PhysioNet/CinC Challenge 2012: Predicting Mortality of ICU Patients
- Provider
- Ikaro Silva, George Moody, Roger Mark, Leo Anthony Celi, PhysioNet / Computing in Cardiology
- License
- ODC-By-1.0 (source)
- Stored at
- Amazon S3, s3://physionet-open/challenge-2012 (us-east-1)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Physiology
- Source
- https://physionet.org/content/challenge-2012/1.0.0/
About PhysioNet/CinC Challenge 2012: Predicting Mortality of ICU Patients
The 2012 PhysioNet/Computing in Cardiology Challenge asked for methods that predict in-hospital mortality from the first 48 hours of an ICU stay. The data are 12,000 ICU stays of adults admitted to cardiac, medical, surgical and trauma ICUs, with stays under 48 hours excluded. Up to 42 variables were recorded at least once in the first 48 hours: six general descriptors collected on admission (such as age, gender, height, ICU type and weight) and time series of vital signs and laboratory values with irregular timestamps. Training set A has 4,000 records with outcomes (in-hospital death, length of stay, survival, SAPS-I and SOFA scores); sets B and C were the test sets. Each record is a CSV text file of time, parameter and value. Measured 2026-11-13: set-a/ holds 4,000 per-patient .txt files, set-b/ is present, Outcomes-a.txt, Outcomes-b.txt and Outcomes-c.txt sit at the top level, along with set-a.tar.gz, the scoring code (physionet2012.m, physionet2012.c, physionet2012.R, saps_score) and the CinC papers. Load records with pandas and pivot to wide time series. A standard benchmark for ICU mortality prediction, irregularly sampled clinical time series and missing-data methods. Released under the Open Data Commons Attribution License v1.0; cite the challenge paper and the standard PhysioNet reference (Goldberger et al., 2000) when using the data.
How to access PhysioNet/CinC Challenge 2012: Predicting Mortality of ICU Patients in Python
import boto3
from botocore import UNSIGNED
from botocore.client import Config
s3 = boto3.client('s3', region_name='us-east-1', config=Config(signature_version=UNSIGNED))
s3.list_objects_v2(Bucket='physionet-open', Prefix='challenge-2012/', Delimiter='/')Lists the files with anonymous access. Runs on any machine, with no cloud account.
Data structure
Files
Example file: challenge-2012/1.0.0/papers/0245.pdf
Analyze this dataset without downloading it
On Scigantic, PhysioNet/CinC Challenge 2012: Predicting Mortality of ICU Patients opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Physiology datasets
- Regulation of Brain Cognitive States through Auditory, Gustatory, and Olfactory Stimulation with Wearable Monitoring
- Brno University of Technology ECG Quality Database (BUT QDB)
- Long-term Recordings of Gait Dynamics
- Haaglanden Medisch Centrum sleep staging database
- MICRO Motion capture data from groups of participants standing still to auditory stimuli (2015)
- GRABMyoFlow: Forearm and Wrist sEMG Extension of GRABMyo with Dynamic Gestures
- Respiratory and heart rate monitoring dataset from aeration study
- St. Vincent's University Hospital / University College Dublin Sleep Apnea Database