How Much Overtesting Is Needed to Safely Exclude a Diagnosis? A Different Perspective on Triage Testing Using Bayes'...

Area of Science:

• Diagnostic test evaluation
• Medical decision-making
• Health services research

Background:

• Diagnostic testing often involves triage and confirmation steps.
• Minimizing invasive or costly confirmation tests is desirable.
• However, too few negative confirmation tests may indicate missed diagnoses.

Purpose of the Study:

• To quantify the necessary proportion of negative confirmation tests to safely exclude disease.
• To apply this to pulmonary embolism (PE) diagnosis.
• To establish a quality benchmark for diagnostic processes.

Main Methods:

• Utilized Bayes' theorem to model the relationship between test results and disease probability.
• Applied the model to pulmonary embolism (PE) diagnosis using Clinical Decision Rule (CDR), D-dimer, and CT-angiography (CTA).

Main Results:

• A minimum of 67% negative CT-angiography (CTA) scans are required to ensure less than 1% of pulmonary embolism (PE) cases are missed in triage-negative patients.
• This necessitates that no more than 24% of patients with PE undergo triage testing.
• Pre-test probability, triage test accuracy, and confirmation test negativity are interconnected.

Conclusions:

• The proportion of negative confirmation tests is a critical, yet often overlooked, factor in diagnostic accuracy.
• Maintaining an optimal, not too high or too low, rate of negative confirmation tests can serve as a quality indicator.
• This principle is vital for safe and effective disease exclusion, as demonstrated in pulmonary embolism diagnosis.