Blood Clots and the Death of Common Sense

A couple of days ago I was on national TV being asked the usual round of questions about an NFL player who’d apparently developed blood clots. How can a young, fit athlete get a pulmonary embolism (PE)? What are the symptoms? Can you fly with a clot? Could it have been picked up earlier?

All perfectly sensible questions, especially for a young athlete, who faces far more scrutiny than an ordinary patient would. Within the hour,  the story was all over TV station websites. . Obviously, he’s got the country’s attention.

I’ve spent much of my career answering requests like this type. After doing emergency, pulmonary, and critical care for so long, you’d think you’d know everything about thromboembolic disease by now, and I really don’t find anything too surprising in that field anymore. I’ve seen PE kill patients who were absolutely stable just minutes before. I’ve also seen patients who were absolutely convinced they had a nasty clot, despite their presentation being very different. Both experiences have shaped how I think about the disease and reinforced that medicine needs the human element, something no machine, test, or algorithm can ever provide: Judgment.

The more I thought about that interview, the more I realized it wasn’t really about a footballer or pulmonary embolism. It was an interview about where modern medicine has gone. We have such wonderful diagnostic tools now, and at the same time we have created a situation where doctors are increasingly unwilling to use their own clinical judgment without some form of technological validation.

We can now pick up abnormalities that would have been completely invisible a generation ago, yet we can find ourselves in a position where we’re not sure whether that abnormality is of any significance. We’re absolutely superb at detecting disease, and we’re surprised by how uncertain we can be about when to stop looking.

At one time, a doctor’s most important tool was his own mind. That’s still true today, though I’m not sure we believe it anymore.

An Old Problem with a Modern Twist

About 180 years ago, Virchow examined the mechanisms of thrombosis and pulmonary arterial obstruction. His 1846 work and the papers collected in 1856 formed the pathological basis of what we now call thromboembolic disease [1,2]. Every medical student has heard of him at some stage, invariably in the context of the classic triad of venous stasis, endothelial injury, and hypercoagulability. The triad as we know it wasn’t actually Virchow’s. Historians have examined how later interpretations and reassemblies of his original observations produced the concept that now bears his name [3].

The historical detail is of course interesting, but the real issue is the physiology. Clots form because the biological milieu is pro-thrombotic. Flow may be slow. The vascular endothelium may be damaged or activated. The coagulation system may shift to a pro-thrombotic state. All of these can happen in isolation or in combination, and their relative importance varies from patient to patient.

A post op patient who is bedbound is not in the same physiological situation as a young athlete who has had an injury. A patient with active cancer is not in the same situation as a patient who has simply spent several hours sitting on an aircraft. They may all go on to have a venous thromboembolism (VTE), but the circumstances in which it occurs and the risks involved will be very different.

That’s why you can’t just see a symptom and order a test. You have to understand the patient, the patient’s situation, and the biology behind it. A swollen leg could be venous thrombosis, chronic venous insufficiency, a drug effect, trauma, inflammation, or a hundred other things. Shortness of breath could be a pulmonary embolism; it could be cardiac disease, it could be a pulmonary infection, it could be obstructive airways, it could be anemia, or it could be anxiety. The presence of a symptom is the first thing you consider in the diagnostic process; it is not the final one.

Virchow had no CT scanner, no D-dimer, no modern ultrasound. His was by no means a golden age of medical care. I wouldn’t trade the toolkit we have today for what was available in the 19th century. One thing to keep in mind, though, is the intellectual underpinning of his work. He looked at disease through the lens of observation, physiology, and pathology. All the modern technology we’ve developed has enormously improved our ability to look at the human body, but it hasn’t taken away the need to understand what we’re looking at.

A doctor who can order a test perfectly well but has absolutely no idea why he’s ordering it has missed a fundamental aspect of the job.

The Pulmonary Embolism We Must Not Miss

In 1998, Robert Fromm and I published a very brief paper in the Journal of Emergency Medicine called ‘Syncope: The Forgotten Sign of Pulmonary Embolism’ [4]. We drew attention to a sign that is easily overlooked when assessing a patient who has had a syncopal episode. Most of the time, syncope is not due to PE, but in a small number of cases it can be the initial presentation of a significant obstruction of the pulmonary circulation. Failure to appreciate that possibility can be very dangerous.

I still think we were absolutely right at the time. PE can present in all sorts of ways, and you have to watch for the possibility that a fairly common presentation could be a manifestation of a very serious underlying problem. Anyone who has managed a patient with massive PE will understand what I mean. A large PE can result in acute right ventricular failure, profound hypotension, circulatory collapse, and death. In some situations, minutes really do count, and you simply can’t afford to be slow.

We certainly didn’t suggest that every patient who fainted should have a pulmonary CT angiogram. That would misrepresent the clinical situation. The point is to identify which patients’ presentations are concerning. A fit young adult having a classic vasovagal episode after standing for a long time is a very different proposition to a post op patient who suddenly goes off the rails and falters at the same time as developing unexplained hypoxemia and tachycardia. Both can cause syncope, but the probability of PE differs vastly.

This distinction may seem fundamental, but it’s becoming increasingly important in an age where diagnostic technology can lead clinicians to simply order the test rather than engage in the appropriate reasoning first. We’re frequently in the position of wanting to investigate every single possibility just because we can. The result is not necessarily better medicine. It may be more expensive, more complex, and in some cases, actually worse medicine.

We all have to live with a rather uncomfortable reality as doctors. We will never catch every disease in every patient on the first go. No diagnostic system can be that precise. The aim of clinical medicine is to detect significant disease with enough accuracy to improve outcomes, and to do so without intervention when the potential harm outweighs the potential benefit. That, in itself, requires an appreciation of probability, and it seems to me that appreciation is now less in vogue than the pursuit of absolute certainty.

When a Diagnostic Test Becomes a Screening Test

In 2019, my colleagues and I published an editorial in Current Respiratory Medicine Reviews asking whether computed tomography with angiography (CTA) in PE was used as a diagnostic test or had, in fact, become a screening test [5]. The distinction is more significant than it initially appears.

CT pulmonary angiography is one of modern medicine’s great advances. It has completely transformed how we approach suspected pulmonary embolism, allowing us to rapidly and in great detail visualize the pulmonary arterial circulation. In the right clinical situation, it can provide information that directly affects management and can, in fact, save lives. I’ve used it throughout my career and have no desire to go back to a time when PE was so much harder to diagnose.

The problem is that we order the test simply because it’s available. A patient turns up in the ED with chest discomfort. The HR may be slightly elevated. The patient may have had a recent flight, or may have noticed leg discomfort. None of that is meaningless, and none of it alone is sufficient to establish a diagnosis of PE. You have to assess the history, examination, risk factors, and overall clinical probability.

There’s a very real temptation to order a D-dimer in this scenario. If that’s positive, even for some unrelated reason, the next step is almost certainly CT angiography. You’re off on a diagnostic investigation quite quickly without really knowing how likely the disease was in the first place.

This isn’t some fictitious scenario cooked up by the detractors of modern medicine. Wiener and collaborators examined trends in pulmonary embolism diagnoses in the US and found a substantial increase in cases diagnosed during the period when CT pulmonary angiography was widely available. Mortality increased much less [6].

The point is that more sensitive imaging detected many more emboli than had previously been found. It didn’t show that all the small emboli now being detected were non-pathological. It certainly didn’t suggest that we should stop using imaging in patients who needed it. It did show, however, that an increase in the number of cases diagnosed does not necessarily mean an increase in deaths.

That observation should have prompted a much wider debate about what we actually mean by a successful diagnosis. If we’ve found an abnormality that would never have harmed the patient absolutely, have we actually improved their health? If the finding leads to some form of intervention which itself may be potentially harmful, have we in fact done the patient any good at all? These questions are very difficult to answer, and the answer isn’t always obvious. The whole history of medicine has tended to celebrate disease detection, and we’ve rarely stopped to think about its consequences.

The same thing happens in so many other specialties. You’ll find little thyroid nodules, incidental pulmonary lesions, minor vascular abnormalities, and all sorts of imaging findings of uncertain significance. Some will require further investigation or, indeed, treatment. Many won’t. Once you’ve picked up an abnormality, though, it can take on a life of its own. Follow-up imaging is arranged. A specialist is seen. A procedure is discussed. The patient becomes very anxious (in some cases for years) about a finding which may have had absolutely zero impact on their health.

It’s a rather strange paradox, really. The more sophisticated our diagnostic tools become, the greater the responsibility we have to use them judiciously. Instead, we have tended to respond to increased sensitivity by lowering the testing threshold until investigation becomes routine.

The Fear of Being Wrong

I’d be ignoring a massive element of this issue if I didn’t talk about defensive medicine. Doctors don’t operate in a vacuum. They work within hospitals, healthcare systems, insurance schemes, regulatory frameworks, and a legal environment, all of which affect their behavior. Some of those influences are positive. Others create incentives which have very little to do with the needs of the individual patient.

I’ve been in medicine long enough to know what it’s like to have a clinical decision second-guessed once the outcome is known. The difference between assessing a patient prospectively (when the diagnosis is in doubt) and then reviewing the same patient retrospectively (after an adverse outcome has occurred) is enormous. Once the outcome is known, the warning signs can look a lot more apparent than they did at the time. That’s one reason it can be so misleading to judge a clinical decision purely on the basis of the outcome.

A doctor could well make a sensible decision on the evidence and still end up with a bad outcome. On the other hand, they could order an entirely unnecessary test, subject the patient to unnecessary risk, and still get kudos for being thorough because no harm was done. Neither of those scenarios in isolation tells us anything about the soundness of the original decision.

In 2005, Studdert and co-workers published a JAMA paper examining defensive medicine across specialties known to be high risk for malpractice. Of the physicians who responded, 93% reported practicing at least one form of defensive medicine (unnecessary testing, referrals, etc) [7]. Again, this is self-reported, so the study cannot be taken as evidence that all the extra investigations were, in fact, inappropriate. It does, however, demonstrate the extent to which fear of litigation can affect how doctors go about their work.

I think many doctors will be familiar with that internal dialogue. The clinical probability of a serious illness is low. The history and examination are all reassuring. The evidence we have strongly favors a conservative approach. Then of course comes the thought that completely alters the picture: what if I’m wrong?

There’s absolutely nothing unreasonable in considering that possibility. Good doctors will question themselves. The problem arises when the possibility of being wrong starts to outweigh the probability of being right. A doctor may order a test not because of the patient’s clinical situation, but simply because an abnormal result might later be taken as proof the test was a good idea. That’s a completely different kind of medical reasoning.

Incentive is particularly strong given that the harms of unnecessary testing are often diffuse and/or delayed, whereas the harms of missing a dramatic diagnosis are immediate and very much in the patient’s mind. An unnecessary CT scan may cause no obvious harm. An incidental finding may trigger months of further investigation, but the link between the original test and the subsequent chain of events may not be apparent. A patient may be put on anticoagulation and have bleeding weeks later. The harms are real, just not necessarily as readily apparent as the harm of missing a PE.

We have a system where doing more is far easier to justify than doing what is actually required.

The Patient Pays for Our Uncertainty

It would be reassuring to think that further testing is in fact innocuous even when it’s not indicated at all. That is not the case. Every investigation has potential ramifications, and every diagnosis opens the door to treatment. The more aggressively we search for abnormalities, the more we will locate findings of uncertain significance.

Take the patient with a small pulmonary embolus. Anticoagulation may be clearly indicated in some circumstances. In other circumstances (particularly in certain carefully selected cases of isolated subsegmental PE), the balance between treating and monitoring may require a more individualized approach.

The presence or absence of proximal deep venous thrombosis (DVT), the patient’s cardiopulmonary reserve, the risk of recurrence, the bleeding risk, and the reliability of follow-up all come into play. It would be wrong to suggest that small emboli are inconsequential and that anticoagulation can be safely omitted in all cases. It would also be wrong to pretend that every imaging finding has the same clinical significance.

Anticoagulants have saved many lives. Modern agents are far more convenient than older ones and, in many cases, safer. They are not without risk, though, and bleeding is still a significant complication. The decision to subject a patient to that risk must be one based on a reasonable expectation of benefit.

This is where the difference between spotting an abnormality and treating a patient becomes important. We don’t give anticoagulation to make the CT look better. We give it because we believe that by preventing further thrombosis and embolization we will improve the patient’s outcome. If that benefit is in any doubt, then it’s a decision that needs to be weighed up rather than one that you just go with.

The same applies to so many other interventions. A treatment can be technically successful and still fail to produce any meaningful improvement in the patient’s health. A procedure can correct an anatomical abnormality while exposing the patient to a level of risk greater than the expected benefit. An investigation can give us more information without altering what we should do.

One of medicine’s simplest questions is also one of its most frequently overlooked: what will I do differently when I get the result? If the answer is nothing, then the test in question probably needs a second look. That doesn’t mean the test is inappropriate. There may well be value in the information (in helping to rule out a serious condition for instance) or it may simply be that it will inform future management. The issue is one of purpose rather than of routine.

Patients deserve that. They deserve doctors who realize their bodies aren’t just a series of lab results and radiology reports. They deserve care that considers the benefits and harms of any intervention, including the harms of doing too much.

The Algorithm Is Not the Physician

I have absolutely no problem with clinical decision rules. Indeed, some of the best evidence for more sensible testing comes from developing and validating such rules. The Pulmonary Embolism Rule-out Criteria (PERC) and the YEARS strategy both attempt to apply clinical probability systematically and reduce inappropriate investigations in the right sort of patient [8,9].

These approaches are useful because they acknowledge that not every patient will be managed the same way. They don’t replace clinical judgment; they work with it. PERC, for instance, isn’t a tool you apply to every patient who walks into the ED. Its appropriate use starts with an assessment of low clinical probability. The same is true of a D-dimer threshold. You can’t properly interpret a D-dimer threshold without knowing the diagnostic strategy, the assay, and the patient population that you’re looking at.

There’s a difference between using an algorithm to structure clinical reasoning and allowing it to take over clinical reasoning. The first can improve care. The second can give a false sense of precision. A patient doesn’t get any better or worse because a scoring system assigns them a particular score. The score attempts to quantify risk based on associations observed in a studied population.

It’s useful when used appropriately and can be misleading if used outside the circumstances for which it was developed. The clinician still has to decide whether the patient represents the population in which the rule was validated, whether any important findings fall outside the rule, and whether the proposed diagnostic pathway makes sense.

The 2026 multidisciplinary guideline on the assessment and management of acute pulmonary embolism is a very sophisticated document and rightly so. Pulmonary embolism is not a single disease entity with a single level of severity. A hemodynamically stable patient with a relatively small embolic burden is not the same as a patient in obstructive shock. Management must take account of that.

No guideline can cover every scenario you will encounter in practice. Guidelines represent what the evidence generally supports. You will have to work out how that evidence applies to the individual patient in front of you. That process requires knowledge, experience, and the readiness to acknowledge that sometimes a case that looks quite simple on the surface is, in fact, not simple at all.

I’m concerned about a generation of doctors who will be excellent at ticking boxes but will become increasingly uneasy making decisions that require some interpretation. That’s not a criticism of the young doctors per se. That’s a concern about the environment in which we are training them. If we consistently reward checklist completion over the quality of clinical reasoning, we shouldn’t be surprised when checklists become the focus of clinical practice.

The best doctors I’ve known weren’t necessarily the ones who ordered the most tests or knew the most guidelines by heart. They understood their patients, could see the danger, knew the limits of their own knowledge, and could explain why they made the decisions they did. They were comfortable using the technology, but they didn’t equate having the information with having wisdom.

A Culture That Cannot Tolerate Uncertainty

Perhaps the most significant change I’ve seen in medicine is the growing belief that uncertainty itself is a form of failure. Patients want answers, doctors want reassurance, organizations want predictable results, and regulators want demonstrable compliance. All of this is perfectly reasonable, but it can lead to an impossible standard. We start to treat every adverse event as preventable and every missed diagnosis as a failure to do enough testing. Medicine has never worked that way.

Even the most rigorously assessed patient carries some residual risk. Any diagnostic test has inherent limitations. Clinical pictures can be very similar. Diseases progress. A treatment that clearly benefits one patient may harm another. You can be as evidence-based as you can be and still encounter an unexpected result. Accepting these realities is no excuse for poor care. It is absolutely the essential precursor to good care.

The alternative is an unending quest for certainty, where every symptom invites you to consider every possible disease. A patient with a bit of leg discomfort is going to be thinking about DVT. A brief episode of breathlessness will make you worry about a PE. A slight variation in heart rate can trigger a cardiac emergency. Some of these symptoms warrant investigation, and some do not. It all comes down to the full clinical picture.

The problem is of course exacerbated when fear starts to color the patient’s experience. Any doctor who has been in the game for any length of time will have seen patients who, once told they have a serious illness, become intensely sensitive to all the normal things the body does. The sensations are genuine, but their interpretation is completely different. An ache that would have been ignored before is now something to be deeply concerned about. A normal fluctuation in heart rate now prompts the patient to measure it again and again. The more closely they watch the body, the more “abnormalities” they seem to see.

Physicians can actually unwittingly perpetuate this. If every low-probability worry leads to further investigation, the patient can conclude that the only way to be safe is through more testing. A negative study gives you a moment of peace, but the next symptom gives you a new worry. You end up in a situation where the quest for reassurance is in fact generating more anxiety.

This is by no means an argument for dismissing symptoms as psychological or for skipping appropriate investigations. It simply argues that reassurance isn’t just about ordering tests. Patients need to be told what is going on. They need an explanation of why a diagnosis is likely or unlikely, what findings would lead you to reassess the diagnosis, and what level of uncertainty is acceptable.

Sometimes the most useful intervention is a well-thought-through conversation. The problem is that our system makes it far easier to order an expensive test than to spend 20 minutes explaining why the test may not, in fact, be needed. That should trouble us.

What Are We Teaching Our Students?

I’ve spent a fair bit of my career teaching medical students and junior doctors. I still enjoy it because it gets us to think not just about what we know, but how we know it. A student can memorize Virchow’s triad in an afternoon. Learning how to assess a patient with suspected pulmonary embolism is a lot more time-consuming.

The key is integration. The student has to learn to identify risk factors, know the symptoms, understand what the physical examination findings mean, and apply the diagnostic evidence. The real difficulty in medicine is deciding when you don’t have all the information. That is the difficult bit and that is the bit you can’t just learn by memorizing another algorithm.

I frequently wonder what happens when a junior doctor sees an experienced doctor ordering a test simply because they’re afraid of being criticized. The junior doctor may not hear that rationale at all. They will see that the standard response to uncertainty is to order further tests. That becomes the learned behavior. The lesson is: when in doubt, order more. If the first test is inconclusive, order another. If the test is abnormal but the clinical significance is unclear, refer the patient. Eventually, no one seems willing to decide that the information we have is sufficient.

I don’t want doctors to be reckless. I want them to be thoughtful. I want them to understand that failing to consider a potentially dangerous possibility can be a very serious error, but failing to consider it may also be a serious error. I want them to appreciate that clinical reasoning is about balancing those two concerns rather than simply eliminating one in favor of the other. A doctor should be able to explain why the test is being ordered, what the result might mean, and how that might change management. If the doctor can’t answer those questions, then there’s a problem that needs to be sorted out before ordering the test.

This is absolutely not about nostalgia for a previous age of medicine. I have no desire to go back to a time when we simply didn’t have the means to diagnose and treat the diseases we have today. What concerns me is that the amazing accomplishments of modern medicine are blinding us to the intellectual responsibilities that come with them. We should teach students to think with technology rather than hand over their thinking to it.

The Courage to Practice Medicine

It takes guts to recognize a life-threatening PE and act. It may take equally good judgment to assess a low-risk patient, decide further imaging isn’t required, explain that decision to the patient, and arrange appropriate follow-up. Neither approach is superior. The right approach is the one that is appropriate in the circumstances.

The issue with defensive medicine is that it subtly alters the question. Rather than asking what is in the patient’s best interests, the doctor may end up asking what decision will be easiest to defend if something goes wrong. The answer to that question may be the same as the answer to the original question in some cases, but not in all.

I’ve seen medicine done under immense pressure. In an ICU you’re constantly making decisions with incomplete information and very little time. There may be no opportunity to get every test, to consult every specialist, to wait for the absolute certainty. You have to assess the available evidence, identify the most pressing threats, and act. You also have to be prepared to change your mind if new information comes in.

That’s not a failure of evidence-based medicine. It’s evidence-based medicine in the real world.

The same principles apply outside the ICU in less urgent circumstances. The doctor has to work out what is possible and what is probable, what can be seen and what needs to be treated, and what gives you a momentary sense of reassurance and what actually does any good.

None of this means that experience makes a doctor infallible. An experienced clinician can be wrong, indeed can be very wrong. Experience can lead to insight, but it can also reinforce habits and biases. The answer is not to substitute evidence with intuition. The answer is to combine evidence with judicious judgment and appropriate humility about the limits of both. A doctor who ignores the warning signs because of his own confidence is not exercising sound judgment. A doctor who orders every investigation he can think of out of fear is also not exercising sound judgment. Both have abandoned the judicious approach that patients are entitled to.

There is a middle ground (though it is increasingly hard to find). It is about taking responsibility for your decisions without pretending that you can remove all of the uncertainty. It is about being able to tell the patient why you think a particular approach is appropriate, even if that approach doesn’t involve the most exhaustive possible investigation. It is about the institutions and professional bodies recognising that the quality of care cannot be judged by the number of tests you do or by the absence of any surprises.

What the Blood Clot Story Really Taught Me

When I was talking to the media about the professional athlete who’d had a pulmonary embolism, I didn’t have access to his full medical record. I could talk about the pathophysiology of thrombosis, the potential implications of a PE and the issues around travel and sport. I simply couldn’t in any responsible way say whether the doctors had made the right decisions given only the public information available.

That distinction matters, particularly in the current media climate, where people are so quick to draw conclusions. But the episode brought home a much wider issue. We live in a society that increasingly expects medicine to predict and prevent every single catastrophe. When something goes wrong the first question is invariably ‘Why didn’t we do more?’ We rarely ask whether doing more in every other comparable situation would have actually led to a better outcome. That question does deserve an answer.

Back in the very early days (nearly 30 years ago), I wrote about a particular manifestation of PE that doctors should watch for. Over 20 years later, I was in a position to raise concerns about the increasingly wide use of CT pulmonary angiography. To someone who believes medicine must choose between an aggressive investigative approach and a more conservative one, those two positions may seem at odds. To me they are not. They are the same principle. The doctor has to know when to look for disease and when further investigation is unlikely to help.

There will always be patients with very subtle presentations and difficult to make diagnoses. There will always be diseases that we will fail to pick up at an early stage despite what we would consider to be appropriate care. Some patients will also be harmed by the investigations we do, by incidental findings, and by the treatments we initiate after discovering something whose importance we never really considered.

Good medicine requires appreciating both realities. I’m still very excited about the future of medical tech. AI, better imaging, molecular diagnostics, and increasingly sophisticated therapies all have the potential. Some will undoubtedly improve our ability to detect disease and to treat it in a personalized way. Others may well lead to new forms of overdiagnosis and inappropriate intervention if we adopt them without properly considering the clinical implications. The answer isn’t to resist progress. The answer is to remember what progress is meant to achieve.

Our aim is not to produce the most medical information possible. It’s to help people to live healthier, better lives. That may involve an extraordinary intervention at times. It may involve a simple test. It may involve doing nothing at all at the moment and simply observing very carefully whilst remaining ready to act if necessary.

After 30 years in medicine, I think one of the most important questions a doctor can ask is also one of the simplest: will what I’m about to do actually help this patient? It’s a question that absolutely deserves to be asked before every test, procedure, and prescription.

We now have machines that can see things previous generations could only dream of. We can look at the pulmonary circulation in exquisite detail and spot clots which would have been completely invisible before. That’s all very impressive and we should be proud of it. It also comes with a responsibility. The more we can see, the more careful we have to be about what we make of that seeing.

Medicine doesn’t need fewer scientific advances. It needs doctors who know how to use them wisely. It needs the courage to investigate when warranted, the discipline to know when not to, and the humility to accept that, however much technology we have, there will always be an element of uncertainty. Above all, remember that the patient is more than a set of abnormalities to be found.

Common sense was never the enemy of scientific medicine. In fact, properly understood, it’s one of the things that makes scientific medicine worthwhile.

References

1. Virchow R. Die Verstopfung der Lungenarterie und ihre Folgen. Beiträge zur experimentellen Pathologie und Physiologie. 1846:1-90.
2. Virchow R. Thrombose und Embolie. Gefässentzündung und septische Infektion. In: Gesammelte Abhandlungen zur wissenschaftlichen Medicin. Frankfurt am Main: Meidinger Sohn & Comp; 1856. p. 219-732.
3. Bagot CN, Arya R. Virchow and his triad: a question of attribution. Br J Haematol. 2008;143(2):180-190. doi:10.1111/j.1365-2141.2008.07323.x.
4. Varon J, Fromm RE Jr. Syncope: the forgotten sign of pulmonary embolism. J Emerg Med. 1998;16(1):117-118. doi:10.1016/S0736-4679(97)00249-7.
5. Alshami A, Alhillan A, Varon J. Computed tomographic angiography in pulmonary embolism: diagnostic or a screening tool? Curr Respir Med Rev. 2019;15(3):159-160. doi:10.2174/1573398X1503191125144633.
6. Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med. 2011;171(9):831-837. doi:10.1001/archinternmed.2011.178.
7. Studdert DM, Mello MM, Sage WM, DesRoches CM, Peugh J, Zapert K, et al. Defensive medicine among high-risk specialist physicians in a volatile malpractice environment. JAMA. 2005;293(21):2609-2617. doi:10.1001/jama.293.21.2609.
8. Kline JA, Courtney DM, Kabrhel C, Moore CL, Smithline HA, Plewa MC, et al. Prospective multicenter evaluation of the pulmonary embolism rule-out criteria. J Thromb Haemost. 2008;6(5):772-780. doi:10.1111/j.1538-7836.2008.02944.x.
9. van der Hulle T, Cheung WY, Kooij S, Beenen LFM, van Bemmel T, van Es J, et al. Simplified diagnostic management of suspected pulmonary embolism (the YEARS study): a prospective, multicentre, cohort study. Lancet. 2017;390(10091):289-297. doi:10.1016/S0140-6736(17)30885-1.