The Pneumonia Severity Index (PSI), also known as the PORT (Patient Outcomes Research Team) Score, is a validated clinical prediction rule developed by Fine et al. (1997) that stratifies adults with community-acquired pneumonia (CAP) into five risk classes based on predicted 30-day mortality. It was derived from a prospective cohort of 14,199 hospitalised pneumonia patients and validated in a further 38,039 patients — making it one of the most rigorously validated scoring systems in infectious disease. The PSI incorporates 20 variables across five domains: demographic characteristics (age, sex, nursing home residency), comorbid illnesses (neoplastic disease, liver disease, CHF, cerebrovascular disease, renal disease), physical examination findings (altered mental status, respiratory rate ≥30/min, systolic BP <90 mmHg, temperature <35°C or ≥40°C, pulse ≥125/min), laboratory results (BUN ≥9 mmol/L, sodium <130 mmol/L, glucose ≥14 mmol/L, haematocrit <30%, arterial pH <7.35, PaO2 <60 mmHg or SpO2 <90%), and chest radiograph findings (pleural effusion). The scoring system begins with a three-step algorithm to identify very low-risk patients (Class I) based on clinical criteria alone — without requiring laboratory tests — enabling rapid outpatient decision-making. Classes I and II (≤70 points) have very low 30-day mortality (<1–2%) and are suitable for outpatient treatment. Class III (71–90 points, approximately 2.8% mortality) may be treated as brief inpatient or observation cases. Classes IV (91–130 points, 8.2% mortality) and V (>130 points, 29.2% mortality) require hospital admission, with Class V patients often requiring ICU-level care. The PSI was specifically designed to safely identify low-risk pneumonia patients for outpatient management, which has been shown to reduce hospital admissions without increasing mortality.
PSI Score = Age (years) − 10 (if female) + points for comorbidities, physical exam, labs, CXR; Class I (no risk factors + age exclusions): outpatient; Class II (≤70): outpatient; Class III (71–90): consider admission; Class IV (91–130): admit; Class V (>130): ICU consider
- 1Step 1: Assign Class I if patient is <50 years old, has none of the listed comorbidities, and has normal vital signs and mental status — these patients can be treated as outpatients without further scoring.
- 2Step 2: For all other patients, begin scoring. Start with age in years (males: age; females: age − 10).
- 3Step 3: Add comorbidity points: nursing home resident +10; neoplastic disease +30; liver disease +20; CHF +10; cerebrovascular disease +10; renal disease +10.
- 4Step 4: Add physical exam points: altered mental status +20; respiratory rate ≥30/min +20; systolic BP <90 mmHg +20; temperature <35 or ≥40°C +15; pulse ≥125/min +10.
- 5Step 5: Add laboratory points: BUN ≥30 mg/dL (≥9 mmol/L) +20; sodium <130 mEq/L +20; glucose ≥250 mg/dL (≥14 mmol/L) +10; haematocrit <30% +10; PaO2 <60 mmHg or SpO2 <90% +10; arterial pH <7.35 +30.
- 6Step 6: Add CXR point: pleural effusion +10.
- 7Step 7: Total score → Class I (algorithm), II (≤70), III (71–90), IV (91–130), V (>130); correlate with 30-day mortality estimates and make admission decision.
No further lab or CXR scoring required for Class I determination
A young patient with no comorbidities and normal vital signs is Class I by the three-step algorithm — the safest risk class.
Class III: 30-day mortality approximately 2.8%; clinical judgement for admission
An elderly female with a comorbidity and borderline labs scores in Class III, where brief hospitalisation or observation is often chosen.
Class V patients require aggressive management, early ICU assessment, and consideration of vasopressors
Multiple severe physiological derangements produce a very high PSI score. Class V pneumonia carries approximately 1 in 4 mortality and requires intensive care.
CURB-65 is simpler; PSI more validated for outpatient selection; use both for cross-validation
CURB-65 and PSI generally agree in directing hospital admission for severely ill patients. PSI is more sensitive for safely identifying low-risk outpatient cases.
Emergency department triage: rapidly identifying low-risk CAP patients (Class I–II) safe for outpatient treatment., representing an important application area for the Pneumonia Severity Psi in professional and analytical contexts where accurate pneumonia severity psi calculations directly support informed decision-making, strategic planning, and performance optimization
Hospital admission decision-making: directing moderate-risk patients (Class III) to observation or short-stay units., representing an important application area for the Pneumonia Severity Psi in professional and analytical contexts where accurate pneumonia severity psi calculations directly support informed decision-making, strategic planning, and performance optimization
ICU referral: Class V patients with multiple physiological derangements require intensive care assessment., representing an important application area for the Pneumonia Severity Psi in professional and analytical contexts where accurate pneumonia severity psi calculations directly support informed decision-making, strategic planning, and performance optimization
Antibiotic stewardship: correlating PSI class with pathogen probability to guide antibiotic spectrum selection., representing an important application area for the Pneumonia Severity Psi in professional and analytical contexts where accurate pneumonia severity psi calculations directly support informed decision-making, strategic planning, and performance optimization
Quality improvement: tracking PSI-guided admission rates as a measure of appropriate hospitalisation in pneumonia care programmes., representing an important application area for the Pneumonia Severity Psi in professional and analytical contexts where accurate pneumonia severity psi calculations directly support informed decision-making, strategic planning, and performance optimization
Immunocompromised Patients
In immunocompromised patients (HIV with CD4 <200, solid organ transplant, haematological malignancy, long-term corticosteroids), the PSI underestimates severity because these patients are at high risk of opportunistic pathogens (Pneumocystis jirovecii, Cryptococcus, CMV, fungi) that cause different disease patterns. Immunocompromised patients with pneumonia should generally be admitted regardless of PSI score for consideration of BAL and non-standard antibiotic cover.'}
Healthcare-Associated Pneumonia
{'title': 'Healthcare-Associated Pneumonia', 'body': 'Patients from nursing homes, dialysis centres, or with recent hospitalisation are at risk for multidrug-resistant (MDR) organisms including MRSA and Pseudomonas — organisms not covered by standard CAP antibiotic regimens. The PSI does score nursing home residency (+10 points) but may still underestimate the risk from MDR pathogens in these patients, who may need broader empirical antibiotic coverage.'}
Legionella Pneumonia
{'title': 'Legionella Pneumonia', 'body': 'Legionella pneumophila classically produces a high PSI score rapidly due to hyponatraemia, haematuria, elevated creatinine, and altered mental status. Legionella urinary antigen testing should be requested in any PSI Class III–V patient admitted with pneumonia. Treatment requires fluoroquinolone or macrolide (beta-lactams are inactive against Legionella). Standard CAP regimens (amoxicillin alone) are ineffective.'}
Pneumonia in Pregnancy
{'title': 'Pneumonia in Pregnancy', 'body': 'Pregnant women are at higher risk of pneumonia (particularly influenza-related) and have different physiological parameters (lower baseline pCO2, lower HCO3, higher respiratory rate at baseline) that affect PSI scoring. Pneumonia in pregnancy often requires inpatient assessment regardless of PSI score, and antiviral treatment (oseltamivir) should be started early for influenza pneumonia without awaiting test results.'}
| Risk Class | Score | 30-Day Mortality | Recommended Site of Care |
|---|---|---|---|
| I | Algorithm (no score) | <0.1% | Outpatient |
| II | ≤70 | 0.6% | Outpatient |
| III | 71–90 | 2.8% | Observation / brief inpatient |
| IV | 91–130 | 8.2% | Inpatient ward |
| V | >130 | 29.2% | Inpatient (ICU consideration) |
What is the difference between PSI and CURB-65?
PSI (20 variables) is more complex but more sensitive for identifying low-risk outpatient cases (Classes I–II), reducing unnecessary hospitalisation. CURB-65 (5 variables: Confusion, Urea, Respiratory rate, Blood pressure, age ≥65) is simpler and faster to apply at the bedside. CURB-65 is better at identifying severe pneumonia requiring ICU care. Most guidelines recommend using both: PSI for outpatient/inpatient decisions and CURB-65 for severity assessment. Patients with CURB-65 ≥3 should be admitted regardless of PSI.
Can the PSI be used for atypical pneumonia pathogens?
The PSI was developed for community-acquired pneumonia overall and is not specific to any pathogen. It is broadly applicable regardless of whether the pathogen is typical (Streptococcus pneumoniae) or atypical (Mycoplasma, Legionella, Chlamydophila). However, Legionella pneumonia can cause very high PSI scores rapidly due to multi-organ involvement (renal failure, hyponatraemia, deranged LFTs), and these patients often progress to Class IV–V quickly.
Should all Class II patients be treated as outpatients?
Class II PSI (score ≤70) patients have a predicted 30-day mortality of 0.6–0.9% and are generally suitable for outpatient treatment with oral antibiotics. However, clinical judgement must be applied: poor social circumstances, inability to take oral medications, failure to improve within 24–48 hours, or hypoxia on room air may justify admission even with a low PSI score. The PSI guides but does not replace clinical decision-making.
Why does the PSI subtract 10 points for female sex?
The Fine et al. cohort data showed that for the same clinical severity, females had lower 30-day mortality than males — possibly due to hormonal differences, different comorbidity patterns, or healthcare-seeking behaviour. The sex adjustment corrects for this empirically observed survival difference. It does not reflect any difference in physiological severity but rather in outcomes within similar clinical presentations.
What antibiotics are used for PSI Class I–II CAP?
For mild outpatient CAP (Class I–II), NICE and BTS guidelines recommend: amoxicillin 500 mg three times daily for 5 days (if no atypical coverage needed), or doxycycline 200 mg loading then 100 mg daily. If atypical pneumonia is suspected (young patient, gradual onset, extrapulmonary features), add a macrolide (clarithromycin) or use doxycycline alone. Fluoroquinolones (levofloxacin) are reserved for penicillin-allergic patients or suspected resistant organisms.
Does the PSI apply to COVID-19 pneumonia?
The PSI was not validated for COVID-19 pneumonia. COVID-19 severity scoring uses specialised tools including NEWS2 (National Early Warning Score), CRB-65, and the WHO severity classification (mild/moderate/severe/critical). Some studies have applied PSI to COVID-19 pneumonia with reasonable performance, but disease-specific risk factors (elevated CRP, ferritin, D-dimer, lymphopenia, bilateral CT changes) are not captured by PSI. Most COVID-19 guidelines use their own dedicated scoring systems.
What is the role of procalcitonin in CAP management?
Procalcitonin (PCT) can help guide antibiotic initiation and de-escalation in CAP. Low PCT (<0.25 ng/mL) suggests a viral or non-bacterial aetiology and supports withholding or stopping antibiotics. However, PCT should not override clinical assessment — a patient with severe respiratory failure and bilateral infiltrates should receive antibiotics regardless of PCT. In low-risk CAP (PSI Class I–II), PCT-guided therapy can safely reduce antibiotic duration.
Can the PSI identify patients who need ICU care?
The PSI Class V (>130 points) has approximately 27–29% 30-day mortality and identifies patients who likely need intensive care. However, the PSI underperforms compared to specific ICU admission criteria (ATS/IDSA minor criteria for severe CAP). IDSA/ATS guidelines recommend ICU admission if a patient meets 1 major criterion (need for ventilation or vasopressors) or ≥3 minor criteria (respiratory rate ≥30, PaO2/FiO2 <250, multi-lobar infiltrates, confusion, uraemia, leucopenia, thrombocytopenia, hypothermia, hypotension requiring fluids).
Pro Tip
The Class I algorithm is the most time-saving step: if the patient is under 50, alert, has normal vital signs, and no listed comorbidities, they are Class I — discharge home with oral antibiotics without needing any labs. This one step alone can prevent hundreds of unnecessary blood tests and admissions annually in a busy emergency department.
Did you know?
The PSI study published in NEJM in 1997 by Fine et al. was based on data from 38,039 patients across 275 hospitals — one of the largest validation studies ever conducted for a clinical prediction rule at the time. When the results showed that 51% of hospitalised pneumonia patients could safely have been treated as outpatients (Class I–II), it fundamentally changed how physicians conceptualised the need for hospital admission in pneumonia and led directly to significant reductions in hospitalisation rates across the USA and Europe.