Vertical People Have Wants; Horizontal People Have Needs
Healthcare is a consumer product like any other in that customer satisfaction depends upon the fulfillment of wants and needs. The majority of emergency room patients arrive vertically—the walking wounded, the worried-well, and the mildly ill. Their wants are many; their needs are few. As an emergency physician, my attention is rightfully directed toward the horizontal patient, the one too sick to sit upright.
Despite being an emergency physician, a good many of the patients I evaluate arrive with chronic medical conditions. When a patient with a five-year history of abdominal pain arrives in my department—a patient who has already seen her PCP multiple times, in addition to three different gastroenterologists and a gynecologist—it is all but certain that she will leave frustrated and dissatisfied. Most of these folks have already completed the entire gamut of available tests; often many times over, and yet they still hold out hope that a community ER doc will be able to diagnose what has eluded multiple specialists. Sorry, but that isn’t going to happen. Chronic, undiagnosed medical conditions are rarely, if ever, solved in the emergency room. Usually, the best I can do is to address the patient’s immediate need for pain control.
So it’s a given that some of my patients will leave dissatisfied. These are often the same people who come to the hospital believing that they are paying for a diagnosis and anything short of that is a disservice. Many also believe they have the right to dictate their treatment. They’re wrong on both counts. Emergency room patients pay for my expertise in the recognition of emergency conditions. My obligation is to render an expert opinion, and then to act in a manner consistent with my medical ethics. I am under no obligation to provide anybody with narcotics, antibiotics, or work notes. Having said that, I also recognize that mine is a service industry. In general, when assessing the wants and needs of patients, I allow myself to be guided by the Latin maxim: primum non nocere (first, no harm).
If a patient requests a low-yield, low-radiation x-ray after twisting an ankle, I’ll order it. Similarly, if someone asks for the more expensive Zithromax over the cheaper amoxicillin for an ear infection, I’ll almost always prescribe it.
But what if a patient requests an antibiotic for a viral infection, where the treatment will provide no benefit and may cause harm? Here, the divide between patients’ wants and needs begins to yawn wide. While an unnecessary antibiotic is unlikely to provoke anything more serious than a case of diarrhea, catastrophic allergic reactions do occasionally occur. And, there is the larger concern of antibiotic resistance, which comes into play after millions of doses of unnecessary antibiotics are taken by the population at large.
When faced with such a quandary, many physicians simply prescribe the antibiotics anyway believing it to be simpler, faster, and more likely to result in a satisfied customer. At least that’s the prevailing dogma, though the majority of folks are quite willing to forego an antibiotic as long as their concerns have been addressed. Providing such care takes more time, and I believe it is time well spent. Unfortunately, when hospital administrators review metrics at the end of the month, this extra time spent with a patient isn’t reflected as a quality measure. If I see 0.2 fewer patients per hour than my colleagues, it just might be enough to get me fired—and at one facility where I used to work, this seemingly trivial difference was enough to do just that.
Harder to deal with is the narcotic seeker. Most start out with legitimate pain-causing conditions, like low back pain, migraine, irritable bowel syndrome, or endometriosis, but somewhere along the way, their ongoing need for narcotics begins to trump their medical troubles. Such patients now have two problems—their original condition and narcotic addiction. And while it’s certainly faster and easier to give such patients what they want, to do so is akin to placing a salt lick in the yard and wondering why your shrubbery is being eaten.
Sometimes you’re damned either way. A few years ago, I had to respond in writing to my director after a nurse complained about a delay in my ordering narcotics for a well-known frequent flyer who was making her life miserable by incessantly ringing the call bell. Would I please justify the reason for the delay? No mention was made of the fact that this same patient had already taken several doses of her own narcotics at home and had received a whopping dose of Valium prior to my arrival. No sooner had I fired off a response than a letter arrived in my mailbox from her insurance company asking me to justify my prescription of narcotics to her. Wasn’t I aware of her history of prescription narcotic use? The letter included a list of dates and providers who had written her prescriptions in the past. My name was the last in a long line. The insurance company was very concerned. Would I please curb my liberal prescribing habits?
The fulfillment of wants and needs is a tricky business, and nowhere is that more true than within the sphere of testing. Patients expect physicians to use the best, most sophisticated, technically advanced studies available to them. They anticipate that everything that can be done to ensure their well-being will be done. And here is where the warm glow of CT scanning often represents the fulcrum upon which a patient’s wants and needs balance. Many folks simply aren’t satisfied that they have received the full benefit of all that modern science can provide without the reassurance of a CT scan, and physicians, whose incomes are increasingly tied to patient satisfaction scores, are all too willing to comply. The use of CT scanning has increased more than twenty-fold over the last quarter century. In 2007, emergency physicians ordered more than sixteen million of them. The odds of a patient receiving a CT scan during an emergency room visit, regardless of complaint, now stands at an astonishing one in eight. Despite this dramatic increase in utilization, the diagnosis of serious illness hasn’t increased and the rate of hospitalizations hasn’t decreased. In today’s world, CT scanning is far more often used to assure a patient that a medical condition isn’t present than to actually diagnose a condition that is.
So why is this a problem? Computed tomography involves placing a patient on a motorized table, which then moves through a large donut-like structure where a series of narrow, rotating, fan-shaped x-rays are taken at distances ranging from one to twenty millimeters apart. A processor then analyzes the data to create a cross-sectional image of the scanned body part, which, when viewed in sequence, provides an incredible amount of anatomic detail. Its major advantage over plain x-rays is its ability to image not just bone, but the surrounding vessels, tissues, and organs. The price for this detail, however, is a marked increase in radiation exposure.
A good bit of what we know about radiation’s harmful effects comes from studies conducted by the ABCC (Atomic Bomb Casualty Commission), and later by RERF (Radiation Effects Research Foundation) to monitor the effects of the Hiroshima and Nagasaki bombings. Based on this data, an increased lifetime cancer risk was detected in survivors exposed to as little as 5 mSv of radiation—a dose comparable to that obtained from today’s CT scans. Radiation is actually a waveform of energy, and its ability to penetrate the human body is the property that allows us to create x-ray images. As the beam travels, dissipating its energy into the surrounding tissue, miniscule amounts of charged ions known as free radicals are formed. These, in turn, are capable of damaging cellular DNA, inducing mutations that ultimately lead to replication errors tied to the formation of cancer cells. These effects are cumulative over time and increase with repeated exposure. (Ironically, when administered in higher doses, radiation has the ability to kill cells outright, which explains why it is also employed as a treatment against the very same cancers it is capable of inducing.) While a one-view chest x-ray results in an effective organ dose of just 0.06 mSv of radiation, CT scans generate far higher doses. The effective radiation exposure from a standard CT trauma series is equivalent to that of one thousand chest x-rays!
What follows are three clinical scenarios where CT scans are typically employed. The patients depicted are composites of folks I encounter on a daily basis.
Let’s say that you’re the parent of a bouncing bundle of energy affectionately known as a toddler. You take your child to the grocery store and place her in the cart, where she babbles and smiles at all the bright colors parading past her eyes, but you are distracted with the thought of preparing dinner and buying two weeks’ worth of groceries. As you pass the frozen food section, you suddenly remember the Cool Whip. Looking back, you spot it no more than a dozen feet behind you. Meanwhile, your daughter, enamored of the coupon dispenser attached to a nearby freezer door, reaches for the lever and stretches just a bit … too far. She falls, head first, onto the linoleum four feet below.
You hear it before you see it—the dull thud like a melon hitting the floor. Before your conscious brain is even aware of where you are looking, your line of vision has traveled downward to where your daughter lies prostrate. From your vantage point, she doesn’t appear to be breathing. It’s one of those frozen moments that seems to last forever, but is broken soon enough by the sound of your own shriek, and then suddenly you can’t move fast enough. You rush to scoop her up. Her eyes are glassy; the lights are on but nobody is home. Her body is limp. You panic and begin to shout her name.
Suddenly, she takes a deep gasping breath and starts to cry. It is not the usual, fussy, I-want-to-watch-Sponge-Bob sort of cry, but rather a wail of startled fright. Having registered your panic, she is panicked as well.
You begin shouting, “Are you okay? Are you okay?” but your daughter’s crying precludes any meaningful exchange. You examine her scalp. Thank goodness, there isn’t any blood. You start touching her all over, feeling for signs of injury, which only serves to make her cry louder. By the time you get to the checkout line, a large “goose egg” has welled up on the back of her head. The store manager asks if you want an ambulance, but since your daughter is awake and not bleeding, you tell him: “No, no—it’s okay.”
Maybe you overreacted. Maybe everything is fine. By the time the groceries are bagged, your daughter’s crying has subsided to a mere whimper punctuated by the occasional staccato gasp that accompanies a hard crying session. You live ten minutes away and decide to head for home.
You gently place your daughter into her car seat, taking extra care to ensure that she is strapped in properly. Then you give her a juice box. You start the Disney songs CD and adjust the rear view mirror so as to see her better. Within minutes she is asleep, but rather than reassuring you, this scares you silly. You call her name, but she doesn’t answer. She’s out. The drive home seems interminable.
By the time you pull into the driveway, you’re a nervous wreck. There are tears streaming down your face. Your panic increases when your daughter fails to wake up despite additional pleading and gentle shaking. She wrinkles her forehead, yawns, but doesn’t open her eyes. This is bad, isn’t it? Why won’t she wake up? Is she in a coma? What is a coma, anyway?
Less than a minute later she stirs, frowns impressively, and proceeds to vomit the apple juice. There isn’t much and it doesn’t spew out, more-or-less dribbles down from the side of her mouth. Still, it scares you more than any vomiting she’s ever done in the past because vomiting means a head injury doesn’t it?
You call 9-1-1 and minutes later the ambulance arrives.
The medics couldn’t be nicer. Your daughter has her eyes scrunched down tight; she wants to be left alone. They look her over and check her vital signs without removing her from the car seat. The blood pressure cuff wakes her up, and she starts to cry. She reaches for you and you reach back, but stop when the medic tells you that the safest thing is for her to remain strapped in the car seat. He recommends an emergency room evaluation, and asks if you still need the ambulance? Wracked with guilt and fear, you aren’t in any condition to drive, so, yes, you still need the ambulance. The driver says, “Okey-doke,” and you’re off. You call your husband who agrees to meet you at the hospital.
With lights flashing and sirens blaring, you skirt the local traffic and get to the hospital in no time flat. You’ve never ridden in an ambulance before and find it simultaneously exciting and calming. You whisper a silent prayer of thanks as your daughter is ushered past the waiting area straight to an exam room set up for children. The walls are painted with balloons and bears, and a crib stands in the corner right beside the rocking chair. There is a flat-screen TV mounted on the wall, the Disney channel playing with the sound down.
A nurse arrives and performs a rapid assessment. She is kind and gentle, but not overly reassuring. When you inquire about your daughter’s condition, she says: “She looks okay to me, but the doctor is going to have to make that determination. He’ll probably order a CT scan to be sure.” Minutes later a registration clerk arrives, followed shortly thereafter by your husband. He looks worried and flustered. He, too, is out of his element.
Free of the car seat, your daughter cries “Daaa-ddy,” and runs into his arms. He picks her up and surveys the swelling on her scalp. Then he asks: “Has the doctor been in yet?”
“Not yet,” you reply.
“Well, where the hell is he, then?” he asks.
By the time the doctor arrives, your daughter appears fine. She is running around, playing. The physician is younger than you would like, and you can’t help but wonder if he has any children of his own. Your husband appears put-out. You can’t decide whether he’s angry with you for allowing this to happen or with the doctor for making you wait.
The physician senses your distress and apologizes for the delay. He keeps his distance, giving your daughter space, as he takes the history. The exam, when he gets around to it, takes a mere two minutes. He shines a penlight in her eyes. He feels along her neck, back, hips and arms, before gently palpating the swelling on her scalp. He taps her knees with a reflex hammer and smiles when she giggles. He watches her walk before declaring that your daughter’s exam is normal except for the bruise.
Then he spends a few minutes talking to you about pediatric head injuries, before adding: “Kids are tougher than we give them credit for. They almost always do fine.” He suggests a period of watchful waiting at home. “Are you okay with that?” he asks.
Sure, your daughter looks better now, but the doctor wasn’t there when she looked like death. He seems overconfident. How can he be so sure? What if she throws up on the way home? What if her skull is cracked? Despite the doctor’s attempt at reassurance, you aren’t reassured. What about the CT scan? Didn’t the nurse say the doctor would order one? Fortunately, the doctor has seen this look on a number of parents’ faces in the past and quickly adds: “Well, if you aren’t comfortable with that, we could do a CT scan.”
Yes! A CT scan. Thank you!
“That would make me feel better,” you say.
“Yes, let’s do it,” your husband adds.
The scan takes just a couple of minutes. Your daughter doesn’t require any sedation. The possibility of radiation risk never enters your mind. After all, if it were dangerous, wouldn’t the doctor have mentioned it? When he returns thirty minutes later with the good news that the scan is normal, you feel an immediate sense of relief. When he tells you that it’s okay to feed your daughter and that it’s not necessary to wake her through the night, you believe him. You leave highly satisfied. The CT is covered by your insurance. You consider the one hundred dollar co-pay a bargain. Your daughter is fine.
Now, let’s consider a different scenario—assume that you’re a forty-year-old man who has just been hired to a new job. Your medical benefits haven’t kicked in yet. Your wife’s job doesn’t offer health insurance, but you pay your bills on time and have a little money squirreled away. You are the father of a twelve-year-old boy who spends his after-school hours playing video games and networking on Facebook. He does well in school, but eats too much junk food and drinks too many sodas.
For convenience, your family eats out at fast food restaurants two or three times a week. Your son’s favorite is Chipotle, where he eats burritos and chips with an ease that arouses a curious sense of envy in you. Still, all this bad food and lack of activity hasn’t gone unnoticed. He’s overweight, and no longer plays soccer because he doesn’t like the running. He’s had constipation problems for as long as you can remember. Several times a year he suffers from stomach cramps. Visits to the pediatrician have failed to render a diagnosis, but he always gets better after a day or two. A dose of laxative usually does the trick, so when he first starts complaining about the pain, you think it’s the same old thing.
It started after breakfast and hasn’t gone away. He didn’t eat lunch at school and begs off dinner, as well, something highly unusual for him. He gets angry when you offer him the Milk of Magnesia. He retreats to his room without drinking it, and you return to your ballgame. You think nothing more of it until he wakes you in the middle of the night.
“My stomach hurts,” he says, with tears in his eyes.
You give him a hug. He doesn’t feel feverish, but neither is he consoled by your show of affection. Your wife rolls over and asks, “What’s the matter?”
“My stomach—it really hurts! It’s worse than ever,” he says.
You’ve heard this before and once even rushed him to the emergency room, where the tests revealed nothing amiss.
“Come on,” you say. “Let’s try the Milk of Magnesia. It doesn’t hurt to try.”
This is followed with the usual whining but, in the end, he drinks down an ounce of the stuff. You decide to sit with him for a while, and when you push on his belly, it hurts all across his lower abdomen. Eventually he falls to sleep, but in the morning he’s no better. In fact, he’s worse. You check his temperature. It’s normal. He throws up a bit of the laxative. Your wife calls the pediatrician’s office and is advised to take him to the ER. “We can’t do anything for him here,” the nurse says.
The department doesn’t look busy when you arrive, but it still takes twenty minutes to get him registered and another thirty before they call him to the back.
The physician is a middle-aged woman who looks a lot like a teacher you never liked back in grammar school. You’d prefer a male doctor, but in the ER you take whoever is on duty. She greets you with a warm handshake and leans against the counter. There is no place for her to sit. You tell her the story, including the fact that he has had this kind of pain multiple times in the past. Your son now has a low-grade fever of 99.7, and he is very tender in the right lower quadrant of his abdomen. When she releases her hand from his belly, your son winces.
“He has rebound tenderness,” she says. “And the story is good for appendicitis.”
You’re a little shocked. You expected her to say something else. At worst you thought she might suggest an enema.
“Are you sure? The last time we came they said the same thing, but all his tests were normal and he was fine by the next day.”
“Yes, that can happen,” she replies. “But I’m pretty certain that’s not what would happen this time. By tomorrow, it’s more likely that his appendix will be ruptured if we don’t take it out. Still, let’s run some blood work and a urinalysis, and see what turns up.”
She leaves and a few minutes later one of the nurses returns with a handful of tubes. Your son starts crying—he doesn’t want the needle. You tell him it has to be done. The nurse is excellent, and not only starts the IV with a single stick, but is also able to draw off the blood through the IV catheter before hooking it up to a bag of saline. This makes you feel better. It increases your trust that the staff is competent. You believe the blood work will yield a definitive answer as to whether or not your son has appendicitis.
A short while later, the same nurse returns with medication for pain and nausea. This works wonders, and your son is soon asleep. You believe this means that he doesn’t have appendicitis after all. You call your wife and tell her everything is okay.
About forty-five minutes later, the doctor returns.
“Well, the blood work didn’t help us much,” she says. “His white blood cell count is just a little above normal, which doesn’t rule in, or rule out, the possibility of appendicitis. But I still think he has it.”
She re-examines your son, and he winces through his sleep when she presses against his right side.
“You see. He’s still tender,” she says.
“So now what?” you ask.
“He needs to go to the pediatric hospital.”
Suddenly, the thought of what all of this is going to cost flashes through your mind. It’s not that you don’t want the best for your son, but you can’t afford another trip down Goose Chase Lane.
“Look, doc,” you say. “I don’t have any insurance. Do you have any idea how much all this is going to cost?”
She admits to not knowing, but guesses a ballpark figure of “twenty thousand dollars,” before adding, in a tone which sounds more like a threat: “But if the appendix ruptures and he develops peritonitis, it can cost a lot more.” The doctor is advising you to spend all the money you have in the world based on what sounds like nothing more than an educated guess. This doesn’t seem right.
“Before you go calling for an ambulance, isn’t there some other test you can do to be sure?” you ask.
She says there are two options: a CT scan or an ultrasound. Unfortunately, at the adult hospital where you are now, the radiologists don’t have much experience with pediatric ultrasound interpretation. She says they’re better at CT. The subject of radiation exposure doesn’t come up. She explains that even if the scan is normal she still wants to send your son downtown, but you trust the technology more than her opinion. If the scan is normal, you’re taking your son home.
“Let’s get the scan,” you say.
The study requires a dose of oral contrast. Your son drinks it reluctantly. Then you wait for another hour, allowing time for the contrast to line the bowel before completing the scan. Two hours after your initial conversation, the report comes back demonstrating: “a plump, swollen appendix consistent with acute appendicitis.” Your son receives additional pain medicine and is transferred by ambulance to the children’s hospital, where he undergoes a laparoscopic appendectomy later that afternoon. He is discharged uneventfully the next day.
The various bills come out to more than $52,000. To cover the cost, you take a second job at the local Home Depot working weekends and evenings. Your son is well, but your savings are bankrupt. You tell your Canadian cousin this and he laughs, noting that when he had his appendix out four years ago the only cost he incurred was the hospital parking.
Lastly, let’s mull a scenario in which you are the patient. Assume that you’re a thirty-year-old man who owns a landscaping business. You have health insurance. After a hot, hard day cutting grass, mulching beds, and pruning the gardens of the well-heeled, you stop off at the local watering hole for a few beers. While you’re there, it starts to rain.
You live two exits down the interstate and don’t feel the slightest bit impaired when you climb behind the wheel of your truck to head for home. You pick up a burger and fries from the local drive-through. It’s twilight when you pull onto the interstate, the rain barely a drizzle.
Suddenly the truck hits an oil slick lifted off the highway by the rain. The truck spins. You instinctively hit the brakes. The front tire blows and your truck careens off the interstate down an embankment, coming to rest with the driver’s side down in a ditch. The airbag deploys and smacks you in the face. Your seat belt pulls tight, but not tight enough to prevent your left shoulder from slamming into the door or your head from crashing through the window. You are dazed but conscious.
Soon you are aware of two discordant sensations: one cold, the other hot. The first is from the Pepsi spilled over your lap; the second, from the warm flow of blood streaming down your face. Surprisingly, the burger is still safely cradled in your right hand. After releasing the seatbelt, you begin the arduous process of extricating yourself from the vehicle. The truck is a mess. With all the adrenaline flowing, you don’t realize how much your shoulder hurts until you’ve made it up the embankment. Now, sitting on a guardrail, the pain starts to set in. Although your head is cut, your thinking is startlingly clear. You even remembered to snag the Tic Tacs from the glove compartment. Now, using your good arm, you pop a half-dozen into your mouth. The last thing you need is a DUI, but when the trooper arrives, he seems more interested in surveying the wreckage than in giving you a ticket.
The paramedics arrive minutes later and immediately place a hard collar around your neck before securing you to a backboard. To stanch the flow of blood, they wrap a roll of gauze around your head like a turban before checking your chest, abdomen, arms, and legs. While one medic cuts off your T-shirt and places your left arm into a sling, the other starts an IV in your right arm. The entire process takes fifteen minutes.
To your great surprise, you are met in the emergency room by an entire team of healthcare workers fully gloved and gowned in surgical garb. They descend on you like vultures. The doctor at the head of the gurney tells you that a lot of things are going to be happening all at once, but that they’re all “routine.” Your remaining clothes are cut off. A second IV is started, the pain interrupted by an even sharper stick in your right groin, where a medical student has poked you with an enormous needle to obtain blood from your femoral vein. Secured to the board with the collar on, you can’t see what is happening down below, but your attention is fully aroused when a gloved hand grabs your penis and begins lathering the tip with a swab of cold liquid soap. A woman’s voice tells you that she is about to place a catheter into your bladder, and then before you can protest, she shoves the sucker in.
“Shit!” you cry, but no one seems to notice. They’ve heard it all before and are intent on what they are doing. Even after your bladder empties a prodigious amount of urine, you still feel like you have to pee, although the doctor assures you that you’re “good to go.”
Portable x-rays are taken of your chest and pelvis before the chief surgical resident squirts a generous amount of clear jelly across your abdomen. Everything is cold—from the backboard to the iodine soap to the belly jelly—and you begin to shiver, but no one offers you a blanket. Instead, their eyes are focused on the monitor transmitting the ultrasound images from the probe the doctor is running across your abdomen. He tells you through his mask that he doesn’t see any evidence of internal bleeding, but is going to order a series of CT scans to rule out the possibility of a hidden injury. The ultrasound was done only to determine if you needed to go immediately to surgery.
You overhear him say to one of the nurses: “Well, he’s drunk, so we’ll need the full series—head, neck, chest, abdomen and pelvis.”
Then a staff person arrives to take down your demographic information. Without further explanation, she asks you to sign a “consent for treatment” form. No doctor has explained the risks and benefits of anything. After slurping down the CT contrast, you are off to CT Land. The scans take only a couple of minutes. You return to the trauma bay where an intern preps your temple with more liquid soap before injecting you with a local anesthetic. Seconds later your wound is closed with a series of staples. Ba-da-bing, ba-da-boom—all done!
The scans are normal, with the exception of a broken left collarbone. They place your left shoulder into an immobilizer and give you an injection of Dilaudid for pain. The team elects to admit you overnight. After observing you through the next morning, you are discharged to home. There are no complications.
You take the next day off from work to nurse your injuries with Budweiser and Percocet. The pain is manageable, and you’re back in the riding mower before the week is out. Your business hardly misses a beat. The headaches resolve after a day or two. You go back to the trauma clinic in a week, and they pop out your staples. Everything is copacetic.
The bill is outrageous, but the insurance covers everything except the ER co-pay. You don’t even get charged with drunk driving. Three weeks later you buy yourself a new truck with the insurance money. You can’t help but think how lucky you are and what great care you received—all except for the catheter.
In October of 2009, the FDA issued a warning about CT-associated radiation risks after more than two hundred California patients received excess radiation during their scans—some more than eight times the recommended dose—resulting in hair loss and skin reddening. The FDA recommended that facilities audit their results, review radiation-dosing schedules, implement quality control protocols, perform routine surveillance, and adjust the dosages given for patients receiving more than one scan. This was followed a short time later by the release of a study that found huge variances in the amount of radiation exposure associated with the CT scanners at four San Francisco institutions; on average there was a thirteen-fold difference between the highest and lowest amount of radiation delivered. Other authors, looking at risk models based on the National Research Council’s report on the “Biological Effects of Ionizing Radiation,” have projected an additional 29,000 future cancers attributable to the CT scans completed just during the year 2007.
The first prospective study of cancer risk from medical imaging wasn’t published until 2011. Prior to this, all the studies concerning radiation risk had been retrospective reviews subject to various forms of bias. Here, the authors found that during a five-year follow up, patients had a three percent increased risk of cancer for every ten mSv of radiation exposure—less than that associated with a single CT scan of the abdomen and pelvis, or CT angiogram of the chest. Extrapolate these sorts of number over several decades and the real toll of CT-associated radiation becomes frightfully apparent. The problem facing emergency physicians is one of how to best balance a patient’s need for diagnostic accuracy today versus the possibility of a radiation-induced cancer tomorrow.
With this dilemma in mind, let’s return to our three clinical scenarios to sort out whether the CT scans were really necessary. In the first scenario, just how likely is it that the toddler in question has an intracranial injury? Accounting for more than 7,400 deaths a year, pediatric head injury remains the leading cause of trauma mortality in children. The vast majority of head injuries are non-fatal, but still result in more than 600,000 annual trips to the emergency room. How many of these patients need a CT scan?
Emergency rooms currently scan half of the head-injured children they evaluate. The problem is not in identifying the high-risk patients—they are clinically obvious—but rather in identifying those who appear normal, or close to normal, while actually harboring a life-threatening injury. They are precisely the ones with the most to gain from early CT scanning but who are also the hardest to pick out from the crowd.
Evidence-based medicine, as defined by Dr. David Sackett, an expert in data analysis, is the “conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients.” Unfortunately, the use of best evidence remains an interpretive art and, in this case, has resulted in a number of different guidelines for physicians who must rely on the presence or absence of physical findings to guide their decision-making process.
For a clinical finding to be highly sensitive, there must be very few cases where the finding is absent when the disease is present (i.e. the “false negative” rate is low). On the other hand, if a clinical finding is highly specific then it is rarely encountered when the disease is absent (i.e. there are few “false positives”).
Consider the clinical finding of coma following a blow to the head. Since coma is virtually never present in the absence of a serious head injury, it is highly specific for head injury, but because many patients with serious head injuries are not comatose, it is not sensitive. Thus, while the finding of coma helps to rule in the patient with serious head injury, its absence fails to rule it out.
Now, consider the presence of a scalp hematoma, otherwise known as a “goose egg.” The vast majority of kids with a serious head injury have a goose egg, so the presence of a scalp hematoma is highly sensitive as a marker for underlying head injury. It is rarely absent when a serious head injury is present. But as most kids with goose eggs don’t have a serious head injury, the finding is very non-specific. There are lots of false positives. Thus, while the finding of a hematoma does not rule in the presence of a serious head injury, its absence helps to rule it out.
In the case of coma, you have a finding that is highly specific but not sensitive, whereas for goose eggs the reverse is true. And this is where it gets dicey because most kids presenting to the emergency room have a goose egg but aren’t comatose, meaning that the majority have clinical findings which are neither sensitive nor specific for the disease in question—namely, the presence of a serious head injury. So how do you decide which child to scan?
Unfortunately, there is no single clinical finding such that, when the finding is present, the child always merits a scan, and, when the finding is absent, the child is at low enough risk to forego the study, which has led researchers to examine whether findings in combination may be used to determine the need for CT scanning. What if none of the clinical findings individually associated with serious head injury are present? Is it okay to forego CT scanning in the child who only vomits once? How about twice?
Over the last decade at least eight different sets of clinical decision-making guidelines have appeared in the medical literature, which tells you there are no easy answers. One of the largest and best of the studies was published in the Lancet in 2009, enrolling more than 42,000 head-injured children. Researchers first looked retrospectively at a set of patients in order to derive a decision instrument, and then applied the protocol to a second set of patients. The authors concluded that CT scanning could safely be avoided in children less than two years of age provided that none of the following were present:
- A severe mechanism of injury
- A loss of consciousness greater than five seconds
- An obvious skull fracture
- Abnormal behavior
- An altered level of alertness
- A hematoma (“goose egg”) anywhere other than the forehead
Unfortunately, because these findings are so broad and open to interpretation (for instance, parents and doctors won’t necessarily agree on what constitutes a minor versus a severe mechanism of injury), it seems unlikely that their widespread application will result in a significant decrease in the number of CT scans ordered.
What about using skull x-rays as a screening tool, and then obtaining CT scans only in the kids with fractures? Unfortunately, this decision instrument doesn’t work either, as 30 to 50 percent of children with significant brain injuries don’t have concomitant skull fractures.
It is interesting to note that in the Lancet study, a mere 0.14 percent of the head-injured children underwent neurosurgical intervention, while other studies looking at pediatric radiation exposure have estimated that the lifetime risk of a fatal cancer based on a single childhood CT scan of the head is on the order of 0.02-0.1 percent. In other words: the likelihood of a CT scan discovering an injury requiring neurosurgical intervention is almost the same as the chance that the scan itself will someday induce a fatal cancer. I’d say that’s a pretty high price for reassurance.
For the record, I’m not suggesting that seriously injured children forego head CT scanning. In 2010, my fifteen-year-old son fell from an overhanging rooftop and splattered open the back of his head. He was transiently dazed but never lost consciousness. He also broke his great toe and sustained a number of “road rash” abrasions. As I sewed up his scalp, which looked like a piece of pounded flank steak, I wanted the reassurance of a normal CT scan. But I also didn’t want to expose him to radiation unnecessarily, so I had my partner look at him, as well. In the end, we decided not to order the scan. My son soon recovered, but my wife was a nervous wreck for days. If I had seen this type of injury in someone else’s child, I would have almost certainly proceeded with the scan if the parents weren’t vocal about skipping it.
Now let’s turn our attention to the second scenario involving the twelve-year-old boy with appendicitis. The results of a questionnaire published in 2011 clearly demonstrated that when it comes to abdominal pain, patients are far more confident in their doctor’s diagnostic acumen when backed by a CT scan. Patients rated their confidence in the doctor’s diagnosis when based on the history and physical alone at just twenty on a scale of one to one hundred where zero represented no confidence and one hundred complete confidence, but the score soared to ninety when the evaluation also included blood work and a CT scan. Not surprisingly, the ER evaluation of abdominal pain now includes CT scanning in a quarter of patients, up from one in ten just a few years ago. But does a CT scan really help in making the diagnosis of appendicitis?
In the pre-scan era, the decision to operate was based largely on the presence of exam findings in combination with an elevated white blood cell count. Although not perfect, the clinical acumen of an experienced examiner is pretty darn good. But since we now live in an age where technological certainty has largely supplanted clinical gestalt, a pre-op CT scan is de rigueur before proceeding to operate—at least in adults. Nonetheless, misadventures in the diagnosis of appendicitis continue to be the leading cause of malpractice payouts for older children and pre-adolescents. Why is this diagnosis so difficult?
First of all, the problem is common. The lifetime incidence of appendicitis is 7 percent. Second, the “classic” presentation occurs only half of the time in pediatric patients, and the younger the child, the more likely it is that the presentation will be atypical. Third, the disease still carries significant risk. Before the advent of surgery, the mortality rate for appendicitis was roughly 50 percent. And while today’s mortality for simple appendicitis is a mere 0.1 percent, this number increases to 5 percent when the appendix ruptures. Despite the best efforts of pediatricians, emergency physicians, and surgeons, the rate of ruptured appendicitis remains 15 percent or more in adolescents and higher still in younger children. The earlier the surgeon intervenes the better, but there is a downside to early exploration. About 20 percent of the time, the appendix is “normal” and surgical exploration for false positive appendicitis carries with it a one-in-714 chance of death from anesthesia and surgical complications.
This combination of commonality, diagnostic difficulty, and high complication rate explains why doctors continue to seek a test of certitude before operating. The goal of CT scanning is to enable early surgical intervention in those who have the disease, while decreasing the number of operations in those who don’t.
What about ultrasound, which involves no radiation? On the down side, the procedure is operator-dependent (meaning that the quality of the images depends upon the skill of the technician) and not as widely available during non-business hours. Additionally, the study isn’t as accurate in obese children, where the presence of adipose tissue limits the ability to image the appendix. Drawbacks aside, the diagnostic accuracy of ultrasound rivals that of CT and, therefore, should be the screening modality of choice for suspected pediatric appendicitis.
In the adult hospital where I work, however, the radiologists and surgeons have a clear bias for CT. In general, you won’t be offered a choice. Unless you take your child to a pediatric hospital, the emergency physician will most likely order a CT scan. In our case scenario, the father was concerned about the expense of an unnecessary surgery, and this provided justification for the scan. But was it medically necessary?
Based on his age, the patient’s lifetime risk of developing a radiation-induced cancer from the scan is on the order of one in five hundred, exceeding the mortality rate of a negative appendectomy (one in 714). Since clinical scoring systems already rate his risk of appendicitis as high, the CT adds little and is not justified.
In a study of nearly two hundred children with abdominal pain published in 2005, the addition of a pre-operative CT scan failed to reduce either the number of negative appendectomies or the number of children with perforation at the time of surgery. In fact, the number of negative surgeries in the CT scan group was actually higher due to the tendency of radiologists to “overcall” appendicitis in iffy cases. In another study, CT scanning delayed the time to surgery by more than three hours and resulted in a near doubling of the perforation rate. Given that there is little benefit and clear potential harm, why are so many kids with abdominal pain still being irradiated? It goes back to our love of technology and the need for perfection. It’s far easier for surgeons and emergency physicians to defend their actions when they’re based on the results of a CT scan rather than an educated opinion. If your child has suspected appendicitis and the physician feels additional studies are needed to confirm the diagnosis, request an ultrasound. Or better yet, demand that the surgeon examine your child before proceeding.
Finally, let’s examine the third scenario involving the drunk driver with a scalp laceration and broken collarbone. Motor vehicle crashes remain the leading cause of death for every age between three and thirty-three years of life. Most urban areas are now served by Level One trauma centers, and if you are seriously hurt in a car crash, these are the people you want taking care of you.
Although sometimes derided as “cookbook medicine” because of its heavy reliance on protocols, emergency trauma care in the United States is second to none. The downside to protocol-driven medicine, however, is that it over-tests the many in order not to under-test the few. Critically injured patients need the scans; minimally injured folks don’t. In between are the majority of trauma patients, who arrive with a mechanism of injury suggesting the possibility of serious harm, but who otherwise appear okay. These are the patients where CT scanning has, all too often, come to supplant, rather than augment, the physical exam.
In our trauma patient, the application of clinical prediction instruments indicates that his likelihood of survival is very high—on the order of 98 percent. Still, the trauma team behaved in a perfectly perfunctory manner in obtaining a series of x-rays and performing a bedside ultrasound exam to look for evidence of internal bleeding. In this case, the screening tests were negative, suggesting a margin of safety before proceeding.
Given that our patient was intoxicated and had a distracting injury in the form of a clavicle fracture, the decision to scan the neck is appropriate. Unfortunately, once that decision was made, the inclusion of the head followed reflexively, but is still not unreasonable. The decision to include scans of his chest, abdomen, and pelvis, however, are clearly problematic. Here, the justification was based solely on the mechanism of injury without regard to his clinical examination.
The controversy over the broad application of CT scanning based on mechanism of injury intensified after the publication of a trial from one of the nation’s busiest trauma centers, L.A. County, in 2006. The authors reported on a thousand trauma victims who were “pan scanned” (head, cervical spine, chest, abdomen, and pelvis), despite stable vital signs and the absence of any outward visual signs of chest or abdominal trauma. They found CT abnormalities in 20 percent of the patients. The injuries were frequently serious and included cervical spine fractures, rib fractures, collapsed lungs, and traumatic injuries to the liver, spleen, kidney, and bowel. A significant number of the patients were intoxicated and/or had unreliable exams due to head injury. In their discussion, the authors note that: “Some would argue that a CT scan has replaced physical examination in trauma.” They then use their findings to support this practice.
Remind me to drive carefully in Los Angeles.
I guess if you aren’t going to bother examining the patient, then you should do something. Just because a patient doesn’t have a rib sticking out of his chest doesn’t mean he’s not hurt. A simple physical exam would have uncovered the vast majority of the injuries noted in the study. I have yet to encounter a non-comatose patient with a rib fracture who didn’t say “Ouch!” when I touched it. Similarly, if a patient’s neck is broken, it hurts—a seemingly obvious notion that required a large multi-center trial to validate as true. With regard to hidden abdominal injuries, a different study of more than four hundred trauma victims found that simply examining the belly, and asking, “Does it hurt?” correctly identified all those who ultimately required surgery.
In the L.A. County study, patients merely had to lack “visible signs” of trauma to the chest and abdomen for inclusion. I view this trend toward pan scanning based on the mechanism of injury alone to be reckless and irresponsible, not to mention more than a little bit lazy. Unfortunately, when universal protocols are adopted, this is what you get. Five years after the initial study, a second study from L.A. County appeared in the Journal of Trauma, noting that a doubling of the radiation dose per trauma patient between 2002 and 2007 “had no impact on the missed injury rate, length of stay, or mortality.” Similar American reviews from Johns Hopkins and Case Western and an Australian study from Sydney likewise failed to note any improvement in diagnosis or outcome associated with increased CT scanning.
So, despite plenty of data refuting the notion that CT scanning is the panacea for all diagnostic dilemmas, its use continues to rise. Contributing factors include patient expectations, lack of confidence in the physical exam, the need for perfection, financial incentives, physician fear of litigation, and widespread ignorance as to the risk. More than 90 percent of doctors admit to practicing defensive medicine at least some of the time—and the other 10 percent are simply in denial. One of the ironies of our growing dependence on technology is that the more we use it, the more we rely on it at the expense of our own examination skills.
Consider too, that bad outcomes happen in real time, whereas radiation-induced cancers are likely to occur decades down the road. Victims of the Hiroshima and Nagasaki bombings exposed to low doses of radiation didn’t necessarily develop cancers at an earlier age, but developed more cancers over their lifetimes. By the time our head-injured toddler develops her brain tumor at age fifty, the physician who ordered the study would have long since retired, but if she has a poor outcome tomorrow due to a missed brain hemorrhage then the likelihood of a near-term lawsuit is extremely high. In a recent survey, nearly 50 percent of parents expected their child to receive a CT scan as part of their evaluation after a head injury, so from a medicolegal standpoint, it makes no sense not to order the scan.
In short, there are no incentives for doctors to curb their use of CT scans other than the foreknowledge that the more scans we order today, the more cancers we will face in the future. Unfortunately, most doctors remain blissfully unaware of the risks. A survey published in 2004 revealed that fewer than half the radiologists and only one in eleven emergency physicians believed there was an increased cancer risk associated with CT scanning. Although risk awareness is certainly greater today, this hasn’t done a thing to lower the number of scans being ordered. Experts project that up to 2 percent of all future fatal cancers will occur as a result of CT scan-related radiation. Now, the next time you go to an emergency room are you sure you want the doctor to order a scan just to be safe?
As MRI technology (which involves no radiation) advances, I expect this concern to become moot, but that doesn’t mean that the problems associated with the expanded use of technology will go away. As our ability to find minute pathology increases, we will confront growing concerns over what to do with that information, and how much money we need spend chasing down abnormal test results in the absence of clinical disease. In the meantime, we all need to do a better job. Radiologists are the experts here and, frankly, have performed poorly in getting the word out. The government also has a role to play and needs to institute universal emission standards and controls. As physicians, we have an obligation to use technology responsibly, not reflexively. Lastly, you—the patient—need to stop pushing your doctors to continually do more and more. Life involves risk, and very few conditions require an emergency CT scan of anything. Be aware that what you want may not be what you need.