How this instrument works
Sequential Organ Failure Assessment turns exam findings and lab draws from six organ systems into a single running number. Respiration is graded from the PaO₂/FiO₂ ratio, coagulation from the platelet count, the liver from bilirubin, the kidneys from creatinine, and the central nervous system from the Glasgow Coma Scale — each on a 0-to-4 band, worst function scoring highest. The cardiovascular piece works differently: instead of a lab value, a clinician picks one of five descriptions ranging from a mean arterial pressure of 70 mmHg or better with no drugs on board, up to high-dose dopamine or a meaningful dose of epinephrine or norepinephrine. Add the six parts and the total runs from 0, no measurable dysfunction anywhere, to 24, severe failure across every system.
Jean-Louis Vincent and the European Society of Intensive Care Medicine's Working Group on Sepsis-Related Problems published the tool in 1996, built to describe — not predict — how sick a patient's organs were on a given day. That distinction matters in practice: the assessment is meant to be recalculated repeatedly, often once daily through an ICU stay, so a clinician can watch the trend rather than lean on a single reading. A number that climbs over 48 hours flags worsening dysfunction even if the absolute figure still looks moderate, while a falling trajectory is one of the more reassuring signals available at the bedside. The Sepsis-3 consensus definitions, published two decades later, built directly on this behavior, defining sepsis in part as an acute rise of two points or more in a patient with suspected infection.
Two of the six bands here are honest simplifications of Vincent's original table, made so the form stays usable without a chart of drug infusion rates in front of you. The cardiovascular category is a pre-built five-option list rather than a field for the exact microgram-per-kilogram dose, and the respiratory band does not separately branch on whether a patient is receiving ventilatory support at its two worst tiers the way the full published criteria do. Neither change alters the arithmetic within the bands actually offered; both simply trade a sliver of granularity for something fillable in under a minute. This instrument is also a different tool from qSOFA, also on this site: qSOFA is a fast, three-item bedside screen with no labs at all, meant to flag which patients need a closer look; the full assessment built here is the detailed, lab-and-vasopressor-dependent version used afterward, once someone is already being monitored closely, to characterize and track how organ dysfunction is actually progressing.
- Enter the PaO₂/FiO₂ ratio in mmHg — arterial oxygen tension divided by the fraction of inspired oxygen.
- Enter Platelets in thousands per microliter and Bilirubin in mg/dL from the same lab panel.
- Select the Cardiovascular category matching the patient's current pressor support, from no vasopressors up to high-dose therapy.
- Enter the Glasgow Coma Scale total (3-15) and Creatinine in mg/dL.
- Read the five sub-scores plus the SOFA total (0-24); each updates independently as its inputs change.
Worked example — six sub-scores of 2, totaling 12
A patient with a PaO₂/FiO₂ of 250, platelets of 80, bilirubin of 3.0 mg/dL, low-dose dopamine (the ≤5 category), a GCS of 12, and creatinine of 2.5 mg/dL lands in the middle band on all six measures. A PaO₂/FiO₂ of 250 sits in the ≥200 tier for a respiration reading of 2; platelets of 80 sit in the ≥50 tier for a coagulation reading of 2; bilirubin of 3.0 sits under the 6.0 cutoff for a liver reading of 2; the dopamine category enters directly as 2; a GCS of 12 sits in the 10-12 tier for a neurological reading of 2; and creatinine of 2.5 sits under the 3.5 cutoff for a kidney reading of 2. Six readings of 2 add to a total of 12 — meaningful, multi-system dysfunction, though nowhere near the ceiling.
The two extremes bracket that middle case cleanly. Normal oxygenation, a healthy platelet count, low bilirubin, no pressors, a GCS of 15, and normal creatinine all read 0, for a total of 0. Push every input to its worst band instead — a PaO₂/FiO₂ under 100, platelets under 20, bilirubin at 13.0, high-dose vasopressors, a GCS of 5, and creatinine at 5.5 — and every reading maxes out at 4, for a total of 24, the worst the scale allows.
Questions
Why does the SOFA score need to be checked more than once?
Because it was built to describe organ dysfunction on a given day, not to forecast an outcome from a single reading. ICU teams typically recalculate it daily, and a rising total over 48 hours is itself an important signal of worsening dysfunction — often more informative than where the figure happens to sit at any one moment. A falling trajectory is one of the more reassuring signs available at the bedside.
How is this different from qSOFA?
qSOFA, also on this site, is a fast three-item bedside screen — respiratory rate, mentation, blood pressure — that needs no lab work and exists to flag patients who might need closer evaluation. This full assessment needs six lab-and-exam inputs including platelets, bilirubin, and creatinine, and is meant for patients already being monitored closely, to characterize and track dysfunction severity rather than screen for it in the first place.
Why is the cardiovascular field a dropdown instead of a dose entry?
The published table scores this component from the exact vasopressor and dose a patient is on, which this calculator simplifies into five pre-built categories matching those thresholds. It keeps the form usable without a drug-rate chart in hand, at the cost of the fine gradation the original table allows within each category.
What does a total of 12 mean for a patient?
It reflects meaningful dysfunction spread across roughly half the six systems tracked, though the figure alone doesn't translate directly into a survival estimate — this tool was designed to describe organ dysfunction, not to serve as a standalone mortality calculator. Higher totals correlate with worse outcomes in the critical-care literature, but the trend over time and the clinical picture around the number both matter more than any single reading in isolation.
Is this score used to diagnose sepsis?
It's part of how sepsis is defined under the Sepsis-3 consensus criteria, which describe sepsis as suspected or confirmed infection accompanied by an acute rise of two points or more on this scale. The reading by itself doesn't diagnose infection — it quantifies the organ dysfunction that, combined with infection, defines sepsis under that framework.
Does the respiratory sub-score account for whether a patient is on a ventilator?
Not separately, and that's a real simplification here. The original table splits the two worst respiratory tiers by whether a patient is receiving mechanical ventilatory support; this calculator scores the PaO₂/FiO₂ ratio alone across all five bands, trading that branch for a simpler form that takes under a minute to complete.
References
- Vincent JL et al., Intensive Care Med, 1996 (PubMed)
- Singer M et al., Sepsis-3 consensus definitions, JAMA, 2016 (PubMed)
- CDC — Sepsis
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.