Blog

What opioid-induced respiratory depression can teach us about clinical surveillance during the pandemic 

COVID-19 has created an unprecedented demand for ventilators, which continues today as infection surges emerge in different parts of the country. As hospitals introduce new ventilators for these patients – or any new devices – it increases the number of alerts to potentially distract clinicians, risking potential alarm fatigue and cognitive overload.  

Considering the fact that 85% to 99% of alarms per day do not require clinical action, COVID-19 creates added urgency around the need to reduce such nuisance alerts to protect the health and safety risks of the respiratory therapists or nurses who investigate alarms in person while wearing personal protective equipment (PPE), which is also still in short supply. 

Lessons From OIRD 

Fortunately for clinicians, reducing alarm fatigue and improving the relevancy of such alerts has been a priority for hospitals for years and we have already learned quite a bit about how to reduce nuisance alarms. For example, research I collaborated on a few years ago that involved patients with an increased risk for opioid-induced respiratory depression (OIRD) offers lessons that can be applied today.  

Lesson #1: Go beyond the basics 

Basic alarm filtering based on time thresholds for different devices has been occurring since the early 2000s. While somewhat effective, the alarm management strategy did not take into account patients‘ other vital signs. This meant that a sustained alarm could be because the patient is experiencing a cardiac condition, or they could just be out of bed brushing their teeth. Regardless, the clinician has to respond to the alert in person, interrupting her task and contributing to cognitive load. 

Our study, which involved a cohort of patients with sleep apnea recovering from surgery and administered opioids for pain management, attempted to advance that basic filtering technique exponentially forward by first analyzing raw data from multiple device sources. 

Lesson #2: Take alerts out of isolation 

Rather than responding to one metric in isolation, the solution we researched analyzed data – not just alarms – from capnographs and pulse oximeters in real-time that measured patients’ pulse (HR), oxygen saturation (SpO2), respiratory rate (RR), and end-tidal carbon dioxide (ETCO2). At the start of our research, our clinical team faced as many as 427 bedside respiratory depression alarms per hour for just one patient, although the average for patients studied was 182 per hour, or 22,812 for the entire study.  

Applying basic alarm filtering delay techniques was able to reduce the number by 42%, still resulting in more than 13,000 alerts. To build context around these alerts so that the clinician would only be notified when an event was actionable, we configured multiple alert-threshold times through a multivariate rules engine that monitored the values for all four metrics – HR, RR, SpO2 and ETCO2. Such highly customizable continuous clinical surveillance analytics and smart alerts are available today through Capsule Technology’s Medical Device Information Platform (MDIP). 

By analyzing the raw data and only alerting clinicians when an event was actionable, our investigated solution was able to decrease alerts to just 209 for the entire study period – a 99% reduction.  

Lesson #3: Make it Mobile 

In our study, patient alerts were forwarded to the nurse’s mobile phone instead of sounding at the bedside only, which further reduced alarm fatigue and cognitive load. Despite this massive reduction in alerts, our research team independently verified that no actual clinical events were overlooked, and clinicians responded successfully in time to deliver Naloxone to counteract OIRD for affected patients.  

Similarly, Capsule’s MDIP combines high-fidelity device data with multivariate, EHR information for a holistic and complete source of objective information on patients to eliminate non-clinically actionable alerts, but also to support decompensation prediction and clinical decision making prospectively.  

This insight is crucial for surveillance of patients with COVID-19 so clinicians can don PPE and respond in-person only when necessary. Over the longer term, a continuous clinical surveillance strategy with the smart alerts available through MDIP improves the clinicians’ experience – which promotes retention – and enables them to work more efficiently without risk to patient safety.  

Alerts aside, whether it is COVID-19, OIRD or any other condition, the ability to predict a patient’s clinical deterioration and intervene before an adverse health event occurs – even saving a life – is the ultimate goal for any clinical surveillance strategy. 

About the author: 
Sarah Williams, RRT, is Director of Product Management—Surveillance.