Machine learning for predicting severe dengue in Puerto Rico
Infectious Diseases of Poverty,
Год журнала:
2025,
Номер
14(1)
Опубликована: Фев. 4, 2025
Distinguishing
between
non-severe
and
severe
dengue
is
crucial
for
timely
intervention
reducing
morbidity
mortality.
World
Health
Organization
(WHO)-recommended
warning
signs
offer
a
practical
approach
clinicians
but
have
limited
sensitivity
specificity.
This
study
aims
to
evaluate
machine
learning
(ML)
model
performance
compared
WHO-recommended
in
predicting
among
laboratory-confirmed
cases
Puerto
Rico.
We
analyzed
data
from
Rico's
Sentinel
Enhanced
Dengue
Surveillance
System
(May
2012-August
2024),
using
40
clinical,
demographic,
laboratory
variables.
Nine
ML
models,
including
Decision
Trees,
K-Nearest
Neighbors,
Naïve
Bayes,
Support
Vector
Machines,
Artificial
Neural
Networks,
AdaBoost,
CatBoost,
LightGBM,
XGBoost,
were
trained
fivefold
cross-validation
evaluated
with
area
under
the
receiver
operating
characteristic
curve
(AUC-ROC),
sensitivity,
A
subanalysis
excluded
hemoconcentration
leukopenia
assess
resource-limited
settings.
An
AUC-ROC
value
of
0.5
indicates
no
discriminative
power,
while
values
closer
1.0
reflect
better
performance.
Among
1708
cases,
24.3%
classified
as
severe.
Gradient
boosting
algorithms
achieved
highest
predictive
performance,
an
97.1%
(95%
CI:
96.0-98.3%)
CatBoost
full
40-variable
feature
set.
Feature
importance
analysis
identified
(≥
20%
increase
during
illness
or
≥
above
baseline
age
sex),
(white
blood
cell
count
<
4000/mm3),
timing
presentation
at
4-6
days
post-symptom
onset
key
predictors.
When
excluding
leukopenia,
was
96.7%
95.5-98.0%),
demonstrating
minimal
reduction
Individual
like
abdominal
pain
restlessness
had
sensitivities
79.0%
64.6%,
lower
specificities
48.4%
59.1%,
respectively.
Combining
3
improved
specificity
(80.9%)
maintaining
moderate
(78.6%),
resulting
74.0%.
especially
gradient
algorithms,
outperformed
traditional
dengue.
Integrating
these
models
into
clinical
decision-support
tools
could
help
identify
high-risk
patients,
guiding
interventions
hospitalization,
monitoring,
administration
intravenous
fluids.
The
confirmed
models'
applicability
settings,
where
access
may
be
limited.
Язык: Английский
Associations of resuscitation fluid load, colloid-to-crystalloid infusion ratio and clinical outcomes in children with dengue shock syndrome
PLoS neglected tropical diseases,
Год журнала:
2025,
Номер
19(1), С. e0012786 - e0012786
Опубликована: Янв. 10, 2025
Background
Severe
respiratory
distress
and
acute
kidney
injury
(AKI)
are
key
factors
leading
to
poor
outcomes
in
patients
with
dengue
shock
syndrome
(DSS).
There
is
still
limited
data
on
how
much
resuscitated
fluid
the
specific
ratios
of
intravenous
types
contribute
development
severe
necessitating
mechanical
ventilation
(MV)
AKI
children
DSS.
Methodology/principal
findings
This
retrospective
study
was
conducted
at
a
tertiary
pediatric
hospital
Vietnam
between
2013
2022.
The
primary
were
need
for
MV
renal
function
within
48
h
post-admission.
A
predictive
model
developed
based
covariates
from
first
24
PICU
admission.
Changes
analyzed
using
linear
mixed-effects
model.
total
1,278
DSS
complete
clinical
included.
performance
volume
administered
yielded
an
AUC
0.871
(95%
CI,
0.836–0.905),
while
colloid-to-crystalloid
ratio
showed
0.781
0.743–0.819)
(both
P
<
0.001).
optimal
cut-off
point
cumulative
infusion
181
mL/kg,
whereas
that
1.6.
Multivariable
analysis
identified
female
patients,
bleeding,
transaminitis,
excessive
resuscitation,
higher
proportion
colloid
solutions
as
significant
predictors
patients.
demonstrated
high
accuracy,
C-statistic
89%,
strong
calibration,
low
Brier
score
(0.04).
Importantly,
more
pronounced
decline
glomerular
filtration
rate
observed
who
required
than
those
did
not.
Conclusions/significance
provides
insights
into
optimizing
management
protocols,
highlighting
importance
monitoring
during
early
resuscitation
improve
Язык: Английский
Prediction of depressive disorder using machine learning approaches: findings from the NHANES
BMC Medical Informatics and Decision Making,
Год журнала:
2025,
Номер
25(1)
Опубликована: Фев. 17, 2025
Depressive
disorder,
particularly
major
depressive
disorder
(MDD),
significantly
impact
individuals
and
society.
Traditional
analysis
methods
often
suffer
from
subjectivity
may
not
capture
complex,
non-linear
relationships
between
risk
factors.
Machine
learning
(ML)
offers
a
data-driven
approach
to
predict
diagnose
depression
more
accurately
by
analyzing
large
complex
datasets.
This
study
utilized
data
the
National
Health
Nutrition
Examination
Survey
(NHANES)
2013–2014
using
six
supervised
ML
models:
Logistic
Regression,
Random
Forest,
Naive
Bayes,
Support
Vector
(SVM),
Extreme
Gradient
Boost
(XGBoost),
Light
Boosting
(LightGBM).
Depression
was
assessed
Patient
Questionnaire
(PHQ-9),
with
score
of
10
or
higher
indicating
moderate
severe
depression.
The
dataset
split
into
training
testing
sets
(80%
20%,
respectively),
model
performance
evaluated
accuracy,
sensitivity,
specificity,
precision,
AUC,
F1
score.
SHAP
(SHapley
Additive
exPlanations)
values
were
used
identify
critical
factors
interpret
contributions
each
feature
prediction.
XGBoost
identified
as
best-performing
model,
achieving
highest
highlighted
most
significant
predictors
depression:
ratio
family
income
poverty
(PIR),
sex,
hypertension,
serum
cotinine
hydroxycotine,
BMI,
education
level,
glucose
levels,
age,
marital
status,
renal
function
(eGFR).
We
developed
models
for
interpretation.
identifies
key
associated
depression,
encompassing
socioeconomic,
demographic,
health-related
aspects.
Язык: Английский
In-Hospital Mortality in Mechanically Ventilated Children With Severe Dengue Fever: Explanatory Factors in a Single-Center Retrospective Cohort From Vietnam, 2013–2022
Pediatric Critical Care Medicine,
Год журнала:
2025,
Номер
unknown
Опубликована: Март 19, 2025
Objectives:
Severe
dengue
fever
complicated
by
critical
respiratory
failure
requiring
mechanical
ventilation
(MV)
contributes
to
high
mortality
rates
among
PICU-admitted
patients.
This
study
aimed
identify
key
explanatory
variables
of
fatality
in
mechanically
ventilated
children
with
severe
dengue.
Design:
Retrospective
cohort,
from
2013
2022.
Setting:
PICU
the
tertiary
Children’s
Hospital
No.
2,
Vietnam.
Patients:
Two
hundred
who
received
MV.
Interventions:
None.
Measurements
and
Main
Results:
We
analyzed
clinical
laboratory
data
during
stay.
The
primary
outcome
was
28-day
in-hospital
mortality.
Backward
stepwise
multivariable
logistic
regression
performed
associated
dengue-associated
at
initiation
model
performance
assessed
using
C-statistics,
calibration
plot,
Brier
score.
patients
had
a
median
age
7
years
(interquartile
range,
4–9).
Overall,
47
(24%)
200
died
hospital.
Five
factors
were
greater
odds
mortality:
transaminitis
(aspartate
aminotransferase
or
alanine
≥
1000
IU/L),
blood
lactate
levels,
vasoactive-inotropic
score
(>
30),
encephalitis,
peak
inspiratory
pressure
on
training
(test)
sets
C-statistic
0.86
(0.84),
good
slope
1.0
(0.89),
0.08.
Conclusions:
In
our
center,
2022,
MV-experienced
rate.
main
death
(related
liver
injury,
shock,
MV)
may
inform
future
practice
such
critically
ill
Язык: Английский