Twelve ratios and three scores recomputed on every change from values already recorded. This makes no forecast — it applies published thresholds to numbers you supply and reports which have left their band, in which direction, and by how much. Departure counts both ways: a shock index of 0.24 in bradycardia is as abnormal as 1.4 in shock. Every threshold below carries its source and the populations in which it is unreliable. A separate biomarker panel informs without changing any score.
Load a picture, then adjust from there.
In a real deployment these would arrive from the monitor. Here you type them.
Changes the SOFA respiratory and cardiovascular components, which cannot score above 2 without them.
No equipment needed. ISCCM names both capillary refill and urine output.
In practice these arrive with delay — see turnaround below.
Feed the biomarker panel only. Changing them moves no score, ratio or alert tier.
Recomputed on every change. Identical at 4 AM and 4 PM — which is the whole argument for computing them continuously.
Only what has left its normal range. Each is arithmetic on values already recorded — the kind nobody computes continuously on every patient.
Everything within range.
Protein C, CRP, procalcitonin and lactate. Shown to inform the clinician, not to define sepsis: Sepsis-3 defines it by organ dysfunction, and the Surviving Sepsis Campaign 2026 states that no single biomarker rules sepsis in or out. Lactate enters the definition only as the septic-shock criterion.
Informs · does not change qSOFA, SOFA or the alert tier
How long until a result exists, and how long before the number on screen should no longer be acted on. Times run from specimen receipt; add 20–40 minutes for collection, transport and reporting.
| Test | Expected | Stale after | Note |
|---|