Physician Practice Patterns: Chest X-Ray Ordering for the Evaluation of Acute Cough Illness in Adults

Abstract

Objectives. The authors examine which clinical factors contribute to the clinician suspicion of pneumonia, as well as the relationship between clinical factors, clinician suspicion of pneumonia, and ordering chest X-rays (CXR). Methods. Three hundred consecutive adults presenting to the clinic with acute cough in the winter of 2003 were studied. Using standardized encounter forms, data were collected on sociodemographics, illness impact, symptoms, tobacco use, past medical history, vital signs, physical examination findings, chest X-ray result, and clinician diagnoses. Clinicians rated their suspicion of pneumonia on a 5-point Likert scale. Multivariable logistic regression analysis was used to determine independent predictors of clinician suspicion of pneumonia and of ordering of CXRs. Results. Clinician suspicion of pneumonia was low in the majority of patients presenting for evaluation of cough (63%). Higher clinician suspicion of pneumonia was predicted by advanced patient age (odds ratio [OR]: 4.6; 95% confidence interval [CI] [1.2–18.1]), shortness of breath (2.4; [1.0–6.0]), fever (5.5; [1.8–17.5]), tachycardia (3.8; [1.1–13.1]), rales (23.8; [5.7–98.7]), and rhonchi (14.6; [5.2–40.5]). CXRs were ordered in 19% of patients presenting with acute cough. Intermediate clinician suspicion of pneumonia (OR: 7.9; 95% CI: [2.8, 22.5]) (v. low suspicion), advanced patient age ([.greaterequal] 65 years) (9.2; [2.7, 31.6]) (v. ages 18–44 years), and decreased breath sounds on examination (5.1; [1.8, 14.3]) are independent predictors of ordering a CXR. Among patients with a clinical diagnosis of pneumonia (n = 31), CXRs were ordered in only 61%. Conclusions. Advanced patient age and physical findings on chest examination influence clinician practice in obtaining CXRs, beyond their contribution to clinician suspicion of pneumonia. Physicians do not appear to endorse recommendations that the diagnosis of community-acquired pneumonia be based on or confirmed by CXR.