Written by James Welch, CEO Arc Biomedical Consultants (firstname.lastname@example.org)
Mr. Welch is a Clinical Engineer with 17 yrs experience in hospitals and over 24 yrs as an executive in the medical device industry. His focus has been on applying technologies to improve patient safety through continuous surveillance monitoring. Mr. Welch has ten patents and articles in the field of wireless physiologic monitoring, surveillance systems and alarm management. He regularly contributes to the AAMI Foundation on alarm safety and is a voting member on a number of International Standards committees.
Early detection of physiologic deterioration is essential in improving patient safety in acute care hospital settings. Patients in non-ICU settings who are recovering from surgery or special procedures are especially vulnerable because of private or semi-private room settings prevents direct observation and nurse to patient ratios are often 1:6. Experts in Rapid Response Systems (RRS) have arrived at a consensus that strengthening early detection through continuous monitoring is essential in improving the effectiveness of RRS but only if such systems do not impose a burden on the clinical staff. The high incidence of nuisance alarms and cost are two of the major barriers preventing broader adoption of continuous monitoring on the general care floor.
All electronic vital signs monitoring originated in intensive care settings, primarily anesthesia and critical care. Alarms were set to identify a change in the “condition” of the patient. Early technologies often triggered false alarms due to signal noise and limitations in first generation software algorithms. Technology advancements over the past decade have significantly lowered these sources of false alarms. In addition, many true alarms that are of short duration do not require immediate attention in non-ICU acute care settings. These non-actionable clinical alarms contribute to nuisance alarms, a dangerous condition where true alarms are lost in a cacophony of non-actionable alarm events. The modest addition of a few seconds of alarm delay will filter non-actionable events. Alarm effectiveness becomes then a tradeoff between relaxing threshold settings and adding delays. The combination determines at what physiologic condition persistence should a clinician be alerted. Narrow thresholds and short delays are appropriate for intensive care settings where patients are unstable and undergoing aggressive treatment, but not for recovering patients where early ambulation is important for recovery. These patients require continuous surveillance to detect unexpected deterioration in their vital signs. This is especially true for patients on opioid pain management.
The purpose of surveillance monitoring is to detect physiologic deterioration early enough for a successful clinical intervention. Alarm configurations in these settings must have high specificity (a true actionable event) and acceptable sensitivity. Taenzer et al reported on the adoption of continuous pulse oximetry on a post-surgical ward at Dartmouth Hitchcock Hospital with significant improvements in reduced ICU transfers. The researchers compared an 11-month pre and post time period where they reported a 48% reduction in ICU transfers and later reported an estimated annual savings of $1.48M. Patients who were transferred to the ICU had a 20% reduction in overall length of stay . Alarm management was a key performance indicator that contributed to the sustainability of the system. SpO2 alarms were set at 80% SpO2 with an alarm holdoff delay of 15 seconds which resulted in 4 alarms per patient per day.
A barrier for other hospitals to adopt the default Dartmouth alarm settings was the dramatic departure of the low SpO2 alarm threshold setting from the traditional value of 90%. We developed an evidence-based alarm methodology using captured high fidelity numeric SpO2 data to determine the effects of various SpO2 alarm settings and applied this to data from multiple hospitals using the same device as Taenzer. A matrix showing the combined effects of alarm threshold settings and alarm delays provided clinicians quantitative choices to reduce non-actionable alarms. A setting of 85% low SpO2 and 15 seconds delay achieved significant reduction in nuisance alarms similar to the Dartmouth results but at a more clinically acceptable alarm setting. There was no report of any missed respiratory deterioration either of these studies.
We expanded the methodology to include all vital signs using an alternative technology to Taenzer. Results for SpO2 alarm rates were comparable. The combined alarm rates for all continuous vital signs (Heart Rate, Resp Rate, Blood Pressure, SpO2) was 10 alarms per patient per day. Further improvements in this technology has lowered alarm rates to 6 alarms per patient per day.
An alarm occurrence of 6 alarms per day is equivalent to routine vital signs every 4 hrs. The difference is that continuous vital signs alarm warns the nurse when to respond to a patient rather than rely on routine checks. Automated entry into the Electronic Health Record (HER) further improves workflow and clinical documentation.
Limitations to adoption:
Broad adoption for continuous monitoring for the detection of deterioration is less a technology limitation and more a change management challenge. Continuously monitoring patients who previously received only spot check monitoring requires changes in nursing practice. New policies and governance is also needed from the executive suite to the individual clinical settings in order to achieve a sustainable solution. AAMI Foundation has developed a framework for hospitals to navigate towards a more evidence-based determination for appropriate alarm setting. This capability maturity model is meant to engage hospital leadership in a systematic approach to sustainable reduction in nuisance alarms throughout the institution and thereby improve patient safety.
A significant reduction in avoidable adverse events can be realized with continuous electronic vital signs surveillance. Hospitals have several technology choices, each of which can adopt evidence-based practices to reduce the occurrence of nuisance alarms. The barriers to adoption are breaking down as more peer reviewed articles demonstrate the clinical and financial effectiveness of such systems. The AAMI Foundation coalition for alarm management are developing tool kits for hospitals to successfully reduce alarm hazards. These efforts address the technical limitations of current technologies. Our research and development has demonstrated technology can be successfully deployed to significantly reduce avoidable patient safety events. Adoption therefore is a function of the hospital’s culture of patient safety and the willingness to invest in technology solutions.