Clinical Indicators of Radiographic Findings in Patients With Suspected Community-Acquired Pneumonia: Who Needs a Chest X-Ray?

This was a retrospective study designed to evaluate the efficacy of six clinical indicators to predict the presence of pulmonary infiltrates on chest radiography.

The study population will include a convenience sample of David Grant Air Force Medical Center patients that meet study criteria. Adult patients (18 years of age or older) with acute respiratory symptoms and positive chest radiographs from October 2004 through April 2005 will be enrolled as positive cases. There will be an equal number of controls with acute respiratory symptoms but negative radiographs. Controls will be matched by age and date of radiograph. Patients/Cases with suspected hospital-acquired (positive radiographs within 10 days of hospital discharge) or aspiration pneumonia will be excluded

Six clinical indicators were selected based on the cumulative results of seminal studies that included: cough, sputum production, temperature, heart rate, respiratory rate, and findings on physical examination. Abnormal vital signs were defined as temperature of 100.5 degrees Fahrenheit (38 degrees Celsius) or greater, heart rate greater than or equal to 100 beats per minute, and respiratory rate greater than 20 breaths per minute. Positive physical examination findings consisted of the presence of crackles, decreased breath sounds, dullness to percussion, egophony, or rhonchi

Outpatient emergency room physicians will record data regarding the clinical indicators prior to obtaining chest radiographs; thus, they will be blinded to the radiographic findings. All chests radiographs will be subsequently interpreted by a board-certified radiologist blinded to the clinical findings. Positive radiographic findings will be defined as the presence of a new air space opacities in the setting of acute respiratory symptoms. Patients with equivocal radiographic findings interpreted as "possible pneumonia" will be considered positive cases. For each positive case, a single control was randomly selected from a list of eligible controls with acute respiratory symptoms, negative chest radiographs, and a date of birth within five years of that of the positive case.

Once patients are appropriately identified as cases or controls, outpatient charts will be reviewed to gather data regarding the six clinical indicators will be entered into the study database for analysis upon completion of the study. The sample size necessary to predict the proportion of patients with CAP who have all six clinical indicators was calculated as 385, assuming 95% confidence level and a 5% margin of error. Sample size estimates assume a random sample, which is not possible for us to obtain in this study; any conclusion that we draw will thus e cautiously interpreted. We will conduct this study for a minimum of one year with the intent to continue until a sample size of at least 385 is obtained. Upon completion of the study sensitivities and specificities and accompanying 95% confidence intervals for clinical indicators will be made to determine which combination of indicators best predicts the presence of pulmonary infiltrates in patients with CAP. Further analysis will be made within specific age groups (e.g., 18-40, 41-60, older than 60) and amongst those patients with pre-existing pulmonary disease