sapsii-score

name	sapsii-score
description	Calculate SAPS-II (Simplified Acute Physiology Score II) for ICU patients. Use for mortality prediction, severity assessment, or international ICU benchmarking.
tier	validated
category	clinical

SAPS-II Score Calculation

The Simplified Acute Physiology Score II (SAPS-II) is a severity scoring system developed from a European/North American multicenter study. Calculated from the worst values in the first 24 hours of ICU stay.

When to Use This Skill

• Hospital mortality prediction
• Severity stratification
• International benchmarking (widely used in European ICUs)
• Research cohort matching
• Quality improvement initiatives

Score Components (First 24 Hours)

Variable	Scoring Range	Points
Age	<40 to >=80	0-18
Heart Rate	<40 to >=160	0-11
Systolic BP	<70 to >=200	0-13
Temperature	<39 or >=39C	0-3
PaO2/FiO2 (if ventilated)	<100 to >=200	6-11
Urine Output	<500 to >=1000 mL/day	0-11
BUN	<28 to >=84 mg/dL	0-10
WBC	<1 to >=20 x10^9/L	0-12
Potassium	<3 or >=5 mEq/L	0-3
Sodium	<125 or >=145 mEq/L	0-5
Bicarbonate	<15 to >=20 mEq/L	0-6
Bilirubin	<4 to >=6 mg/dL	0-9
GCS	<6 to >=14	0-26
Chronic Disease	AIDS/Hematologic Malignancy/Metastatic Cancer	9-17
Admission Type	Scheduled Surgical/Medical/Unscheduled Surgical	0-8

Critical Implementation Notes

1. PaO2/FiO2 Scoring: Only scored for patients on mechanical ventilation OR CPAP/BiPAP. Non-ventilated patients get 0 points for this component.

2. Mortality Probability: Calculated using:

   sapsii_prob = 1 / (1 + exp(-(-7.7631 + 0.0737*sapsii + 0.9971*ln(sapsii+1))))
   

   This 1993 formula tends to overestimate mortality in modern ICU cohorts (calibration drift). Consider recalibration for contemporary populations.

3. Time Window: Uses worst value from ICU admission to 24 hours after.

Dataset Availability

MIMIC-IV

SAPS-II is available as a pre-computed derived table. Materialize with:

m4 init-derived mimic-iv          # All derived tables including sapsii

The derived table provides the total score, predicted mortality probability, and all 15 component sub-scores (age_score, hr_score, sysbp_score, temp_score, pao2fio2_score, uo_score, bun_score, wbc_score, potassium_score, sodium_score, bicarbonate_score, bilirubin_score, gcs_score, comorbidity_score, admissiontype_score).

BigQuery users already have this table via physionet-data.mimiciv_derived.sapsii without running init-derived.

MIMIC-IV implementation details:

• Comorbidity Definitions (ICD-based):
  • AIDS: ICD-9 042-044, ICD-10 B20-B22, B24
  • Hematologic Malignancy: ICD-9 200xx-208xx, ICD-10 C81-C96
  • Metastatic Cancer: ICD-9 196x-199x, ICD-10 C77-C79, C800
• Admission Type: Classified using admissions.admission_type (elective vs not) + surgical service flag from services table.
• GCS Handling: GCS < 3 is treated as null (erroneous or tracheostomy). Sedated patients use pre-sedation GCS.

MIMIC-IV limitations:

• Urine output is summed over available hours, not extrapolated to 24h. For stays <24h, this may overestimate severity.
• Follows MIT-LCP/mimic-code canonical implementation.

eICU

SAPS-II is not pre-computed in eICU (only APACHE IV is provided in apachepatientresult). It must be calculated from raw tables. 13 of 15 components are straightforward:

Component	eICU Source
Age	patient.age (string; "> 89" for elderly)
Heart Rate, Systolic BP, Temperature	vitalperiodic / vitalaperiodic
PaO2, FiO2	lab
Ventilation status	respiratorycare / respiratorycharting
Urine Output	intakeoutput
BUN, WBC, K, Na, HCO3, Bilirubin	lab (text labname, not numeric itemid)
GCS	nursecharting

eICU limitations:

• Comorbidities: The diagnosis table uses free-text diagnosisstring with variably populated icd9code across the 208 sites. ICD-based extraction (AIDS, hematologic malignancy, metastatic cancer) may have incomplete capture. Consider supplementing with pasthistory table or text matching on diagnosisstring.
• Admission type: No direct elective/surgical classification as in MIMIC. Must approximate from patient.unitadmitsource and service fields.
• Ventilation detection: Different logic than MIMIC — uses respiratorycare and respiratorycharting tables rather than procedural events.

See scripts/mimic-iv.sql for the full MIMIC-IV implementation. An eICU script is not yet available.

Example: Severity Distribution

SELECT
    CASE
        WHEN sapsii < 25 THEN 'Low (<25)'
        WHEN sapsii < 50 THEN 'Moderate (25-49)'
        WHEN sapsii < 75 THEN 'High (50-74)'
        ELSE 'Very High (>=75)'
    END AS severity_category,
    COUNT(*) AS n_patients,
    ROUND(AVG(sapsii_prob), 3) AS avg_predicted_mortality
FROM mimiciv_derived.sapsii
GROUP BY 1
ORDER BY 1;

Example: Component Analysis

-- Which components contribute most to high SAPS-II?
SELECT
    'Age' AS component, AVG(age_score) AS avg_score FROM mimiciv_derived.sapsii WHERE sapsii >= 50
UNION ALL
SELECT 'Heart Rate', AVG(hr_score) FROM mimiciv_derived.sapsii WHERE sapsii >= 50
UNION ALL
SELECT 'Systolic BP', AVG(sysbp_score) FROM mimiciv_derived.sapsii WHERE sapsii >= 50
UNION ALL
SELECT 'GCS', AVG(gcs_score) FROM mimiciv_derived.sapsii WHERE sapsii >= 50
UNION ALL
SELECT 'Comorbidity', AVG(comorbidity_score) FROM mimiciv_derived.sapsii WHERE sapsii >= 50
ORDER BY avg_score DESC;

References

• Le Gall JR, Lemeshow S, Saulnier F. "A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study." JAMA. 1993;270(24):2957-2963.