Lessons learned from Johns Hopkins’ alarm reductions

The task force that reduced clinical alarms at The Johns Hopkins Hospital learned many valuable lessons along the way, says Maria Cvach, MSN, RN, CCRN, assistant director of nursing clinical standards.

Cvach offers these tips to risk managers who want to duplicate the success of Johns Hopkins:

  • Understand the problem and state the goal. For example, the goal can be “to eliminate 30% of alarm conditions throughout the hospital.”
  • Share your goals with key hospital staff, including clinicians, administration, clinical engineers, and biomed technicians.
  • Recognize the problem as institution-wide
  • Recognize the resolution of the problem as long-term and ongoing.
  • Obtain the support of your administration to achieve your goals.
  • Engage a multi-disciplinary team to study and address the problems. Include nursing staff, clinical engineers, biomed technicians, device vendors, and others as appropriate.
  • Analyze the problem. Access the right data, and know how to extract it. Identify key data such as (but not limited to) bed number, purpose, and timeframe/length or alarm condition.
  • Conduct a fault tree analysis to understand the failures to respond to critical physiologic alarm conditions in a timely manner.
  • Identify a key metric, such as the average number of alarm conditions per bed per day.
  • Implement safety checks on alarm settings.
  • Revise alarm default parameters in each unit to actionable levels. Recognize that settings might vary from one unit to another.
  • Implement revisions or changes incrementally.
  • Prioritize and differentiate between actionable alarm signals in each unit, such as visual vs. audible. Recognize that settings might not be the same from one unit to another.
  • Define alarm types as false, true, nuisance, actionable, or other categories, and ensure that definitions are understood by unit staff.
  • Gather quantitative baseline data to evaluate alarm conditions.
  • Examine logs from the network that track alarm messages from devices in order to capture the quantitative data.
  • Compare pre- and post-data to measure changes.
  • Ask the right questions and gather the right data:
    • Where the alarms are coming from? What is the bed number?
    • Who is the patient?
    • What is the cause?
    • How long are alarms sounding?
    • How many alarms are occurring in units?
    • When an alarm signal goes off, what do you do?
    • When an alarm goes off, how do you hear it?
    • What is the average number of patient alarms per bed, per day?
    • What is the workflow of a clinical unit e.g. backup notification, nurses per unit, assignments, etc.?
    • What is the clinical significance of an alarm? What are the high/low priority alarms along with high/low risk alarms?