Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Overview of the input variables for both models predicting IHCA.

Six sets of eight-hourly measured Vital signs in the last 48 hours were used in both models. Two sets of most recent Laboratory results in the last seven days were used in Lab model. Demographic data remains static whereas vital signs and laboratory results, other data, and output are tracked at regular intervals. Abbreviations: Alb, albumin; APTT, activated partial thromboplastin time; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; BMI, body mass index; BNP, brain natriuretic peptide; CK, creatine kinase; CKMB, creatine kinase–muscle/brain; Cre, creatinine; CRP, C-reactive protein; eGFR, estimated estimated Glomerular. Filtration Rate; GGT, gamma-glutamyl transferase; Glu, glucose; Hb, haemoglobin; Hct, haematocrit; IHCA, in-hospital cardiac arrest; LD, lactate dehydrogenase; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular haemoglobin concentration; MCV, mean corpuscular volume; Plt, platelets; PT, prothrombin time; PT-INR, prothrombin time international normalized ratio; RBC, red blood cells; TP, total protein; T-Bil, total bilirubin; WBC, white blood cells.

More »

Fig 1 Expand

Fig 2.

Architectural overview of data extraction and representation.

More »

Fig 2 Expand

Table 1.

Characteristics of study population.

More »

Table 1 Expand

Table 2.

Predictive performance of each model for the in-hospital cardiac arrest in the next 0–8 hours.

More »

Table 2 Expand

Fig 3.

Calibration plot.

The x-axis summarizes the predicted probability of having a cardiac arrest in the next 8 hours, whereas the y-axis shows the observed proportions of patients who have a cardiac arrest in next 8 hours. Histogram below the calibration plot summarizes the distribution of predicted probability amongst all the patients in our dataset. Abbreviations: IHCA; inhospital cardiac arrest.

More »

Fig 3 Expand

Table 3.

Predictive performance of each model in sets of sensitivity analyses.

More »

Table 3 Expand

Fig 4.

Feature importance in predicting subsequent occurrence of IHCA.

Importance of each predictor in the 2 different random forest models: Vitals-Only model and Vitals+Labs model. 20 most important variables were summarized. Abbreviations: APTT, activated partial thromboplastin time; BUN, blood urea nitrogen; CRP, C-reactive protein; LD, lactate dehydrogenase; PT, prothrombin time.

More »

Fig 4 Expand