| Learning Objective #1: describe the frequency of patients who have left against medical advice from U.S. emergency departments. | |||
| Learning Objective #2: describe five significant factors that contribute to patients who have left against medical advice from U.S. emergency departments. |
Design: A cross-sectional study was conducted.
Sample: A convenience sample of 444 ED managers throughout the United States participated.
Framework: A systems model was used to identify the patient flow factors in the emergency department (input, throughput, output) that influenced patients to leave AMA.
Methods: A 72-item survey on ED characteristics was developed and reviewed by content experts. Univariate and multivariate analyses examined the relationships between ED characteristics and the annual number of patients who left AMA. Logistic regression analyses were used to predict patients who left AMA.
Results: The backward elimination logistic regression model contained five significant variables. Patients were more likely to have left AMA as the collaboration between physicians and nurses decreased (p=0.023), if the hospital was not JCAHO-accredited (p=0.003), as the waiting time for evaluation for non-urgent care decreased (p=0.009), and as the turnover rate (in a quadratic relationship) for registered nurses decreased (p=0.032, turnover rates <13.0%; p=0.014, turnover rates >13.0%). The sensitivity (78.8%) and specificity (81.2%) of the final model were high.
Conclusions: The strong predictor model identified five factors in the ED system that influenced patients’ decisions to leave AMA. Although research concerning the patient’s actual decision making is warranted, ED nurses and managers might reduce AMA visits and improve patient safety by maintaining stable levels of nurse staffing and facilitating strong nurse-physician collaborative relationships. The influence of JCAHO accreditation and waiting times for evaluation by acuity warrants further investigation.