Understanding the Pneumonia Severity Index

The PSI, also known as the PORT score, was developed to standardize risk assessment in community-acquired pneumonia. Rather than relying on clinical intuition alone, the index systematically weighs 19 clinical and laboratory variables to predict 30-day mortality.

PSI stratifies patients into five risk classes:

  • Class I: Low risk; outpatient management appropriate
  • Class II: Low risk; consider brief observation
  • Class III: Intermediate risk; likely requires hospitalization
  • Class IV: High risk; hospitalization recommended
  • Class V: Very high risk; intensive care consideration

The scoring system accounts for non-modifiable factors (age, sex, comorbidities) and acute presentation markers (vital signs, examination findings, laboratory abnormalities). This multifactorial approach improves prognostic accuracy compared to single parameters alone.

PSI Scoring Calculation

PSI aggregates points across three domains: patient characteristics, physical examination findings, and laboratory values. Each variable contributes a fixed point value based on clinical significance.

PSI = age + sex adjustment + nursing home status + comorbidities

+ neoplasm + liver disease + CHF + cerebrovascular disease

+ renal disease + mental status + respiratory findings + vital signs

+ temperature + pulse + pH + BUN + sodium + glucose

+ hematocrit + partial pressure oxygen + pleural effusion

  • Age — Patient age in years; contributes 1 point per year
  • Sex — Male 0 points; female −10 points
  • Comorbidities — Presence of renal disease, liver disease, CHF, cerebrovascular disease, or malignancy; each worth 10−30 points
  • Vital signs — Respiratory rate, systolic BP, temperature, and pulse; abnormal values add 10−20 points each
  • Laboratory values — BUN, hematocrit, sodium, glucose, arterial pH, and partial pressure of oxygen; each contributes 10−30 points if abnormal

Key Considerations in PSI Interpretation

Several clinical nuances affect how PSI scores guide management decisions.

  1. Age and frailty may diverge from risk score — Elderly patients with minimal comorbidities may score lower than younger patients with multiple chronic diseases. Clinical judgment about functional status, social support, and ability to comply with outpatient care remains essential alongside the numerical score.
  2. Laboratory results must be contemporaneous — Outdated lab values distort the score. Electrolytes, renal function, and oxygenation should reflect the acute pneumonia presentation, not chronic baseline levels. Missing laboratory data may require empirical assumption or repeat testing.
  3. PSI does not account for atypical presentations — Immunocompromised patients, those with severe underlying lung disease, or those with respiratory failure may require ICU admission despite lower PSI scores. The index performs less reliably in immunosuppressed or critically ill populations.
  4. Class II patients still warrant careful observation — Although classified as low risk, Class II patients occupy a grey zone. Factors like age >50, recent illness, poor social circumstances, or difficulty following up may justify admission even when PSI suggests outpatient management feasible.

Clinical Application and Admission Criteria

PSI class directly informs admission decisions. Patients scoring Class I or II are generally safe for outpatient treatment with oral antibiotics and close follow-up. Those in Class III and above typically require hospital admission; Class IV and V patients frequently need critical care assessment.

However, PSI serves as a guideline, not a rigid rule. Several factors override the score:

  • Patient unable to reliably take oral medications or maintain hydration
  • Uncertain diagnosis or concern for atypical organisms
  • Significant comorbidity exacerbation (e.g., acute heart failure, acute renal injury)
  • Lack of reliable outpatient follow-up within 24−48 hours
  • Desire for admission despite low score for patient safety or reassurance

Research supports using PSI to reduce unnecessary admissions while maintaining safety, particularly in resource-constrained settings.

Frequently Asked Questions

What is the difference between CURB-65 and PSI scoring for pneumonia?

CURB-65 uses five simple bedside variables (confusion, urea, respiratory rate, blood pressure, age ≥65) and is faster to calculate; it also predicts 30-day mortality. PSI is more comprehensive, incorporating 19 variables including comorbidities and laboratory findings, offering greater granularity but requiring more data collection. Both are valid; CURB-65 suits emergency settings, while PSI provides detailed risk stratification when full information is available.

Can PSI score guide antibiotic selection?

PSI primarily stratifies mortality risk and admission need, not antibiotic choice. Antibiotic selection depends on suspected organisms (typical community pathogens vs. atypical), severity, local resistance patterns, and patient allergies. A Class I patient still requires appropriate coverage for Streptococcus pneumoniae and Haemophilus influenzae. Severity-based guidelines (e.g., IDSA) alongside PSI inform antibiotic decisions.

How accurate is the PSI score in real-world practice?

PSI demonstrates good discrimination in large populations but imperfect performance in individual patients. Studies show sensitivity 75−90% for identifying low-risk patients but miss some high-mortality cases. Clinician experience and patient-specific factors improve decision-making beyond the score alone. PSI reduces unnecessary admissions by approximately 10−20% when used systematically.

What happens if I'm missing some laboratory values for PSI calculation?

PSI can be calculated with incomplete data using step 1 (demographics and comorbidities alone), though step 2 (vitals) and step 3 (labs) enhance accuracy. Missing labs bias the score unpredictably; if results are unavailable, assume normal values or obtain them urgently. Never knowingly use outdated labs; repeat testing may be needed to reflect acute-phase values.

Does PSI apply to all pneumonia patients?

PSI was designed for community-acquired pneumonia in immunocompetent hosts. It performs poorly in immunosuppressed patients (HIV, transplant, chemotherapy), healthcare-associated pneumonia, ventilator-associated pneumonia, or those with severe immunodeficiency. Use clinical judgment and specialist input in these populations; PSI scores may underestimate risk.

What is the mortality rate for each PSI class?

Approximate 30-day mortality rates: Class I <1%, Class II 1−2%, Class III 2−3%, Class IV 8−15%, Class V >30%. These represent population averages; individual risk varies. Older patients, those with multiple comorbidities, and those requiring ICU support experience higher mortality within each class.

More health calculators (see all)