Interoperability and Post-Crash Care: Knowing the Health Needs of the Humans Who Were Affected by a Crash

Mr. Jonathon Feit

In 2009 – fifteen years ago – USDOT published a little-attended paper called “The Contribution of Medical Conditions to Passenger Vehicle Crashes.” As the title suggests, it looked at the incidence of medical emergencies as a precipitating factor of a crash. The figures were dramatic: Seizures were implicated in 35% of crashes with a medical cause or contributor, diabetes in 20% of crashes, and heart attacks in 11%.

In 2019, Epilepsy Currents published data via the Mayo Clinic, University of Colorado, and others, which indicated that some 36 million people around the world have epilepsy and drive. By the time responders arrive on-scene, the seizure may be over. Indeed, the vehicle’s occupants may be unable to explain the circumstances of what happened. Today, myriad crashes occur that had a medical precipitating factor—however, that fact was unknown until the patient (hopefully) awoke at the hospital. During this process, “offenders” turned into “patients.” Individuals with conditions ranging from uncontrolled blood pressure to multiple sclerosis often appear to be driving under the influence—swerving on the road, slurring their speech—but the medical etiology was only discovered later.

However, damage may have already been done; a failure to identify the medical cause of a crash can lead to delayed care, and “knock-on” effects including the financial and/or reputational damage associated with DUI. The operational impact across the healthcare system is likewise compelling—and insufficiently discussed. According to the former prehospital care director of a major Denver-area hospital system: “Every trauma activation called into our trauma center generates a rapid response of hospital personnel awaiting EMS arrival. Hospital staff access to critical information prior to patient arrival, such as the actual mechanism of the crash, number and demographic of occupants, past medical history, medications, etc., is almost non-existent. Such information known prior to patient arrival would accelerate patient care and treatment while simultaneously reducing the potential for iatrogenic errors -- such as administering medications contraindicated due to hospital staff being unaware of what medications a patient is currently taking.”

People are living longer today, and they want to live and drive independently. Precision medicine is helping to protect and heal people with serious illnesses, but are the tools and training used in the field today sufficiently insightful? For example, are roadside tests clinically accurate enough? What if roadside tests could access health information exchanges to contextualize their findings, e.g., if an individual is found to have a particular substance in their blood after a crash, could interoperability with a health information exchange indicate that the substance may have come from a prescribed medicine?

Are responders of “every patch color”—including Fire, Emergency Medical Services, and Public Safety—sufficiently trained to identify and engage strokes, low blood sugar, psychiatric challenges, and so much more? Or – in an effort to reduce the crisis of roadway deaths – have we “knee jerked” to solutions without collapsing data silos? For example, a national movement is underway to advance whole blood delivery in the field. Unquestionably, bringing blood to the patient is a good idea; propping circulation up buys time after trauma. What if the patient has a POLST or other advance directive, or they are a Jehovah's Witness and would reject a transfusion? (Ambulance services will ask how they can carry multiple blood types to be ready for whatever they will encounter. But they only recently learned about how to deliver Type O, and they can innovate to carry multiple types.) The open question—which is more about workflow than technology—is how crews can know a patient’s identity and needs before they reach the scene, then share their insights with a receiving hospital in real-time.