What's your differential diagnosis
doc? A 67 year old presents to the hospital with a three day history of "shortness of breath" and "pain in my left chest with coughing". The vital signs showed BP 142/78 HR 110 T 99.6 F and an O2 sat of 91% on room. Laboratory showed an ABG of 7.45/33/60 on room, a WBC of 14K and a troponin of 0.6 with normal CPK and MB fractions. The chest xray suggested changes of COPD and an early infiltrate in the left lower lobe, new when compared with a chest xray from six months prior. The physical exam suggested the absence of wheezing and no signs or symptoms of heart failure. The EKG suggested sinus with nonspecific T wave changes and the telemetry showed bursts of paroxysmal atrial fibrillation, presumably from the catecholamine surge.
doc? A 67 year old presents to the hospital with a three day history of "shortness of breath" and "pain in my left chest with coughing". The vital signs showed BP 142/78 HR 110 T 99.6 F and an O2 sat of 91% on room. Laboratory showed an ABG of 7.45/33/60 on room, a WBC of 14K and a troponin of 0.6 with normal CPK and MB fractions. The chest xray suggested changes of COPD and an early infiltrate in the left lower lobe, new when compared with a chest xray from six months prior. The physical exam suggested the absence of wheezing and no signs or symptoms of heart failure. The EKG suggested sinus with nonspecific T wave changes and the telemetry showed bursts of paroxysmal atrial fibrillation, presumably from the catecholamine surge.So the patient got admitted by another hospitalist with a diagnosis of community acquired pneumonia, was placed on Levaquin per evidence based guidelines and routine therapies. And thus began their hospital course. During their stay, they got a cardiac echo because of the elevated troponin. The echo showed a normal wall motion with an estimated pulmonary pressure of 65 mm HG.
While in the hospital for the next three days for their treatment of pneumonia, the patient continued to have brief bursts of rapid atrial fibrillation, resulting in the inability for early hospitalization discharge. However, their WBC count normalized and the subtle infiltrate remained stable to improved by day three while the patient remained on room air with borderline oxygenation saturations of 92-93%. But they continued to have pain in their chest with coughing. Through it all the doctor is constantly reassessing their medical decision making and changing their differential diagnosis.
We are now at a time where the hospitalist goes "off service" and in comes Happy. When I first presented to the patient's care three days after admission, I reviewed all the current data and the things that struck me as the most clinically relevant for my differential diagnosis were:
1) Chest pain with cough
2) Elevated troponin
3) Pulmonary hypertension
4) Paroxysmal bursts of atrial fibrillation
My first concern when seeing these four data points was to rule them out for in indolent pulmonary embolism. I have seen this scenario before with confirmed pulmonary embolism. After pulling up the patients EMR, I discovered they had fallen off a ladder three weeks prior and were seen in the emergency room as a trauma, but sent home. That gave fuel to my fire in my working differential diagnosis as to the possibility of a delayed diagnosis of pulmonary embolism. Was this possible negligence in play here? If the patient had a pulmonary embolism, could my partner's failure to diagnose it be considered compensatory in court if a bad outcome ensued? Despite practicing sound medicine and treating clinical data that pointed to pneumonia, could my partner be sued anyway if the delayed diagnosis of a pulmonary embolism resulted in a bad outcome? Could not putting pulmonary embolism into the differential diagnosis be considered negligence? The answer to that question is unfortunately yes. It is a resounding reality of the broken legal climate we practice in today.
The question I have to ask myself is:
Should I do the CT scan of her lungs to rule out pulmonary embolism?
The issue here is one of clinical judgment. What may be a part of my differential diagnosis may not be a part of my partner's differential diagnosis or the differential diagnosis of the specialist down the street. There isn't a protocol in the world that could guide me toward the right differential diagnosis every time. Only education and experience gets me closer. What I decide to do is entirely dependent on my education and experience. Some want to believe that the standard of care should prevail for my decision making. What would most physicians do in this situation? That is the fatal flaw of our medical-legal complex. Why? Because the standard of care is frequently a legal driven mentality, not based on science. And a standard of care based on fear is at the root of defensive medicine, no matter how rational or irrational it is. The fear drives the the standards. The standards drive the defensive medicine. It is a self sustaining loop of irrational fear based medicine.
In my situation above, I could explain every concerning finding by a diagnosis other than pulmonary embolism. I could have excluded pulmonary embolism from my differential diagnosis on the basis of the sound scientific principles before me.
- What if the chest pain came from the trauma, perhaps a deep muscular bruise?
- What if the elevated troponin came from hypoxemia and known CAD?
- What if the pulmonary hypertension came from the known advanced COPD?
- What if the paroxysmal bursts of atrial fibrillation came from the underlying structural atrial changes and the COPD?
What if the diagnosis was one of cardiac? I could explain every concerning finding by a diagnosis other than pulmonary embolism:
- What if the chest pain was an atypical presentation of cardiac pain?
- What if the elevated troponin indicated a true fixable lesion in the coronary arteries?
- What if the pulmonary hypertension came from the valvular disease?
- What if the atrial fibrillation was secondary to underlying cardiac disease?
But what if the standard in my community said that every physician should order a CT scan to rule out pulmonary embolism because that was the fear driven standard in my community? What if a plaintiff attorney lined up ten highly paid experts who each said the doctor should have ordered the CT scan and if they had, the patient would not have died from an undiagnosed pulmonary embolism? What if they said this despite the hospitalist's sound evaluation of the data on admission which suggested treatment for a primary pneumonia was appropriate and who's accuracy is only known after the fact? Should the hospitalist be considered negligent for not ordering a CT scan and missing the diagnosis because other docs say so, even though treating only for pneumonia is entirely appropriate? Must you put pulmonary embolism in the differential diagnosis, and if you do, must you test for it before a possible bad outcome occurs?
That is the quandary of our current threshold for negligence in our fear driven culture of defensive medicine.
My partner decided to treat the patient as a pure pneumonia. And that was entirely appropriate. However, my partner could have decided to treat the patient as a possible cardiac patient and gone down that route. Or they could have decided to treat the patient as a rule out pulmonary embolism from the day of admission. And all three decision trees and differential diagnosis processing would have been appropriate.
And only one of them would have been known to be correct, something not known until after the fact. Thus is the essence of the differential diagnosis process. You can only know after the fact if you differential diagnosis was correct.
None of the decision trees should be considered negligent even if the wrong decision tree was taken. But we live in a legal climate where they are. If I was to bring twenty doctors the same set of data for evaluation I would get twenty different permutations of care and recommendations. And I could still find ten highly paid doctors that say there is only one correct answer.
Thus is the quandary of our current medical legal climate in the differential diagnosis process.
Some doctors would have trained with a lower threshold for coronary disease and would recommend heart catheterization. Some would have trained with a lower threshold for pulmonary embolism and would have evaluated that decision tree on the day of admission. Some would have trained with a higher threshold to use antibiotics and deferred the antibiotics on day one.
My point is that guidelines can help you decide evidence based approaches to diseases. But there are few if any guidelines that exist in the vast expansive permutational possibilities of clinical medicine that is the essence of differential diagnosis building. Yet we practice medicine as if there is a right answer. That couldn't be farther from the truth. Not a day goes by where the cardiologist of the day changes the orders of their partner just 24 hours before. Not a day goes by where the infectious disease doctor decides to make a blanket antibiotic change due to their more or less experience with certain drugs. Not a day goes by where one hospitalist will see things far different than the hospitalist before them.
What if this patient died suddenly of a pulmonary embolism on hospital day one after being admitted with a diagnosis of pneumonia? What if they died suddenly of an MI? What if I decided not to order the CT scan to look for pulmonary embolism on day three, instead choosing watchful waiting as my decision of choice? And what if they died on day four? What if they died on day two, before I ever had a chance to see them? What if the differential diagnosis did not account for the ultimate diagnosis?
What if that decision was the wrong decision, despite what years of medical education and judgment suggested? What if the wrong decision is only known until after the fact?
Should a physician practice in fear that their decision, which they feel is medically sound, turns out to be the wrong decision, something known only after the fact? Some would argue that a bad outcome would, in retrospect, mean that the decision making wasn't sound. That it failed to rescue the patient from their illness. And therefor constitutes negligence.
To that I say, the differential diagnosis is a moving target through time and space. That is the essence of why failing to diagnose should never be considered negligence. The information obtained from doing something is just as valuable as the information obtained from doing nothing. I cannot predict the future. I can only use my experience and education to make clinical judgments. This is a right I have earned through my education and experience as certified by national boards of specialty. This is a right I have been granted by state licensing authorities. My judgments that will not always be right on day one. They will not always be right on day three. In fact, they may not always be right for weeks, months or years. And that fact does not constitute negligence. That is a part of the differential diagnosis process.
Patients may go misdiagnosed or undiagnosed for long periods of time. Some will have no ill effects. Some will die. But to suggest that the failure to make the diagnosis is negligence when no guidelines exist for exact sets of clinical data is exactly why defensive medicine is alive and well. When I see million dollar workups for common medical conditions ordered by physicians through every spectrum of specialty, one has to ask themselves one simple question:
Is every medical school, residency and fellowship program in this country so bad that they are unable to train doctors that can differentiate the probable from the possible? Or is the legal system we operate in so dysfunctional that it drives highly competent physicians of every specialty to order the possible, instead of the probable.
The fact that I even had to think to myself whether I could be sued (not lose, just be sued), for not ordering a CT scan to rule out a pulmonary embolism three days after a hospital admission, three days after a highly qualified doctor did everything right, says volumes about the internal conflict physicians experience daily in their differential diagnosis process and ultimate decision making trees. It is the essence of a legally driven differential diagnosis process. A process that has been corrupted by the possibility of legal action who's standards are themselves based on fear.
It is the worst of all worlds.
And you wonder why nobody can afford health care. It's all being spent on the possible, while we treat watchful waiting failure to diagnose as negligence.
Unless we can give physicians the freedom to use their education and experience to make medical decisions based on their best clinical judgment, without the fear of being sued for failure to diagnose, the diagnostic decision trees and differential diagnosis process we face hundreds of times a day will always tilt heavily in favor evaluating the possible, whether or not it's highly probable.
Until you change that, nothing changes.
By the way, I ordered the CT scan to rule out the pulmonary embolism because my clinical education and experience said it was necessary, even though the other hospitalist believed otherwise. The CT confirmed the presence of pneumonia without pulmonary embolism, which confirmed the original admitting diagnosis.
But what if it didn't? A situation that happens often...
