWR
Morb
Mortal
Wkly
Rep.
70:
719Crossref
(106)
From
health-policy
perspective,
generated
from
serve
predominately
virtual
testbeds
exploring
chains
events
probable
occur
given
such
transmission
assumed
vaccine
roll-out.
A
taxonomy
design
was
recently
put
forward
based
analogy
experimental
design.7Runge
MC
Shea
al.Scenario
infectious
projections:
integrating
analysis
design.Epidemics.
2024;
47100775Crossref
(1)
considering
along
two
independent
axes—intervention
uncertainty—they
identify
six
classes
design,
sensitivity
analysis,
situational
awareness,
horizon
scanning.
served
another
important
purpose,
namely
For
example,
spring
2020,
governments
appealed
impose
social
distancing
flattening
curve
predictions
projections)
backdrops.
could,
instance,
contrast
absence
admission
distancing,
observing
fell
below
critical
capacity
threshold.
goes
beyond
illustrating
possible
world.
Rather,
implore
presumptive
audiences
act
specific
way.
this
related
Runge
colleagues'
concept
'decision
making'
does
cover
persuasive
aspect.7Runge
Of
note,
device
always
clear
reported
preprints
phases
could
interpreted
calls
heavier
restrictions
Gardner
colleagues),8Gardner
JM
Willem
Van
Der
Wijngaart
W
al.Intervention
against
estimated
impact
Swedish
healthcare
capacity.Int
J
Epidemiol.
2020;
49:
1443-1453PubMed
thus
had
transactional
aim
without
clearly
expressing
this.
aspects
described
performativity
interactive
effects,
which
refer
ability
effect
world.9Oomen
Hoffman
Hajer
MA
Techniques
futuring:
imagined
futures
become
socially
performative.Eur
Soc
Theory.
25:
252-270Crossref
(114)
Here,
modelers
(and
makers)
responsibility
infinite
pick
handful
simulated
communicated.
choices
can
control
communicated
profound
contingency
unfolds.
Given
wide
range
applications
modelling,
presentation
output
should
align
its
purpose.
Modelers
makers
explicit
underlying
and,
additionally
it
results
conditioned
intended
design.
Evaluating
usefulness
complex
compared
forecasts.
Reporting
forecasting
evaluated
comparing
actual
outcome
formal
metrics
mean
absolute
percentage
error
weighted
interval
score
probabilistic
forecasts).10Cramer
EY
Ray
EL
Lopez
VK
individual
ensemble
mortality
United
States.Proc
119e2113561119Crossref
(126)
Projections,
cannot
straight-forwardly
outcomes.
Formal
evaluations
performed
post-hoc
information
made
real-world
framework
applied
context
Hub.2Howerton
might
difficult
implement
there
no
guarantees
gathered
components
build
sufficient
represent
later
occurred
real
if
increase
post
hoc
obtain
accurate
timeframe.
Nonetheless,
found
failed
match
world
still
deployed
If
adherence
recommendations
improved
reduced),
then
fulfilled
although
factual
far
made.
same
true
analyses
application
precautionary
principles
policy
making.
rational
scenarios,
regards
measures
endpoints
adapted
When
presenting
projection,
therefore
utmost
importance
defined
unambiguous
terminology.
communicates
estimate
virulence
agent
short-term
influence
population
behaviour
reduce
spread
disease.
foundations
full,
run
undermine
general
trust
science
institutions.
argue
strict
reporting
evaluating
help
prevent
distrust
facilitate
communication.
declare
competing
interests.
Язык: Английский
Preface: COVID-19 Scenario Modeling Hubs
Epidemics,
Год журнала:
2024,
Номер
48, С. 100788 - 100788
Опубликована: Авг. 24, 2024
Язык: Английский
Design of a Simulation Model for the Diagnosis of Classical Swine Fever Virus in Ecuadorian Farms
WSEAS TRANSACTIONS ON BIOLOGY AND BIOMEDICINE,
Год журнала:
2024,
Номер
21, С. 345 - 355
Опубликована: Ноя. 5, 2024
Classical
swine
fever
(CSF)
is
a
disease
that
slows
down
animal
production
and
international
trade;
therefore,
its
identification
key
in
pig
farms
to
take
the
relevant
health
measures.
Therefore,
objective
of
this
research
was
design
Susceptible-Exposed-Infected-Recovered
(SEIR)
simulation
model
carry
out
epidemiological
modeling
for
outbreaks
classical
Sierra
Region
Ecuador,
using
Python
software
historical
data
on
incidences
provinces
Ecuadorian
highlands,
considering
variables
population,
initial
number
exposed
pigs,
infected,
pigs
removed,
contagion
rate
(α),
transmission
(β),
recovery
(γ).
The
results
show
SEIR
allowed
us
determine
population
susceptible
(healthy)
decreases
over
time
until
reaching
zero.
This
decrease
susceptibility
occurred
during
first
15
days,
which
shows
necessary
infect
entire
with
an
infected
person.
increases
days
total
infection
process
lasts
then
decreases.
It
also
identified
throughout
these
five
years
analysis
past,
it
has
been
increasing
from
2015
2019,
hurt
yields
productivity
mountains.
Язык: Английский