Alarm Fatigue

---

Big Data in practice - making real changes in healthcare delivery and improving patient outcomes.
Big Data is the new buzz but don’t be put off. According to Dr Veena Goel, Paediatrician and Clinical Informatics Fellow at Lucile Packard Children’s Hospital Stanford, amidst the hype are some real applications for Big Data to improve the quality of healthcare delivery. Dr Goel was a keynote speaker at HISA’s Big Data conference in Sydney last year that brought together leaders in healthcare to share their experiences of working to harness the power of Big Data.

---

Dr Goel, when we talk about ‘Big Data’ and ‘analytics’, what are we describing?
“In the current era of rapidly growing technological capacities, our ability to capture, collect, and save data has seen an exponential rise and the result is the concept of ‘Big Data’. I think of analytics as allowing us to process data in such a way that we can make meaningful interpretations from it and it is one of the biggest challenges associated with Big Data.”
How can Big Data improve treatment delivery and patient outcomes?
“In healthcare, the gold standard we aspire to use to drive medical decisions is evidence-based randomised control trials. However, the majority of decisions that clinicians are faced with must occur in the absence of such evidence. I think this is where we have an opportunity to leverage Big Data within our learning healthcare system. Meaningful analysis and interpretation of Big Data can allow us fill in so many of the gaps in what we do not know in healthcare.”
Can you share some examples of how Big Data and analytics are being used to make a difference in the Australian healthcare system?
“Absolutely. I had the privilege of participating in the Big Data 2015 conference in Sydney, where I learned about so many innovative examples of the use of Big Data and analytics in Australia. From the national e-health strategy to genome sequencing, to population health efforts, the contributions to the field are invaluable. The work of Professor Enrico Coiera and colleagues in the Centre for Health Informatics at Macquarie University has made profound contributions to helping advance our understanding of issues with healthcare delivery and improving healthcare safety and quality.”
All over the world we are hearing stories about how Big Data is being used to predict epidemics and prevent avoidable deaths. Can you comment on this?
“The possibilities and opportunities for using Big Data in population health efforts on a global scale are real and exciting. For example, the Google Flu Trend functionality showed great accuracy in predicting the influenza season. Mobile phone tracking efforts have been employed to track cholera in Haiti after the earthquake and to predict the spread of the Ebola epidemic in Africa.”
“But as these Big Data applications are being discovered and implemented, we are learning a lot about the challenges and limitations in its use as well. Issues such as false or inaccurate predictions and conclusions and management of public expectations plus data security and privacy all have ethical implications that we must be mindful of.”
Accumulating the data is one thing - analysing and applying it effectively is another. Would you agree?
“Yes and accumulating tremendous quantities of data brings up issues of data integrity such as ‘how accurate is the data and can we trust it?’ ‘Where do we keep the large amounts of data?’ ‘How do we keep it secure and ensure patient privacy?’ And these issues are just skimming the surface!”
“Once the data is collected and accumulated then we try to analyse it in such a way that we can draw conclusions that can be applied back within the healthcare setting. But the data comes in so many different forms, can be fraught with bias and can be very difficult to draw accurate and reliable conclusions from. And once we get to the point of having achieved some sort of reliable analysis, managing change within the healthcare setting poses its own unique challenges.”
“Managing change amongst clinicians is not an easy task, especially within the complexities of the healthcare system, where day-to-day decisions are often life-or-death!”
Big data has the potential to drive real change in models of treatment delivery. Can you use your experience with alarm fatigue to illustrate this?
“There are so many treatment models that are currently employed with little underpinning evidence. Let me give you an example. The parameters used to monitor children in the hospital setting have for the longest time been based on anecdotal and small-sample observations.”
“In the alarm fatigue trial my colleagues and I did last year, we leveraged vital sign data from over 7,000 patients extracted from the electronic health record and bedside monitors, to better understand and define data-driven thresholds for age-based heart and respiratory rates in children¹. Our analysis revealed that by using data to drive how we define normal vs. abnormal vital signs, we could safely decrease hospital alarms by over 50 percent.”
“Any of you who have had a loved one in the hospital can relate to the constant noise of alarms and beeps that go off and how often these alarms are not clinically significant or meaningful. We have managed to use Big Data to change our approach to monitoring children in the hospital. This is just one small example of how Big Data has driven a change in the way care is delivered.”
What prompted you to undertake this trial into alarm fatigue?
“I clinically practice medicine as a paediatric hospitalist, so I take care of hospitalised children and deal with monitor-alarms on a daily basis, so this project really aligned well with my interests as a doctor and also with my clinical informatics training.”
“The main reason I chose to undertake the work to reduce alarm fatigue in our hospital was because the Joint Commission for Accreditation of Healthcare Organizations here in the US released a Sentinel Event Alert back in 2013 identifying alarm fatigue as a huge problem in the clinical setting. They quoted that 85-99 percent of alarms do not require clinical intervention and that alarm fatigue can lead to patient death.”²
“Subsequently in 2014, the Joint Commission released a National Patient Safety Goal on Alarm Safety and essentially in it, required that all hospitals establish policies and procedures for managing alarms by January 1, 2016.” ³
What is the difference between alarms in clinical settings for adults and those for children?
“Alarm settings are typically quite different between adult and paediatric clinical settings because what we consider “normal” in the two populations is very different. Not only that, “normal” in children is very dependent upon age - so a normal heart rate in a 2 month old is different from a 2 year old and a 12 year old!”
“Because of this, alarm settings are age-dependent in paediatrics. But in adults, typically the normals are defined the same in all people over 18 years of age. There is just so much more variability in the paediatric populations and there is very little research or literature guiding us with regards to how to go about monitoring them.”
How has reducing alarm fatigue made a difference to the working environment for nurses?
“The work that I did to define new data-driven vital sign parameters for children (in 10 different age categories) for heart rate and respiratory rate has actually been implemented at our hospital. We started by piloting the new HR and RR parameters on one unit of our hospital in October of 2014 and demonstrated just over a 30 percent reduction in heart rate alarms.”
“Since then, we have further revised our respiratory rate parameters and launched the new HR and RR parameters across all of the paediatric wards and the paediatric ICU and cardiac ICU at Lucile Packard Children’s Hospital Stanford. We are currently in the process of analysing alarm data since we went live at the end of August, 2015 but nurses are reporting that they feel fewer alarms are going off, and that the workflow around alarm management has improved.”
What do you see as being the foreseeable challenges and opportunities for Big Data in healthcare?
“We have just begun to realise the opportunities for Big Data in healthcare in enabling us to improve safety and efficiency, fill knowledge gaps, cure disease, predict epidemics, and revolutionise population health. And with each discovery and step we make, we are uncovering and understanding the concurrent challenges.”
“One big challenge I foresee is that the healthcare informatics workforce is still young and underdeveloped. We need more clinical informaticians who are educated in clinical medicine, but who also have intimate knowledge of the technologic systems that we are striving to improve upon.”
Dr Veena Goel
Dr Veena Goel is a board-certified paediatrician completing simultaneous training in the first nationally accredited clinical informatics fellowship and a paediatric hospital medicine fellowship, both at Stanford Medicine. Her scholarship focuses on improving bedside monitor alarm fatigue, which was recently declared a National Patient Safety Goal by the Joint Commission. In this work, she is leveraging a unique big dataset to create and implement data-driven alarm parameters for paediatric inpatients. Dr.Goel’s passion is in harnessing her informatics training and clinical and administrative leadership skills in healthcare IT delivery to improve patient safety and quality of care.

---

“Any of you who have had a loved one in the hospital can relate to the constant noise of alarms.”

---

References
1 “Implementation of Data Drive Heart Rate and Respiratory...” 2015. 21 Jan. 2016 <http://www.researchgate.net/publication/281815113_Implementation_of_Data_Drive_Heart_Rate_and_Respiratory_Rate_parameters_on_a_Pediatric_Acute_Care_Unit>
2 “Sentinel Event Alert Issue 50: Medical device alarm safety in ...” 2013. 21 Jan. 2016 <http://www.jointcommission.org/assets/1/18/sea_50_alarms_4_5_13_final1.pdf>
3 “National Patient Safety Goal - Joint Commission.” 2014. 21 Jan. 2016 <http://www.jointcommission.org/assets/1/18/jcp0713_announce_new_nspg.pdf>