Clinical prediction rule for pulmonary infiltrates

Area of Science:

• Internal Medicine
• Pulmonology
• Clinical Decision Making

Background:

• Pneumonia is a common cause of acute respiratory illness in adults.
• Accurate diagnosis of pneumonia is crucial for timely and appropriate treatment.

Purpose of the Study:

• To derive and validate a clinical prediction rule for identifying pneumonic infiltrates in adult patients with acute respiratory illness.

Main Methods:

• A derivation set of 1134 patients and two validation sets (150 and 152 patients) from emergency departments were analyzed.
• Clinical data including symptoms, signs, comorbidities, and chest roentgenogram results were collected.
• Stepwise logistic regression was used to develop and validate the prediction rule.

Main Results:

• Key predictors identified: temperature > 37.8°C, pulse > 100 bpm, rales, decreased breath sounds, and absence of asthma.
• The rule demonstrated good discriminatory ability in the derivation set (ROC area 0.82) and validation sets (ROC areas 0.82 and 0.76).
• Predicted probabilities closely matched observed pneumonia incidence across settings.

Conclusions:

• A clinical prediction rule based on readily available findings accurately identifies pneumonia in adults with acute respiratory illness.
• The rule maintains discriminatory ability across different patient populations and settings.
• This tool can aid clinicians in pneumonia detection and guide decisions regarding radiologic imaging.