P0457 - Developing age-specific PICU alarm parameters to improve the accuracy of provider orders.

Background: Vital signs are integral to the ongoing assessment of patient’s stability. The monitoring of vital signs can help clinicians anticipate patient deterioration and clinical improvements. Currently, there is little research on what are acceptable vital sign parameters for patients admitted to a PICU. Children admitted to a PICU will have abnormal vital signs due to critical illness. However, established parameters are based on children who are healthy, such as the PALS Vital Sign Reference Chart. In this project, we present a method for personalizing vital sign parameters to our specific PICU population in an effort to improve the accuracy of parameters ordered by resident physicians, improve overall nursing compliance of parameters entered into the patient’s bedside monitor and reduce the total number of non-actionable alarms.

Methodology: This is a retrospective quality improvement project in the Pediatric Intensive Care Unit of an Academic Children’s Hospital. Vital sign data was collected from discharged PICU patient’s over the course of 6 months. Data was then separated out by age and disease state. The data was processed to develop a standardized order set in our electronic medical record in which the age categories followed the age groupings of the PALS Vital Sign Reference Chart. After initiation of our order set we compared the accuracy of vital sign parameters ordered to the patient’s bedside monitor. Additionally, we monitored the amount of provider interventions required after admission. These interventions are the changes required in the ordered vital sign parameters due to increases of non-actionable alarms or inappropriate parameters based on patient’s age or conditions.

Discussion: Pre-implementation results showed that, 279 out of the 749 (37.2%) individual parameters set on the bedside monitor matched the ordered MD vital sign parameters, with only 2 of the 85 bedside monitors (2.3%) matching the MD order exactly. Out of the 749 individual parameters 83 (11.1%) required MD intervention.

Post order set implementation, 658 out of the 977 (67.3%) of the individual parameters matched the MD orders with only 33 out of the 106 bedside monitors (31.1%) matching exactly. Out of the 977 individual parameters 55 (5.6%) required MD intervention.

Conclusion: The introduction of a standardized order set that utilizes data-driven vital sign “normals” from our PICU, resulted in a near 13-fold increase in vital sign parameters matching the bedside monitor exactly. Additionally, it showed a near 180% increase in individual parameter accuracy on the bedside monitors with a 50% decrease in provider interventions. By developing a standardized approach to vital sign parameters in our PICU we improved the accuracy of ordered vital signs by resident physicians as well as improved nursing compliance with the bedside monitor.