Figures¶

Every figure in the atlas, grouped by what it shows. Each links to the page it belongs to.

Drift and resets¶

Physical units, shared drift and time order Physical units, shared drift and time order Physical units, shared drift and time order Physical units, shared drift and time order Records out of order at SNUADC resets, and where the lost time goes Records out of order at SNUADC resets, and where the lost time goes SNUADC stamp arithmetic When each device's clock is corrected, in every recording

Data continuity¶

Gaps and unbroken segments in every waveform Gaps and unbroken segments in every waveform Gaps and unbroken segments in every waveform Gaps and unbroken segments in every waveform Message sizes and the steps between records Message sizes and the steps between records What each break in the stamps is Eight recordings store whole messages twice In some recordings two unrelated streams are interleaved in one track The Primus AWP stops for about 21 s every 7,230 s

Correction intervals¶

Clock corrections in every file Clock corrections in every file

Delay and window fits¶

How the monitors compute their numbers Numeric delays are fixed within a case, different between cases One device delay for the PLETH across the release SpO₂ and pulse rate latency from step events End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment Post-induction hypotension, from the charted start and from the devices Pulse arrival time in two correction steps

Distributions of measured parameters¶

Body mass index, derived from its inputs How the monitors compute their numbers How the monitors compute their numbers The PPG hardware delay and the adopted pulse arrival time The PPG hardware delay and the adopted pulse arrival time The PPG's 30 Hz frame clock The PPG's 30 Hz frame clock What the lab timestamp marks Intraoperative fluid balance as a distribution Intraoperative fluid balance as a distribution Oliguria labels for a prediction model Pulse arrival time in two correction steps Urine output in mL/kg/h as a distribution Urine output in mL/kg/h as a distribution

Agreement and residuals¶

Identifying the recording setup of every case Shared bursts between the ECG and the EEG The stamp sawtooth model for SNUADC and BIS The stamp sawtooth model for SNUADC and BIS Primus waveform gains are wrong in every recording, for CO₂ or for AWP A regression that carries rounding uncertainty through Cuff pressure against the arterial line End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment Minutes of hypotension per patient, from the monitor and from the waveform Pulse arrival time against blood pressure, before and after synchronisation Pulse arrival time against blood pressure, before and after synchronisation Pulse arrival time in two correction steps Pulse pressure variation gated by ventilator breaths

Heaping¶

Anaesthesia record times are rounded up How the anesthesia end time was charted How the anesthesia start time was charted How the operation end time was charted How the operation start time was charted Rounding and heaping of estimated blood loss Rounding and heaping of height Rounding and heaping of intraoperative crystalloid Rounding and heaping of intraoperative urine output Rounding and heaping of propofol bolus Rounding and heaping of weight

Distributions of recorded values¶

ASA Physical status classification Activated partial thromboplastin time Admission time from casestart Age Airway route Alanine transferase Albumin Ammonia Anesthesia end time Anesthesia start time Anesthesia type Arterial oxygen saturation Asparate transferase Base excess Bicarbonate Blood urea nitrogen Body mass index C-reactive protein Calcium chloride Case end Chloride Cormack's grade Creatinine Creatinine clearance Diagnosis Discharge time from casestart Double lumen tube size Emergency operation Endotracheal tube size Ephedrine Epinephrine Erythrocyte sedimentation rate Estimated blood loss Fentanyl Fibrinogen Glomerular filtration rate Glucose Height Hematocrit Hemoglobin ICU days In-hospital Mortality Intraoperative FFP transfusion Intraoperative RBC transfusion Intraoperative colloid Intraoperative crystalloid Intraoperative urine output Ionized Calcium LMA size Lactate Midazolam Operation end time Operation name Operation start time Partial pressure of CO₂ Partial pressure of O₂ Phenylephrine Platelet count Potassium Preoperative ECG Preoperative GOT Preoperative GPT Preoperative HCO3- Preoperative K Preoperative Na Preoperative PT Preoperative PaCO₂ Preoperative PaO₂ Preoperative SpO₂ Preoperative aPTT Preoperative albumin Preoperative base excess Preoperative blood urea nitrogen Preoperative creatinine Preoperative diabetes Preoperative glucose Preoperative hemoglobin Preoperative hypertension Preoperative pH Preoperative platelet count Preoperative pulmonary function Probably phosphorus (unconfirmed): not in lab_parameters.csv Propofol bolus Prothrombin time (%) Prothrombin time (INR) Prothrombin time (sec) Recording Start Time Rocuronium Sex Site of IV line (1) Site of IV line (2) Site of arterial line (1) Site of arterial line (2) Site of central line (1) Site of central line (2) Sodium Surgery type Surgical approach Surgical department Surgical position Total bilirubin Total protein Weight White blood cell count pH CardioQ FLOW uses three scales (a segmented code) SNUADC values are 10-bit codes, though the device is documented as 12-bit

Example traces¶

Arterial line accesses in the pressure waveform Shared bursts between the ECG and the EEG The PPG sawtooth and its correction The PPG sawtooth and its correction The Primus ventilator against SNUADC, by breaths The Primus ventilator against SNUADC, by breaths The stamp sawtooth model for SNUADC and BIS What the lab timestamp marks Where the PPG sawtooth stops End-tidal to arterial CO₂, and pulse oximetry against blood gas, at the right moment Pulse arrival time in two correction steps

Schematics¶

How the monitors compute their numbers

Other¶

The stamp sawtooth model for SNUADC and BIS Reading the stamps

Population results¶

Closing loops in the timing graph The Primus stamp step at each airway-pressure pause The SNUADC clock through the Solar8000 numerics, whole release