dataset_info:
- config_name: cohort_full
features:
- name: subject_id
dtype: int64
- name: ed_stay_id
dtype: int64
- name: hadm_id
dtype: int64
- name: ed_intime
dtype: timestamp[us]
- name: ed_outtime
dtype: timestamp[us]
- name: disposition
dtype: large_string
- name: race
dtype: large_string
- name: arrival_transport
dtype: large_string
- name: first_careunit
dtype: large_string
- name: first_icu_intime
dtype: timestamp[us]
- name: cohort_label
dtype: large_string
- name: gender
dtype: large_string
- name: anchor_age
dtype: int64
- name: anchor_year
dtype: int64
- name: admittime
dtype: timestamp[us]
- name: dischtime
dtype: timestamp[us]
- name: admission_type
dtype: large_string
- name: discharge_location
dtype: large_string
- name: insurance
dtype: large_string
- name: language
dtype: large_string
- name: marital_status
dtype: large_string
- name: ed_stay_id_2
dtype: float64
- name: stay_window_start
dtype: timestamp[us]
- name: stay_window_end
dtype: timestamp[us]
- name: ed_boarding_time_min
dtype: float64
- name: time_steps
dtype: int64
splits:
- name: cohort_base
num_bytes: 108446901
num_examples: 397670
download_size: 25727336
dataset_size: 108446901
- config_name: dispensed_meds
features:
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: hadm_id
dtype: int64
- name: charttime
dtype: timestamp[us]
- name: medication
dtype: large_string
- name: event_txt
dtype: large_string
- name: in_er
dtype: bool
- name: drug_class
dtype: large_string
- name: action_idx
dtype: int64
- name: minutes_into_stay
dtype: float64
- name: time_step
dtype: int64
splits:
- name: er_and_ward
num_bytes: 311238084
num_examples: 2675531
download_size: 33872234
dataset_size: 311238084
- config_name: ecg_results
features:
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: hadm_id
dtype: int64
- name: ecg_time
dtype: timestamp[us]
- name: ecg_acuity
dtype: int64
splits:
- name: full_ecg
num_bytes: 6226839
num_examples: 155186
download_size: 4053167
dataset_size: 6226839
- config_name: empty_patient_state
features:
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: hadm_id
dtype: float64
- name: time_steps
dtype: int64
splits:
- name: empty_patient_state
num_bytes: 1454310797
num_examples: 45270375
download_size: 10265748
dataset_size: 1454310797
- config_name: height
features:
- name: subject_id
dtype: int64
- name: chartdate
dtype: date32
- name: result_name
dtype: large_string
- name: result_value
dtype: float64
splits:
- name: height_full
num_bytes: 22419409
num_examples: 656979
download_size: 3891379
dataset_size: 22419409
- config_name: lab_results
features:
- name: ed_stay_id
dtype: int64
- name: category
dtype: large_string
- name: fluid
dtype: large_string
- name: order_time
dtype: timestamp[us]
- name: subject_id
dtype: int64
- name: hadm_id
dtype: int64
- name: result_time
dtype: timestamp[us]
- name: ordered_location
dtype: large_string
- name: abnormal
dtype: bool
- name: action
dtype: large_string
splits:
- name: labs_full
num_bytes: 269113124
num_examples: 2542661
download_size: 43257327
dataset_size: 269113124
- config_name: medrecon
features:
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: recon_ace_inhibitor
dtype: uint8
- name: recon_analgesic___nsaid
dtype: uint8
- name: recon_analgesic___opioid_nsaid
dtype: uint8
- name: recon_antiarrhythmic
dtype: uint8
- name: recon_antibiotic
dtype: uint8
- name: recon_anticoagulant
dtype: uint8
- name: recon_anticonvulsant
dtype: uint8
- name: recon_antiemetic
dtype: uint8
- name: recon_antiplatelet
dtype: uint8
- name: recon_antipsychotic
dtype: uint8
- name: recon_benzodiazepine___sedative_anxiolytic
dtype: uint8
- name: recon_beta_blocker
dtype: uint8
- name: recon_bronchodilator
dtype: uint8
- name: recon_calcium_channel_blocker
dtype: uint8
- name: recon_corticosteroid
dtype: uint8
- name: recon_diuretic
dtype: uint8
- name: recon_gi___acid_suppression
dtype: uint8
- name: recon_insulin_glucose
dtype: uint8
- name: recon_nitrate
dtype: uint8
- name: recon_n_total_meds
dtype: int64
- name: recon_n_drug_classes
dtype: int64
splits:
- name: medrecon_full
num_bytes: 15100539
num_examples: 296089
download_size: 4092329
dataset_size: 15100539
- config_name: microbiology
features:
- name: subject_id
dtype: int64
- name: ed_stay_id
dtype: int64
- name: hadm_id
dtype: int64
- name: charttime
dtype: timestamp[us]
- name: storetime
dtype: timestamp[us]
- name: action_space
dtype: large_string
- name: culture_result
dtype: large_string
splits:
- name: microbiology_full
num_bytes: 13591066
num_examples: 185552
download_size: 5278498
dataset_size: 13591066
- config_name: rad_results
features:
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: hadm_id
dtype: int64
- name: exam_time
dtype: timestamp[us]
- name: rad_acuity_level
dtype: int64
splits:
- name: rad_full
num_bytes: 10801289
num_examples: 269191
download_size: 6246623
dataset_size: 10801289
- config_name: triage_vitals
features:
- name: charttime
dtype: timestamp[us]
- name: ed_stay_id
dtype: int64
- name: subject_id
dtype: int64
- name: current_temperature
dtype: float64
- name: current_heartrate
dtype: float64
- name: current_resprate
dtype: float64
- name: current_o2sat
dtype: float64
- name: current_sbp
dtype: float64
- name: current_dbp
dtype: float64
- name: current_pain
dtype: float64
- name: acuity
dtype: float64
- name: chiefcomplaint
dtype: large_string
- name: source
dtype: large_string
- name: time_since_last_min
dtype: float64
- name: pain_missing
dtype: int64
- name: heartrate_missing
dtype: int64
- name: o2sat_missing
dtype: int64
- name: acuity_missing
dtype: int64
- name: chiefcomplaint_missing
dtype: int64
- name: resprate_missing
dtype: int64
- name: sbp_missing
dtype: int64
- name: dbp_missing
dtype: int64
- name: temperature_missing
dtype: int64
- name: temperature_delta
dtype: float64
- name: heartrate_delta
dtype: float64
- name: resprate_delta
dtype: float64
- name: o2sat_delta
dtype: float64
- name: sbp_delta
dtype: float64
- name: dbp_delta
dtype: float64
- name: pain_delta
dtype: float64
- name: temperature_rate_per_min
dtype: float64
- name: heartrate_rate_per_min
dtype: float64
- name: resprate_rate_per_min
dtype: float64
- name: o2sat_rate_per_min
dtype: float64
- name: sbp_rate_per_min
dtype: float64
- name: dbp_rate_per_min
dtype: float64
- name: current_mean_arterial_pressure
dtype: float64
- name: temperature_rolling1h
dtype: float64
- name: heartrate_rolling1h
dtype: float64
- name: resprate_rolling1h
dtype: float64
- name: o2sat_rolling1h
dtype: float64
- name: sbp_rolling1h
dtype: float64
- name: dbp_rolling1h
dtype: float64
splits:
- name: triage_vitals_full
num_bytes: 661714092
num_examples: 1811202
download_size: 88016187
dataset_size: 661714092
- config_name: weight
features:
- name: subject_id
dtype: int64
- name: chartdate
dtype: date32
- name: result_name
dtype: large_string
- name: result_value
dtype: float64
splits:
- name: weight_full
num_bytes: 56211758
num_examples: 1647231
download_size: 9166863
dataset_size: 56211758
configs:
- config_name: cohort_full
data_files:
- split: cohort_base
path: cohort/cohort_base-*
- config_name: dispensed_meds
data_files:
- split: er_and_ward
path: dispensed_meds/er_and_ward-*
- config_name: ecg_results
data_files:
- split: full_ecg
path: ecg/full_ecg-*
- config_name: empty_patient_state
data_files:
- split: empty_patient_state
path: patient_state/empty_patient_state-*
- config_name: height
data_files:
- split: height_full
path: height_data/height_full-*
- config_name: lab_results
data_files:
- split: labs_full
path: labs/labs_full-*
- config_name: medrecon
data_files:
- split: medrecon_full
path: medrecon_data/medrecon_full-*
- config_name: microbiology
data_files:
- split: microbiology_full
path: microbiology_data/microbiology_full-*
- config_name: rad_results
data_files:
- split: rad_full
path: rad_data/rad_full-*
- config_name: triage_vitals
data_files:
- split: triage_vitals_full
path: triage_vitals_data/triage_vitals_full-*
- config_name: weight
data_files:
- split: weight_full
path: weight_data/weight_full-*
Cohort Data
Processed ED patient cohort derived from MIMIC-IV. One row per ED visit (ed_stay_id). Built from cohort_base (BigQuery) with the following post-processing steps:
- AGAINST ADVICE discharge_location records removed — patient-driven departures are out of scope.
- Consecutive ED visits sharing a hadm_id collapsed into single rows; second ed_stay_id stored in
ed_stay_id_2. stay_window_start/stay_window_endcolumns added covering ED arrival to final discharge.ed_boarding_time_minadded: minutes in ED after admission decision (null for ED-only patients).
Intended as the primary cohort table for feature engineering and RL state construction.
Dispensed Meds
Dispensed medication records for all cohort patients across two phases and two populations.
ED phase (in_er=True): Pyxis dispenses for all patients (admitted and ED-only) while physically in the ED, derived from ed.pyxis.
Ward phase (in_er=False): eMAR administration records for admitted patients from ED departure up to ICU transfer, derived from hosp.emar.
event_txt filtered to administration and rate-change events only.
Medication names mapped to a 22-class drug vocabulary via regex. Each record includes minutes_into_stay (time since ED arrival) and time_step bucket (floor(minutes / time_block)).
ECG Data
ECG acuity labels derived from mimiciv_ecg.machine_measurements for cohort patients. Each of the 18 report columns is classified independently
via regex with priority order: abnormal (2) > neutral (1) > normal (0) > empty (-1).
Row-level max across all 18 columns gives ecg_acuity: 0=normal, 1=neutral/unknown, 2=abnormal.
Patients with all 18 columns empty are assigned 1 (missing ECG ≠ normal ECG).
Multiple ECGs per ED stay resolved to one row: highest acuity kept, ties broken by earliest ecg_time.
Window covers ED arrival through hospital ward stay (capped at ICU transfer if applicable), with ±1 hour buffer for ECG machine clock drift.
Labs Data
Laboratory results from hosp.labevents for cohort patients during their stay window.
Grouped by category x fluid (19 unique combinations) rather than individual test label to reduce action space.
Each result includes the abnormal flag for worst-case aggregation. Intended as the primary lab state feature source.
Medrecon Data
Medication reconciliation records from ed.medrecon — medications the patient was taking prior to ED arrival.
One row per medication per ED visit. Represents pre-arrival medication state, not actions taken during the visit.
Microbiology Data
Microbiology culture events from hosp.microbiologyevents for cohort patients during their stay window.
Includes culture order time, specimen type, organism name, and antibiotic sensitivity results.
culture_ordered is the real-time state signal; culture_positive is a retrospective label only (~2% of results available before ED discharge).
OMR Data
Height and weight measurement records from hosp.omr for cohort patients — separated to take single measurements of height as an average of all available height measurements for a patient and retaining all separate weight measurements.
Duplicates were dropped. Used to supplement ED and inpatient state features with recent baseline measurements.
Radiology Data
Radiology report text from mimiciv_note.radiology for cohort patients. Covers all imaging modalities (CXR, CT, MRI, ultrasound, etc.).
Primary reports only (note_type=RR). Window covers ED arrival through hospital ward stay (capped at ICU transfer if applicable).hadm_id is NULL for ED-only patients and populated for admitted patients. exam_name and cpt_code included from radiology_detail to identify
imaging modality.
Triage_Vitals
Combined triage and vital signs time series — one row per (ed_stay_id, charttime) timestep.
Triage records (ed.triage) provide the baseline reading at ed_intime; subsequent vitals (ed.vitalsign) provide follow-up readings.
stay_id remapping applied so that vitals from a patient's second consecutive ED visit are attributed to the canonical (first) stay_id.
Rows recorded before ed_intime are dropped; records from vitals with charttime == ed_intime are dropped as duplicates of the triage records.
Preprocessing: pain column normalized and coerced to numeric, physiologically implausible vital values nulled. Missing indicators created before forward-fill.
Vitals forward-filled within stay, then mean-imputed for any remaining NaN.
Feature engineering: time_since_last_min, per-column deltas and rates of change (units/min), 1-hour rolling averages, mean arterial pressure (dbp + (sbp - dbp) / 3).
Triage rows get 0 for all delta/rate columns. Final vital columns renamed to current_{col}.