Emergency medical services (EMS) has had a hard time being considered part of the US healthcare system. At times viewed as transportation, logistics, public safety, or the generic "first responder," the role of EMS providers in regard to patient mortality and morbidity can be elusive when observing various aspects of the US healthcare infrastructure and provider environment.
However, new advances in machine learning and artificial intelligence are shaping a future where EMS providers may solidify their role in the traditional care continuum. In the near future, longstanding provider roles will begin to become more integrated as machine learning is applied more broadly across healthcare organizations and stakeholders.
In this system, all participants in an episode of care will be able to contribute data and insight to AI-assisted algorithms that in turn deliver valuable decision support at the point of care. Mortality and morbidity rates for many conditions could be substantially reduced through earlier identification made possible by AI and machine learning.
While still in its early days, this technology can already benefit one extremely serious condition: sepsis.
Sepsis is the nation’s deadliest condition, responsible for more than 1.6 million deaths a year. In addition to its high mortality rate, sepsis is the leading cause for readmissions and the costliest condition in the US, accounting for as much as 6.2 percent of hospital costs.
The high cost and clinical difficulty of treating sepsis is due to the challenge of diagnosing sepsis early, before a patient’s condition has already deteriorated and they are transferred to the Intensive Care Unit. In fact, 86 percent of physicians state that symptoms of sepsis are often misattributed to other conditions. This is extremely problematic given that the mortality rate increases 7.6 percent every hour that sepsis goes untreated.
Deciphering the often misleading attributes of sepsis is where machine learning and data analytics can be useful to aid clinicians in expediting the identification of conditions that are exceptionally challenging to diagnosis early in the care continuum before critical escalation.
For example, Intermedix’s Condition Awareness uses machine learning to continuously learn from operational and provider data to accelerate the detection of conditions like sepsis. Further, because it leverages a unique mix of socioeconomic information, it is able to predict the risk of sepsis with a high degree of accuracy early in the continuum of care.
While this technology is currently deployed within hospitals, EMS providers also have a unique opportunity to leverage data analytics to identify patients at risk of developing deadly and costly conditions like sepsis early, before they and their patients have arrived at the hospital. In fact, studies show that 80 percent of sepsis deaths could be prevented with rapid diagnosis and treatment. For EMS providers, this is particularly impactful as 85 percent of sepsis patients are admitted through the emergency department, many of whom arrive by EMS transport.
So perhaps there is a different vision of the future, one that is facilitated by a more open definition of healthcare and predictive analytics. In this vision, EMS is an integral part of the healthcare system. From the time of 9-1-1 activation, the healthcare system could use data, provider judgement and the capabilities of machine learning and analytics to positively assess risk, identify patients that should receive a more in-depth assessment and allow providers across the continuum of service delivery to provide interventions that positively impact the patient’s condition, quality of life and recovery. In this vision of the future, one driven by the patient’s perception of their care, we have recognized that it is our responsibility as healthcare providers to leverage every tool at our disposal to improve the care we deliver.