These days you cannot leave a radiology meeting without hearing that radiologists must prove their value. If bold enough to discuss “what is value” over cocktails, you will likely end up in that intractable conversation that college students have when they discuss “free will.” Words will be thrown at you, such as “value equation,” “value proposition” and “outcomes.” You will be left mystified and befuddled.
Despair not. Actually, value of imaging is simple to understand. It is more challenging quantifying value. But in order to quantify value it is important to qualify it. Numbers are meaningless without context and a narrative. Let’s start with a conceptual framework to understand the value of imaging.
Imagine a pill, Clotbuster, which dissolves clot and prevents recurrent thrombus. Clotbuster costs nothing to administer and has no side effects whatsoever. It can be doled out liberally to anyone with even the lowest suspicion of pulmonary embolism (PE). What will Clotbuster do to the value of computed tomography for pulmonary embolism (CTPE)?
CTPE has value because PE is worth diagnosing. PE is worth diagnosing because there are consequences if undiagnosed, notably recurrence of embolism. Recurrence is problematic because the next embolus can be so large so as to cause cardiovascular collapse and death. The value of CTPE lies in the harms of untreated PE. These harms are necessary but not sufficient to give CTPE its value.
Treatment for PE, either thrombolytics for certain situations or anticoagulation for prophylaxis, has complications, such as bleeding. The patient can have a fatal intracranial hemorrhage. Unlike our fictional wonder drug, Clotbuster, anticoagulants can’t be handed out like candy. Imagine if someone who doesn’t have a PE develops intracranial hemorrhage from anticoagulation. That’s two wrongs – harming and harming unnecessarily.
Thus, there is value in discernment. That is there is value in discerning between those who have PE and those who don’t. Discernment has value because treatment for PE can harm. This is an odd calculus between CTPE and anticoagulation, where the value of CTPE is related to the harms of anticoagulation.
The value of CTPE depends on:
a) Harms of untreated PE.
b) Harms of treatment for PE.
c) Ability to distinguish between those with PE and those without PE.
The framework is common sense and you might wonder why it needs any exposition. This needs an explanation for three reasons. First, when we think of value of imaging we should think of the harms avoided as well as gains incurred. This sounds like a clever frame, but arithmetically harms avoided, such as bleeding complications, are relatively easy to measure. Second, the cost of treatment, monetary and physical, determines the value of discernment. Third, the value of imaging must periodically be reassessed and a simple way to reassess value is to revisit the safety of treatment.
Clotbuster dramatically reduces the value of CTPE, because it lowers the stakes of incorrectly treating PE. But don’t worry, if big pharma ever discovered a drug with seamless action and no side effects, you can bet that the price will be so high that CTPE will still have monetary value in discerning.
Sticking to the example of pulmonary embolism, CTPE has value because we do not know for certain who does and who does not have PE by history, physical examination and basic laboratory tests. Not only do we not know who has PE, there is a broad spectrum of our not knowing. What do I mean by that?
Imagine you are arguing with your spouse whether to watch Gone with the Wind or The Usual Suspects. You agree to a coin toss and that if the coin lands on heads you watch The Usual Suspects. You don’t know whether the coin will land on heads or tails but you know that the chances of heads are 50 %. This gives you hope, even if you are still left with the prospect of watching a long, boring movie.
If the chances of heads extend from 25 % to 75 %, you are a little more anxious. But what if you don’t know the chances that the coin will land on heads? I suspect you might want a different method to decide which movie to watch.
In a similar vein, sometimes clinical factors give a reasonable estimate of the probability that a patient has PE. Sometimes, one doesn’t know what the chances are that a patient has a PE. CTPE has value because it converts not knowing in to knowing. CTPE has even more value when we do not know how much we do not know that a patient has a PE. Our ignorance is not homogenous.
The indications for prophylactic statins are expanding. Let’s say in the future that a panel of the wise known as “The Guardians of the Myocardium” decrees that once people turn thirty five they should be on a statin.
Will plaque imaging, such as calcium scoring CT, have any value? Remarkably, yes, although admittedly not a lot. People are individuals. Individuals have preferences. Preferences are not the same – remember, people are individuals. Many will gladly subscribe to statins. Some may worry about the low chances of muscle pain, a side effect of statins. Some may not wish to be conscripted to a pill so early in their lives unless they must, and could decide if they must after the results of the calcium scan.
When there are preferences, there are trade-offs, and vice versa. Were statins an unmitigated good what would be left to prefer? If there are no trade-offs, there are no choices.
Ingredients of Value
Value of imaging lies in the presence of the following.
1. Harms of untreated disease.
2. Harms of treatment.
3. Uncertainty that the disease is truly there.
Of course, value is not limited to these elements. Nor is it presently necessary to argue about the relative contribution of these elements. What is important to appreciate is that value of imaging lies in the fact that clinical medicine is imperfect and imprecise.
If there is a costless anticoagulant without side effects, imaging for PE will be useless.
If a decision rule confers PE with the same accuracy as imaging, there will be little value to CTPE.
If everyone must take statins and everyone feels equally enthusiastic about statins, plaque imaging adds little information.
Click Here for downloadable lecture formats