Transcript Document

Apoio à decisão em
medicina intensiva usando
ECBD
Pedro Gago – I P Leiria
Intensive care
• About 250 variables
are needed to
describe an ICU
patient
• Humans are unable
to cope with more
than seven variables
at a time
2
Objectives
• Assist ICU doctors by providing accurate
and timely predictions for:
– the final outcome
– organ dysfunction or failure
• Must overcome natural physician
resistance
3
Intensive Care Medicine
• Condition is severe to the point where it
is very difficult for doctors to assess the
patient’s state
• Objective is to stabilize in order to allow
transfer to other units
• Highly invasive and very costly
4
Intensive Care Medicine
• Data from bed-side
monitors may
contain useful
information
• Presently such data
is not stored
5
Practical Issues
• Some variables values must be
collected manually
– Urine output
• Data Quality
– Errors caused by human intervention
– Sensor malfunctions
6
Scores in use
• SAPS – indicative of the patient’s
condition severity
• The worst values the first 24 hours of
stay in the ICU are collected and the
score is calculated
7
Scores in use (2)
• SOFA – measures organ
dysfuntion/failure (worst daily values)
– Cardiovascular, hepatic, central nervous
system, respiratory, renal, coagulation
• Worst daily values
• Indicative of patient’s condition evolution
8
INTCare
• Decision Support System to assist ICU
doctors
• Uses available data in order to predict
outcome and organ dysfunction/failure
• Not intended to replace doctors
9
INTCare (2)
• Semi-autonomous – updates its models
as new data arrives
• Performance expected to improve with
time
• Better results through the use of real time
data
10
KNOWLEDGE MANAGEMENT
Architecture
Performance
Data
Performance
Agent
Model
Initialization
Agent
Data
Warehouse
Data Mining
Agent
Knowledge
Base
(PMML)
Bedside monitors
Heart Rate
O2 Sat.
Blood Pr.
Data
acquisition
DATA ENTRY
Preprocessing
Agent
INFERENCE
Monitored
Data
Physiological
data
Clinical
Data
Clinical
Data Entry
Agent
Data
Retrieval
Agent
SAPS
SOFA
Admission
Discharge
Prediction
Agent
Interface
Agent
INTERFACE
Scenario
Evaluation
Agent
Scenarios
11
KNOWLEDGE MANAGEMENT
Architecture (2)
Performance
Data
Model
Initialization
Agent
Performance
Agent
Bedside monitors
Ensemble
Agent
Heart Rate
O2 Sat.
Blood Pr.
Data
acquisition
DATA ENTRY
Data
Warehouse
Data Mining
Agent
Knowledge
Base
(PMML)
Preprocessing
Agent
INFERENCE
Monitored
Data
Physiological
data
Clinical
Data
Clinical
Data Entry
Agent
Data
Retrieval
Agent
SAPS
SOFA
Admission
Discharge
Prediction
Agent
Interface
Agent
INTERFACE
Scenario
Evaluation
Agent
Scenarios
12
EURICOS II
• Data from 42 UCI from 9 countries
• 10 months (1998 and 1999)
13
EURICOS II (2)
• Data available includes:
– case mix (age, origin, etc)
– SAPS score
– daily SOFA scores
– intermediate outcomes
– final outcome
14
INTERMEDIATE OUTCOMES
Critical Event
Suggested Continuously
out-of-range
Range
Intermitently
out-of-range
Event
anytime
BP(mmHg)
90 – 180
≥ 60 mins
≥ 60 in 120
mins
< 60
SaO2(%)
≥ 90
≥ 60 mins
≥ 60 in 120
mins
< 80
HR(bpm)
60 – 120
≥ 60 mins
≥ 60 in 120
mins
< 30 OR
> 180
Diur(ml/hour)
≥ 30
≥ 2 hours
15
Ensemble
• Training
– Each model is trained on different subsets of the dataset
– Each variable has a 70% chance of being selected
– Starts with equal weights
16
Ensemble
• Evolution
– Results from batches of records
– Weight adjustments according to individual model
performance
– Worst performing models are deleted from the ensemble
– New models are trained using the most recent data and
included in the ensemble
17
Ensemble
• Preliminary results (evolution doesn’t include new
models)
– Ensemble trained with all cases still outperforms ensemble
trained with less cases followed by weight adjustment
– Both outperform best individual model
18
Future Work
• Greater volume of data – deployment in other ICUs
• Reduce prediction window (next 6 hours instead of
next day)
• Suggest course of action (must be delayed until
physician resistance is lowered)
19
INTCare
Thank you.
Questions?
[email protected]
20