Medical Examiner


Ambulances can be dangerous places.

Not long ago in Western Pennsylvania, an ambulance was dispatched to help an elderly woman whose heart was beating irregularly. Although the patient was awake and her blood pressure was normal, the paramedics on the scene detected a worrisome cardiac rhythm on their monitor: The heart was beating too fast, and each beat appeared widened on the screen. The patient’s condition was consistent with a serious and sometimes fatal heart rhythm called ventricular tachycardia. One of the paramedics called a local hospital, and a doctor there told him to administer intravenously 100 milligrams of a potent anti-arrhythmic drug, intravenous lidocaine hydrochloride. In the cramped ambulance, the medic grabbed a 2-gram syringe of lidocaine in concentrated form, which must be diluted in a bag of saline and dripped into the vein slowly. Thinking he had a different vial, the paramedic quickly injected the entire syringe into the patient. The woman went into cardiac arrest and died.

In 1999, the Institute of Medicine published its reportTo Err Is Human, which estimated that up to 98,000 patients may die each year because of the mistakes of doctors, nurses, and other hospital workers. But few published studies have tried to quantify or even characterize the injuries to patients that take place before they reach the hospital. How frequent and how serious are the mistakes that take place in ambulances—and are there simple changes that could help prevent them?

EMTs and paramedics (click here for an explanation of the difference; the gist is that paramedics have more training) take care of about 30 million patients in the United States every year. In the past, ambulances did little more than “scoop and run”—pick patients up and speed them to the hospital. But over the last 25 years, the importance of the “golden hour” has been enshrined: Research shows that critically ill patients who get the right medication or procedure quickly have a much better chance of survival. As a result, ambulance workers do a lot more active patient care than they used to. In 2003 in Pennsylvania, for example, EMTs and paramedics performed at least one medical intervention—by administering medication, establishing an IV, performing CPR, or even placing a patient on artificial ventilation—on about 35 percent of ambulance runs.

Based on what we know about hospital-based medical error, ambulances may be one of the more dangerous places to be a patient. Studies have shown that medical error is more common when conditions are variable, like in the emergency room, than it is in other parts of the hospital. The problem likely has little to do with experience or skill. Instead it’s about the lack of predictability: Doctors and nurses make more mistakes when they work under changing conditions. Think about that and compare the working conditions of paramedics and EMTs with an operating room. Before surgery, an entire staff is prepped with information about a patient’s condition, medical history, and the anticipated plan of action. On an ambulance run, there is no plan. Paramedics and EMTs have to improvise as they encounter the obese, frail, terrified, combative, near-dead, stoned, violent, and newly born. And they have to deliver care in a cramped space with relatively few resources.

In hopes of making ambulances safer, some states have tried self-reporting by EMTs and paramedics. The idea is to urge ambulance workers to report their mistakes and those of their colleagues, often anonymously and without risk of punishment. The model is the aviation industry, which has decreased the number of airplane crashes since federal rules began requiring pilots to report all near misses. As yet, however, EMTs and paramedics haven’t gone for self-reporting. They tend to be loath to call each other out. In Pennsylvania, a state-run Web site for reporting ambulance medical errors has seen little traffic.

Since they can’t rely on self-reporting, most ambulance services use a classic quality-assurance model to track patient safety. For example, if a patient with chest pain doesn’t get nitroglycerine or aspirin in the field, a doctor or nurse at the receiving hospital is supposed to recognize the lapse and report the error. These reports might work fine for detecting errors of omission. But they’re not likely to catch an error like a misdosage, which wouldn’t be recorded anywhere and is nearly impossible for a doctor or nurse to identify by examining the patient in the ER. And EMTs and paramedics hate quality assurance, which they call “train and blame.” When I directed a suburban ambulance service, I played a part in the scheme: A medical-error report would come across my desk, and, if it was valid, I was asked to arrange for more training or recommend discipline. The focus was all on what was wrong with the medic—was he or she badly trained, incompetent, lazy, drunk? The answer was usually none of the above. Most EMTs and paramedics save lives daily and do far more good than harm. When mistakes happen, focusing the inquiry on their faults obscures the real question: What was wrong with the ambulance system that allowed a competent paramedic or EMT to commit a significant error?

The field of injury-prevention science offers some answers. The father of this discipline, William Haddon Jr., showed that when people get hurt, a web of contributing factors is almost always at work. Injury researchers have demonstrated that altering human behavior matters less for reducing error than making changes to the physical environment. A classic example: Driver’s ed does not reduce injury from car crashes nearly as much as air bags or highway guardrails do.

In the ambulance, these lessons reinforce what hospitals have learned about reducing medical errors by standardizing medications, automating orders, and building in systemic checks. A 1998 study in the Journal of the American Medical Association showed that when physicians enter medication orders into computers (instead of writing or speaking the orders to a nurse), hospital med errors fall by 17 percent. Redundancy also helps. If a drug must be ordered by a doctor, dispensed by a pharmacy, and administered by a nurse before a patient gets it, then each provides a check against providing the wrong drug or the wrong dose.

Ambulances, of course, don’t have these luxuries. Most do not use computers and are usually staffed by only two people, one of whom has to drive. But it would be fairly simple—and potentially lifesaving—to standardize all the dosages and packaging of the drugs that EMTs and paramedics give out. At the moment, some states maintain master lists of medications that specify which medications ambulances may use. But brand, packaging, and concentration levels often vary. In some communities, ambulances restock as they travel to different hospitals, which means they carry different variations of the same drugs at the same time. Ambulances would be safer for patients if they were all organized in the same way, with cardiac meds on the left and respiratory meds on the right. Syringes could all be prefilled and even color-coded with standard adult and pediatric dosages. 

Sound like a no-brainer? All of this would cost money, and the benefits would be hard to quantify. And some EMTs and paramedics will resist, because they’ll feel that color-coded syringes imply they can’t think independently and well in an emergency. For years, doctors have similarly complained that standardization in the hospital forces them to practice “cookbook medicine.” But boring or demeaning as they may be, these systems work.