Do you know that when you walk into an emergency department, your doctor may run a test for illicit-drug use without telling you?
Yes, this is something we can and often do. And in many cases, it's done for a good reason.
Suppose you come in acting confused or excessively sleepy. Your doctors need to know right now whether your condition is caused by alcohol or drugs, or whether it's something else like a brain infection, a stroke or a seizure. Learning that you have a mind-altering drug in your system is an important piece of the puzzle especially if you are too confused or incoherent to tell us what is going on.
Or let's say you come to the emergency department, flipping out your family says they've never seen you like this before. If we test you for PCP and the result is positive, then we have a reason for your erratic behavior, and now we might be able to spare you a spinal tap (to look for that brain infection) or avoid artificially paralyzing you to keep your head still in the CT scan (while looking for that stroke).
So there it is emergency-care providers must have the ability to obtain rapid laboratory drug-of-abuse tests on demand. These drug screens are an essential tool in our diagnostic kit, enabling us to take care of our patients and protect public health.
Of course, it isn't so simple. There are some real tradeoffs to testing emergency-department patients for illicit drugs. As practitioners we need to pay attention to the downsides of these tests so we don't overuse them.
First there are ethical issues of autonomy and confidentiality. The principle of autonomy states that patients should be able to decide whether or not to undergo testing or treatment for anything. When you check in to a hospital, you sign a form giving consent for routine testing, including blood and urine tests for lots of things. This makes sense it means that as doctors, we don't have to check with you for every run-of-the-mill test we order. But the question here is whether or not testing for drugs and alcohol without your explicit consent should be considered routine.
Confidentiality is, of course, the other major ethical problem with ordering illicit-drug tests on our patients. While the Health Insurance Portability and Accountability Act legally protects all medical information from public disclosure, just ordering the test increases the risk that a breach of confidentiality could expose this sensitive information.
False-positive tests are another concern. Urine drug tests use immunoassays to screen for multiple illicit drugs. While they tend to be quite accurate, cross-reactions with other medications have been demonstrated for example, over-the-counter decongestants have been shown to light up the amphetamine test incorrectly. Also, these tests may identify previous drug use but fail to tell us that the patient was using drugs recently (the marijuana test can be positive over a month after use, for instance). This can be confusing to the diagnosticians who are trying to figure out what is causing today's symptoms.
Last, there are issues of cognitive biases, mental prejudices on the part of doctors that can interfere with our ability to make the best decisions for our patients. Fundamental attribution error is one such bias, in which a health care provider inadvertently and wrongly blames a patient for her illness. Take the case of the patient with belly pain, who tests positive for cocaine: she becomes the "drug user in Room 2."
On one level, this categorization may be helpful it reminds me to discuss drug addiction and counseling options with the patient. But it also can undermine her care. As a practitioner, I may develop a sense of futility: whatever we do to fix her immediate problem isn't going to help because she is just going to go out and continue to abuse her body and mind. Now she is given a lower priority on my giant to-do list, which is wrong. Priorities in care should be determined by an unbiased assessment of urgency. A positive drug test can scramble this order.
The downsides of emergency-department drug screens become even more apparent when we consider that the benefits of ordering the test are often unclear. Take chest pain, for instance, the second most common reason for seeking care in an emergency department. Chest pain associated with cocaine use can be due to spasms of the coronary arteries. These spasms may be made dangerously worse when we give beta-blockers, a commonly prescribed class of medication for patients with heart-attack symptoms that are not caused by cocaine use. Because we know that many patients presenting with chest pain will not self-report cocaine use, conventional wisdom holds that drug-testing these patients is a good idea. It allows us to figure out who should and shouldn't get beta-blockers. The problem is, there's no evidence to suggest that doing clandestine cocaine tests actually saves lives, or even improves patient outcomes.
Similarly, we often run drug screens on patients who come to the emergency department with psychiatric complaints like acute depression. Then we have to wait for hours for the results before we can proceed with their care. Yet studies have shown that getting drug tests makes no difference in the subsequent psychiatric care and treatment of these patients.
So what is the solution? Legal or regulatory barriers to drug tests are a bad idea. When patients are very sick or acting completely wacky, the screens are essential. Sometimes we can give antidotes. Often we can hone in on our diagnoses. But we also need to stop ordering drug screens for patients who don't need them. A good rule: if it isn't going to change your management, don't order the test. And even if ordering that test might change your care, consider other options like simply asking patients if they use drugs before undermining their trust. Just like everything else we do in medicine, the risks and benefits need to be weighed ahead of time. Drug-testing needs to be put to the same measure.
Dr. Meisel is a Robert Wood Johnson Foundation clinical scholar and an emergency physician at the University of Pennsylvania.