Trickle-Down Economics of Medical Imaging

Trickle-Down Economics of Medical Imaging

Saurabh Jha, MBBS, MRCS

February 05, 2015

Medical Imaging and the Price of Corn

After the Napoleonic wars, the price of corn in England became unaffordable. The landowners were blamed for the high price, which some believed was a result of the unreasonably high rents for farm land. Economist David Ricardo disagreed.

According to Ricardo, detractors had the directionality wrong. It was the scarcity of corn (the high demand relative to its supply) that induced demand for the most fertile land. That is, the rent did not increase the price of corn. The demand for corn raised the rent. Rent was a derived demand.

Directionality is important. Getting directionality wrong means crediting the rooster for sunrise and blaming umbrellas for thunderstorms. It also means that focusing on medical imaging will not touch healthcare costs if factors more upstream are at play.

Medical imaging is a derived demand. The demand for healthcare induces demand for imaging. Demand is assured by the unmoored extent to which we go for marginal increases in survival.

The Demand for Imaging in Stroke

The treatment of ischemic stroke using thrombolytics and intra-arterial therapy (IAT) is instructive on how imaging can be induced. Lytics improve outcomes but must be administered relatively rapidly after onset of symptoms. IAT, which is treatment at the site of arterial blockage, allows the clock to tick for a bit longer and has recently been shown to be beneficial.

In the MR CLEAN[1] study, patients with acute ischemic stroke with radiographically proven occlusion in the proximal anterior circulation were randomly assigned to IAT or usual care. IAT included local thrombolysis, mechanical break up of thrombus, or stent placement. In both treatment and control groups, most patients also received thrombolysis with alteplase (Activase®).

The results published in the New England Journal of Medicine [1] showed that patients who received IAT within 6 hours of stroke onset had a clinically significant increase in functional independence at 3 months without higher mortality.

Counting the CTs and CT Angiograms

However, patients don’t walk into the emergency department saying, “Good evening, doctor, I have an occlusion in the proximal internal carotid artery. Can we get moving, please?” Sometimes they have classic signs and symptoms of stroke in the distribution of the target artery. Often they have more vague signs and symptoms that could be due to other causes.

Thrombolytics aren’t candy that you distribute willy-nilly. They have serious side effects such as bleeding in the brain. They can worsen the neurologic deficit. They can kill, even when used correctly.

Patients first need a CT of the head—not to pick up early stroke; that’s a clinical diagnosis. A CT is done to make sure that the stroke is not caused by a bleed. Hemorrhagic strokes are less common than ischemic strokes but common enough to warrant routine use of CT for this distinction.

The other reason for CT is to ensure that if the stroke is indeed ischemic, there is not so much brain edema that thrombolysis will lead to bleeding. The risk for bleeding in an ischemic stroke increases in edematous brains.

For IAT, we need to know who has a treatable lesion. Signs and symptoms don’t tell us with certainty who has a culprit lesion in the proximal internal carotid arteries, and neither does standard CT of the head. Patients need CT angiogram not only to identify those who have a treatable blockage but also to find out who doesn’t have a treatable lesion. We can’t tell this without doing a CT angiogram.

Obviously, this means that more patients will get a CT angiogram than are eligible for IAT. Many more will get a CT angiogram than will actually benefit from IAT. (Not everyone deemed to benefit from IAT will actually benefit.)

Okay, so CT of the head is done in all, and CT angiograms are done in many. Is it time to check out with the radiology cashier? No. We are far from done with imaging.

Imaging for Complication Monitoring

After thrombolysis, patients will have another CT of the head to make sure that the clot buster did not cause a bleed. They will be managed in the neuro-intensive care unit (ICU). The level of care will be higher. The sensitivity to mild variation in objective ICU parameters will be intense. Surveillance CTs of the head will be frequent.

There is a plausible risk for bleeding elsewhere after thrombolysis, such as in the abdomen. How do we know whether the patient has bled? Signs and symptoms aren’t reliable. Remember, these patients have neurologic deficits, which bar some from verbalizing clearly.

A fall in hemoglobin and hematocrit could herald a bleed, even if more often than not these indices fall because of the fluid status. Patients in whom these indices have dropped may need a CT of the abdomen to look for bleeding. The cheaper noncontrast CT will suffice in most but not all patients. Some may have active bleeding warranting interventional radiology, which requires a CT angiogram for diagnosis.

IAT involves a puncture of the femoral artery and then navigation of a catheter through the aorta, which is not without complications such as seroma, which may resolve spontaneously. Less commonly, patients might develop a false aneurysm of the femoral artery or an arteriovenous fistula. Patients need an ultrasound and/or CT or MR angiogram of the femoral arteries when there is suspicion of dangerous complications.

The aorta and carotid arteries are tricky territories to navigate. The surgeon might tear the arterial wall. The tear might propagate. These patients need a CT angiogram of the chest and neck. The surgeon might feel that he or she was a bit too enthusiastic breaking up the clot and fear that the artery has been torn. These patients will get a CT angiogram as well. Better safe than sorry.

Broken thrombus could embolize and infarct new areas of the brain, causing new neurologic deficits. In those instances, an MRI of the brain, a CT angiogram of the head, or, more likely, both may be ordered. Of note, in the treatment arm of the MR CLEAN trial, new strokes were reported in 13 of 233 participants.

We are not yet done with imaging.

The CT dye load and natural variation may elevate the creatinine, prompting an ultrasound of the kidneys to exclude a postobstructive, and treatable, renal impairment. Patients will receive countless chest and abdominal x-rays when support lines are inserted and manipulated, in accordance with principles of safety.

A few CTs of the chest may be ordered because a radiologist could not “exclude pneumonia with absolute certainty” in someone with left partial lung collapse, which everyone in the hospital has, particularly when an intern puts in the clinical information, “new fever, exclude pneumonia, kindly.”

Cost and Quality Considerations

Stroke doesn’t sleep at night, and neither do its costs. Radiology departments must be staffed to provide timely access and reports. CT angiograms are complex studies. Trained radiologists, and perhaps expert 3D technologists, must be available 24/7. The window for IAT is 6 hours from stroke onset—the urgency at 5 hours, 30 minutes will be intense.

The medical imaging consequent to applying thrombolysis and/or IAT for stroke in practice is substantial. Any other industry would proudly showcase this exemplar of trickle-down economics. However, patients don’t pay large marginal costs for small marginal benefit. The costs are diffused. But costs are costs. Healthcare is neither a free market nor a market that is free.

How should payers reimburse the derived demand for imaging? They can say: “Stroke team, here is how many quality-adjusted life years (QALYs) the population gained. Thanks for your contribution, radiologists. Here is your share. Yes, I know it isn’t a lot, and your efforts were enormous, but you must admit, the net QALYs are a bit modest.”

That is if the absurdly but rationally high amount of medical imaging as a result of thrombolysis/IAT for stroke were reimbursed in a bundled payment for stroke. This is fine, but is it fair that the average returns on imaging are diminished because of increasing complexity of treatment and diminishing returns on QALYs, factors over which radiologists have little control?

Or should radiologists say to payers, “You want to improve survival in stroke patients using aggressive, expensive, and dangerous therapy. Here is an estimation of the costs of the imaging. Now put your money where your coverage is.”?

Regardless of the payment method, whether bundled pay for outcomes such as QALYs or itemized fee for service, it should be apparent that thrombolysis and IAT for stroke will substantially increase use of imaging. But imaging should not be blamed for increasing costs; that credit should be given to science and humanity for the high costs of stroke treatment.

When clot busting for stroke becomes practice du jour, the population will no longer resemble the trials. The MR CLEAN study was conducted in The Netherlands. This is the United States. Patients will be older and sicker; they will incur more complications, induce more imaging, and derive even less treatment benefit, statistically speaking. Any bundled payment determined by outcomes will become smaller with more and more activity in the bundle. The math of such a payment scheme can’t last for too long.


  1. Berkhemer OA, Fransen PS, Beumer D, et al; MR CLEAN Investigators. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med. 2015;372:11-20. Abstract


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s