The Multiple Layers of Diagnostic Uncertainty

Author: Justin Bright, MD (EM Attending Physician, Henry Ford Hospital, Detroit, @JBright2021) // Edited by: Alex Koyfman (EM Attending Physician, UT Southwestern / Parkland Memorial Hospital, @EMHighAK)

The young female with lower abdominal pain.  The middle-aged male with atypical chest pain.  The elderly female that presents with vague symptoms of dizziness.  These are just the tip of the iceberg of chief complaints we will see in our emergency medicine careers.  Those with symptoms that don’t fit into a particular diagnostic box or with totally clean workups can be frustrating for patients and physicians alike.  We are taught in medical school that 90% of diagnoses can be made with a very meticulous history and physical.  But, until I was asked to write on the topic of diagnostic uncertainty, I had never really thought about how infrequently we actually make a slam dunk, no doubt about it diagnosis.

Diagnostic uncertainty has multiple implications.  Some physicians have a hard time saying “I don’t know” and more importantly, don’t convey that message effectively to their patients.  How many times have you discussed at length with a patient that x, y, z test is normal, that there’s nothing life threatening going on, and that it’s safe for discharge – only for the patient to say, “so then why do I feel the way I do?”  If this uncertainty is not explained appropriately, will a patient lose confidence in the physician?  Will this in turn lead to poor compliance and follow up by the patient?  Will the uncertainty lead to excessive diagnostic testing?  Does diagnostic uncertainty lead to more patient complaints and lawsuits?  Even if your tests end up coming back totally normal, how certain can you truly be that whatever issue the patient is having isn’t a big deal?

What are some of the barriers to making a certain diagnosis?  A patient’s inability to give a cogent history of their illness is particularly challenging.  Patients often don’t know how to describe the way they feel.  They will bring extra variables into the discussion in a true attempt to help you understand, even though more times than not it can just throw you off the scent.  We all know the patient who tries to explain away the heart attack they’re having as heartburn from a chili dog they ate 4 days ago.  As physicians we know that there is no causal link, but to the patient it totally makes sense.  In the end, it just clouds the picture of what’s going on with them, making it harder to come to a definitive diagnosis.  It is important to recognize that the average health literacy of a patient is that of a 5^th grader.  You can’t expect a patient to be able to tell you what’s wrong.  I liken it to me taking my car to a mechanic.  “I’m not sure what’s wrong with the car – it’s making this weird noise, only some of the time, and it’s been doing it for weeks.”  I don’t see it as my job to figure out what’s wrong – that’s why I brought it to the mechanic.

The next problem is more anecdotal than evidence-based, but I believe the increasing epidemic of obesity and reliance on prescription medications is making it substantially more difficult for physicians to diagnose their patients.  More and more atypical presentations of common illnesses are popping up.  I find the physical exam is becoming more and more unreliable as patients get larger – particularly abdominal pain.  Point tenderness doesn’t seem to correlate the way it used to.  Medications that modulate immune and hemodynamic responses to pain and illness can confound us as we try to figure out the root of a patient’s problem.  As such, we shift from a paradigm of a comprehensive history and physical to one where we overly rely on tests.  It’s generally pretty easy to make treatment and disposition decisions with abnormal testing.  But what do we do when the tests are completely normal?  Does that mean the patient is fine?  Or are we just kidding ourselves and hoping that they are fine?

Physicians contribute to the diagnostic uncertainty as well.  Our biggest hurdle is the manner in which we communicate with our patients.  Physicians often interrupt their patients within the first 7 seconds of the encounter.  This prevents us from fully eliciting the patient’s concerns and understanding just what occurred to make the patient seek out an emergency department on that particular day.   Many of our patients have social concerns as well as medical concerns, which can complicate the encounter and frustrate physicians.  Keeping that frustration in check will allow us to better understand our patients and prevent prematurely cutting off the discussion.

Physicians also aren’t great at communicating with each other.  As an example, we are becoming more and more reliant on radiologic imaging.  This means we are also at the mercy of the radiologists issuing the formal interpretation.  How often do we get studies that only confuse us instead of help us? We know the radiology reads that state “can’t rule out,” “clinical correlation advised,” and “recommend further testing.”  This is by no means meant to criticize our radiology friends, but rather to demonstrate how often we order a substantial amount of testing, only to be no closer to the truth than when we started.

It seems to me that diagnostic uncertainty is always going to be present in our field.  We will never be able to diagnose every single thing that walks through our doors.  Is it even our job to do that?  Here’s my approach to the diagnostic uncertainty we see every day.

First, when I approach the patient, I ask myself, is this patient dying right this instant?  Thankfully, that answer is usually “no,” but if they are, I go into ACLS/life-saving mode and do what needs to be done for the patient.  The next question I ask myself is, is the patient sick enough to stay in the hospital, or at the end of the encounter will they be able to go home?  Many times, this decision can be made before I’ve even spoken to the patient – based on age, chief complaint, vital signs, risk factors, and how the patient looks when I enter the room.  The key is to use this as a starting point for the encounter so you have something you’re working towards, while at the same time being very careful to prevent anchor bias.  You must give your patient the opportunity to declare him/herself.

If I think the patient is ultimately going to be well enough to go home, then I must ask myself – based on their chief complaint – what are the worst possible things that could be going on with them and how do I go about systematically ruling them out?  More important, while I’m going about their diagnostic evaluation, how can I make the patient feel better so they are comfortable with a discharge plan?  Does it require medication?  Will it require me to dedicate a chunk of time of uninterrupted discussion with the patient and family to reassure them and develop a follow up plan?  As I supervise residents and med students, this is something I see repeatedly neglected.  Sometimes we can be so immersed in the diagnostic plan that we forget that all the patient wants is to feel better.  Conversely, if I determine the patient will ultimately need to stay in the hospital, then I ask myself what are the tests I need to order to identify what’s wrong so I can initiate treatment and make the patient look good for whoever will be accepting care of the patient on the floor.

So the workup is back.  It’s totally normal.  We’re not sure what’s going on with the patient.  What is it that makes us admit some patients with normal workups, while we discharge others that have some abnormal finding of unknown significance?  Is that woman with abdominal pain, a WBC of 13 but normal CT actually fine?  Does the chest pain patient with normal ECG and trops really need to stay for a stress test?  Is it our medical knowledge that makes us certain?  Does our ever-growing clinical experience lead to a gestalt where we just know if someone is actually sick or not?

The truth is, I have no idea.  In situations with diagnostic uncertainty, I think our clinical decision making and disposition decisions occur when we find the sweet spot between clinical confidence and risk aversion.  Keeping a patient in the hospital is the easier of the two decisions when faced with diagnostic uncertainty.  You’re giving another physician a chance to evaluate and diagnose the patient.  You’re giving your patient some additional time to declare themselves as sick or not sick.  You’re also just sharing the medicolegal burden with another physician.  We must also ask ourselves, are we really doing our patient any service by keeping them in the hospital.  Will more testing get us any closer to the diagnostic truth?  Will it subject the patient to unnecessary testing and infectious pathogens?  What additional financial burden does this bring on to the patient, the hospital, and the American health care system?

The harder part is when you are uncertain, but want to send the patient home.  Telling the patient your tests turned up nothing while they continue to experience a particular symptom is very challenging.  We all know the patient who says they were seen in another ED and “they did nothing for me.”  We know the physicians who tell their patients “nothing’s wrong” while failing to validate the fact that they are experiencing a particular symptom.  My personal strategy is to tell the patient that I absolutely believe they are feeling a particular way.  I tell them that I don’t know why they are feeling that way.  Then I tell them that I have crossed off many of the life threatening things that could harm them.  The key here is also to assure the patient that you will be giving them medication to continue to use at home to feel better, and that you will help arrange follow up for them so they don’t feel like they’re grasping at air for answers.  It’s not perfect, but it’s what we have.

There are definitely patient encounters where you can tell – almost immediately – that you’re not going to have a definitive answer to their problem that day.  I like to manage the expectation up front.  I will lay out my plan for them, even saying I’m not confident that I will be able to give them a great answer, but I will make sure they are going to be ok.  How do I know they will be ok?  Because when all else fails, I follow the “no nightmare rule.”  Basically, I ask myself, what do I need to do to make sure that I don’t have a nightmare about this patient when I go home for the night?  Then I don’t stop until I’ve reached that point.  I don’t pretend that it’s evidence-based or necessarily good medicine.  But, it’s the world we practice in, and I like to say that I practice reality-based medicine.

I am beginning to wonder if the only certainty in our profession is diagnostic uncertainty.  It is something we are faced with every single shift we work.  Your clinical acumen, diagnostic approach, and communication style are going to dictate how successful you are in taking care of your patients and how confident your patients are in your ability to take care of them.  We as physicians need to be more comfortable saying “I don’t know.”  We also need to have a strategy that allows us – despite our uncertainty – to feel confident that we’re not sending home a time bomb.  The lives of our patients and our own livelihood depends on it.

References / Further Reading

1. Diagnostic Uncertainty: dichotomies are not the answer. Shinkins, Bethany. Br J Gen Pract. 2013 Mar 63(608); 122-123
2. Diagnostic uncertainty in general practice. Green, C. Eur J Gen Pract. 2003 Mar; 9(1): 13-15
3. Communicating diagnostic uncertainties to patients: The problems of explaining unclear diagnosis and risk. Than, M. Evidence Based Med 2009; 14: 66-67
4. Uncertainty and the Diagnostic Leviathan. Saurabah, J. JAMA Intern Med.2015;175(7):1085-1086.
5. Communicating diagnostic uncertainty in surgical pathology reports: Disparities between sender and receiver. Lindley, S. Pathology – Research and Practice. Volume 210, Issue 10, October 2014, Pages 628–633.
6. Coping with uncertainty in primary care. http://patient.info/doctor/Coping-with-Uncertainty-in-Primary-Care. First accessed: 12/26/15.
7. Editorial: “Irritable Bowel Symptoms” in Inflammatory Bowel Disease: Diagnostic Uncertainty Meets Pathological Reality. Eamonn, M. The American Journal of Gastroenterology107, 1483-1485 (October 2012)
8. Understanding laboratory testing in diagnostic uncertainty: a qualitative study in general practice. van der Weijden, T. British Journal of General Practice. December 2002: 974-980.
9. Understanding Medial Uncertainty: A Primer for Physicians. Ghosh, AK. Journal of the Association of Physicians in India (JAPI). September 2004, vol 52: 739-742
10. A Case of Medical Uncertainty. Wellbery, C. Am Fam Physician.2012 Mar 1;85(5):501-508.
11. Diagnosis deferred—the clinical spectrum of diagnostic uncertainty. Lossos, I. Journal of Clinical Epidemiology. 1989 Volume 42, Issue 7, Pages 649–657