Table 4 Instability prediction performance of APRICOT-M compared to acuity baseline for episode and step level

From: Real-time prediction of intensive care unit patient acuity and therapy requirements using state-space modelling

Cohort

Model

Prevalence (episode)

AUROC (episode)

AUPRC (episode)

Prevalence (step)

AUROC (step)

AUPRC (step)

Development

SOFA

0.116

0.51 (0.50–0.52)

0.13 (0.13–0.14)

0.012

0.60 (0.60–0.61)

0.02 (0.02–0.02)

SOFA (≥2 points)

0.49 (0.48–0.50)

0.28 (0.27–0.29)

0.54 (0.54–0.55)

0.17 (0.16–0.17)

APRICOT-M

0.73 (0.72–0.74)**

0.32 (0.31–0.34)**

0.74 (0.73–0.75)**

0.10 (0.09–0.11)**

External

SOFA

0.131

0.46 (0.46–0.47)

0.13 (0.13–0.13)

0.013

0.55 (0.55–0.56)

0.02 (0.02–0.02)

SOFA (≥2 points)

0.46 (0.46–0.46)

0.23 (0.22–0.23)

0.51 (0.51–0.51)

0.13 (0.12–0.13)

APRICOT-M

0.74 (0.74–0.75)**

0.47 (0.47–0.48)**

0.75 (0.74–0.75)**

0.24 (0.23–0.25)**

Prospective

SOFA

0.242

0.61 (0.54–0.68)

0.39 (0.29–0.49)

0.011

0.68 (0.64–0.71)

0.02 (0.02–0.03)

SOFA (≥2 points)

0.52 (0.46–0.58)

0.40 (0.33–0.48)

0.53 (0.50–0.57)

0.12 (0.09–0.16)

APRICOT-M

0.61 (0.53–0.68)

0.41 (0.27–0.48)

0.69 (0.64–0.74)*

0.07 (0.04–0.10)**

  1. AUPRC area under the precision-recall curve, AUROC area under the receiving operating characteristic, SOFA sequential organ failure assessment. Performance is the median AUROC and AUPRC across a 100-iteration bootstrap with replacement, with 95% confidence intervals in parenthesis. P values are based on pairwise two-sided Wilcoxon rank sum tests.
  2. Bold values represent performance metrics that are significantly higher based on statistical analysis.
  3. *p value < 0.05 compared to one of SOFA or SOFA (≥2 points).
  4. **p value < 0.05 compared to SOFA and SOFA (≥2 points).
Back to article page