Table 2. Studies of data-analytic models for prediction of outcomes.
| Study | Patient population | Technology used | Study design | Endpoint predicted | Validity |
|---|---|---|---|---|---|
| Baronov et al. (32) | Pediatric intensive care unit | Model based risk assessment | Retrospective cohort | Probability of inadequate oxygen delivery | AUC for SVO2 <40, 0.79 (95% CI: 0.76–0.82) |
| Moorman et al. (56) | Neonatal intensive care unit | Heart rate characteristic index | Randomized controlled trial | Fold-increase in risk of sepsis | HR 0.78 (95% CI: 0.61–0.99) for neonates monitored with HRC |
| Rusin et al. (52) | Pediatric cardiac intensive care unit | Classification algorithm | Prospective observational | Decompensation (CPR or intubation) | AUC 0.91 (95% CI: 0.88–0.94) |
| Hu et al. (58) | Adult patients with hematologic malignancies | Advanced neural network | Retrospective cohort | ICU transfer and cardiac arrest | Sensitivity: 0.93; specificity: 0.97; PPV: 0.78; NPV: 0.99 |
| Chen et al. (57) | Intensive care unit patients | Data-analytic decision support system | Retrospective cohort | Risk of mortality | AUC: 0.88; sensitivity: 0.83; specificity 0.92; PPV: 0.62; NPV: 0.97 |
SVO2, central venous saturation; AUC, area under the receiver operating curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; CPR, cardiopulmonary resuscitation.