The specific treatments given in the ED

Type II (moderate) exacerbations were defined as occurring when only two of these symptoms were present. Type III (mild) exacerbations included patients with only one of the three symptoms, in addition to at least one of the following findings: upper respiratory symptoms, increased wheezing, increase in respiratory rate or heart rate by 20%, or fever without another cause. Other data collected (if available) included the initial pulse oximeter reading, arterial blood gas, WBC count, and chest radiograph. The specific treatments given in the ED and prescribed at dismissal were also analyzed.

Relapse was defined as a return visit within 14 days of the initial presentation with the patient having persistent or worsening symptoms related to AECB. The time period of 14 days was chosen from previous studies that showed patients with AECB were at excessive risk for persistent symptoms for approximately 14 days and that these symptoms are likely related to the initial exacerbation.

Statistical Analysis

The outcome of interest was the relapse rate, which refers to the proportion of ED dismissals ending in relapse. Historical, diagnostic, and treatment variables in the ED were compared for visits that ended in relapse and those that did not. For differences between groups, y² and analysis of variance were used. The level of significance was set at 5%; p values are two-tailed. Univariate and multivariate analyses were performed to investigate the historical and clinical variables that may be used to predict an increased risk of relapse. A logistic regression model was constructed to determine independent predictors of relapse for AECB. The hypothesis of a significant influence of a variable was assessed by the likelihood ratio test. Data were entered into a customized database, and all statistical analyses were performed using Statistical Analysis Software (SAS Institute, Cary, NC). Results are expressed as mean ± SD.