Clinical Scorecard: Commentary on the PASC Score for Long COVID: Comparing Clinical and Statistical Diagnostic Accuracy

At a Glance

Key Highlights

• The authors argue that recruiting cases from an LC clinic may overrepresent people with multiple symptoms, potentially reducing estimated sensitivity through spectrum bias.
• Requiring cases to have at least one neuropsychiatric symptom may bias sensitivity upward; 62% of the cases reportedly had brain fog, worth 3 PASC score points.
• Using fully recovered individuals and healthy controls for specificity validation may inflate specificity because these groups may not represent symptomatic patients with alternative etiologies.
• The commentary states that case-control validation cannot establish positive or negative predictive values because it does not reflect disease prevalence in a population of interest.
• The authors conclude that the PASC score was not validated by the reported estimates, which they consider biased.

Guideline-Based Recommendations

Diagnosis

• For evaluating an LC classifier, the commentary favors a representative sample of people with LC drawn from the general population.
• For specificity assessment, the commentary favors symptomatic patients with overlapping symptoms but without an etiology involving prior SARS-CoV-2 infection.

Management

Monitoring & Follow-up

Risks

• Clinic-based case recruitment may produce spectrum bias when the cohort overrepresents people with multiple symptoms.
• Requiring a symptom that contributes to the score as part of case eligibility may bias sensitivity upward.
• Using recovered individuals and healthy controls as specificity comparators may inflate specificity.
• Case-control validation does not reflect population prevalence and therefore cannot determine positive and negative predictive values.

Patient & Prescribing Data

The commentary discusses patients assessed for possible Long COVID and critiques a study using LC clinic cases, fully recovered individuals, and healthy controls.

No treatment or prescribing information is provided.

Clinical Best Practices

• When evaluating an LC classifier, select case and comparison cohorts that reflect the intended clinical population and the diagnostic alternatives encountered in practice.
• Interpret sensitivity and specificity estimates in light of case-selection criteria and the symptom spectrum represented in each cohort.
• Do not infer positive or negative predictive values from a case-control validation design.

Related Resources & Content

• Commentary on the PASC Score for Long COVID: Comparing Clinical and Statistical Diagnostic Accuracy