Discussion Board

Expand all | Collapse all

'Diagnostic overshadowing'

  • 1.  'Diagnostic overshadowing'

    Posted 01-18-2021 14:00
    Paralyzed patient with a history of mental illness is assumed to be faking. 

  • 2.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 14:11
    Thank you for posting this, Helen. Standards for psychosomatic diagnosis lead clinicians to believe that the presence of mental health disorders increases the likelihood that symptoms are psychosomatic rather than disease-related. That is straightforwardly false. Those with mental health disorders are actually more likely to suffer from disease.

    These kinds of errors are all over the place in the standards that guide clinicians in deciding which symptoms to take seriously. It is mind-boggling that diagnostic training in this area has never been evaluated for safety.

    Diane O'Leary

  • 3.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 14:25
    It is considered 'pretty normal' for patients to minimize their symptoms, even when they are likely to be part of a health problem.  Ignoring a cough for 3-4 months, assuming that a little blood in the urine is normal, that severe headaches, shortness of breath, dismissing a racing heartbeat while at rest, and enjoying a sudden weight loss is pretty typical behavior for a mentally sound person to make, as he tries to avoid the reality of a potentially serious illness.  

    If a doctor assumes that a mentally sound patient can ignore obvious problems, why is it so hard to think that the 'unsound' patient might do the reverse, i.e., report the symptoms that are indeed problematic?

    Peggy Zuckerman

  • 4.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 16:23
      |   view attached
    Diane: the particular error of under-diagnosing medical illness in the psychiatric patient is common - it is one of a cluster of biases leading to error in psychiatric patients referred to as the 'Psych-Out errors'. Attached is a clinical case from our recent book, The Cognitive Autopsy' in which the cluster is described in more detail.


    TCA Case 5.pdf   250 KB 1 version

  • 5.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 16:40
    Thank you, Pat.  What a terrific resource your book is - I will be making good use of it.  What I'm trying to say, I think, is that the error of under-diagnosing medical illness in psychiatric patients is not a bias.  It's an explicit directive that's entirely standard in textbooks, practice recommendations and guidelines.  In the Oxford Textbook of Medicine, for example, the presence of psychiatric conditions is given as a primary factor to support psychosomatic diagnosis.  

    I'm not sure why we handle this kind of thing as if it's a matter of personal bias, when it's actually a standard practice recommendation.  We do the same with the gender issue.  I have done it myself, and I've just recently realized that this might be how the problem gets perpetuated.  

  • 6.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 16:47
      |   view attached
    Apologies. I think the book chapter on the cluster of Psych-Out errors got dis-allowed. It was in a teenage psychiatric patient whose rapid breathing was attributed to an acute anxiety state. At autopsy she was found to have massive bilateral pulmonary saddle emboli (she was being urged to re-breathe into a paper bag shortly before her arrest). Attached is an abstract from the book describing the cluster of biases that lead to error in psychiatric patients.


  • 7.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 15:12

    Oh Helen!


    The stories I could tell.


    It's not just patients with a history of mental illness.  There seems to be a segment of the healthcare system that has an innate tendency to poo-poo certain patients according to their skin color, the name by which they call god, their age, their personal hygiene, there sex, their political leanings, and on and on.  In the vast majority of cases I witnessed or was involved in correcting, there was a significant medical problem that, in many cases, would have lead to serious morbidity or mortality if the 'mistake' had not been caught in time.


    Perhaps a compendium of cases illustrating each situation and how to teach our fellow professionals to avoid making this type of diagnostic error would be a good project to pursue.




    Mark Gusack, M.D.


    MANX Enterprises, Ltd.

    304 521-1980



  • 8.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 15:22

    So well said, Mark.  Wouldn't it be more effective, though, to correct the original training?  For example, in an overview of "somatic symptom disorder" revised in 2019, American Family Physician recommends a female-to-male ratio of 10:1.  Do women need disease care any less often than men?  Nope.  

  • 9.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 15:45

    Agreed.  We need to incorporate this into original training.


    However, we already have large numbers of active healthcare workers doing this on a daily basis  with immediate harm being done.  So, from my perspective, one begins here and then moves back towards the nursing, PA, medical schools focusing on the faculty and finally, the students themselves.  Otherwise, what gets taught during school tends to be 'forgotten' under the pressures of externships, internships, and residences as I have witnessed.




    Mark Gusack, M.D.


    MANX Enterprises, Ltd.

    304 521-1980




  • 10.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 16:24

    This is really important - the urgency of the problem.  You're definitely right about that.  But there's something very strange about the way the profession deals with this area of practice.  In any other area, we demand scientific support for dx principles that become standard in textbooks.  And when blatant errors persist that affect big portions of the population, there's effort to correct them.

    Whatever the reason, medicine has allowed psychiatry to develop principles for managing dx uncertainty for decades, without ever stepping in to evaluate for safety, even when those principles clearly threaten the lives of disadvantaged groups - particularly women.  I don't understand why there's never been an effort to develop standards for managing dx uncertainty that are evidence-based and safety tested.  

    This discussion happened to catch me at a moment when I'm writing about this problem!


  • 11.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 16:40

    Good Afternoon Diane:


    Perhaps part of the problem is our own fear regarding psychiatric issues which threaten our very self, versus somatic issues which merely threaten our bodies.  We can be more easily detached and 'objective' when talking about a patient's cardiac electrophysiology and how we might 'fix' that whereas, when confronted by a patient with psychiatric problems, it's more difficult.  One result is a type of defensive behavior towards our own fears by exhibiting distain towards this type of patient.


    The answer is to acknowledge this fear and to begin to build a better way to approach these patients in a warm but more 'objective' manner.  One way that I was taught by my father was to focus on the history and physical as if the patient had no prior psychiatric diagnosis and only after assessing the findings, factor in the potential for the psychiatric illness influencing the presentation.  Another factor is the longevity and trustworthiness of the clinician-patient relationship.  If those who are very familiar with this patient are presented with the potentially somatic disorder, they may be able to discern a difference in affect that indicates this really is a separate issue and not the product of the patient's psychiatric state.


    Regardless, of the cause of the bias, it's an ever present threat to patient safety.  So, I advocate starting at 'the top' with the clinical staff and work backwards towards those teaching in our healthcare institutions, and finally, getting the message to the students so that, when they go on the wards, they experience an identity of ethical standards that they were taught.


    Hope you're having a wonderful MLK day!




    Mark Gusack, M.D.


    MANX Enterprises, Ltd.

    304 521-1980



  • 12.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 17:38
    Thanks, Mark.  I like your account of the fear involved with this kind of thing.  I think there's a lot of fear around dx uncertainty generally, and a lot of frustration with conflict that arises with psychosomatic dx.

    Hope you're having a good MLK day too!

  • 13.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 17:01



    To expand on this, as you know there is a lot of literature, and I can certainly testify from personal experience, that when we hear a didactic lecture much of what we have "learned" is forgotten in days. Hence, we believe that longitudinal education in medical decision making and diagnostic error is essential to having a meaningful, sustained impact on trainees.


    We introduce this with some case-based presentations in the first year of medical school. When on their neurobiology block, for example, we have a 1 hour discussion of a common neuro patient problem, focusing on diagnosis and reasoning concepts. The entire 3rd year medical school class takes a ~ 4-5 hour online curriculum we have developed in medical decision making and diagnostic error. There is an elective in the 4th year that repeats that curriculum and then expands on it with more in depth experience. About 30-40 students take that elective each year.


    Our entire Internal Medicine interns and resident class (~ 200) also takes the curriculum. This has now been extended to about half a dozen other residency training programs within our system. We have then integrated medical decision making into their ongoing clinical experience. For example, we have an extremely popular monthly conference in which an experienced faculty person is given a completely unknown case which uses a thinking out loud format for the clinician to explain their reasoning as they work through the case. We have a faculty member from our group who moderates the discussion, focusing not only on the diagnostic reasoning process but also the medical decision making concepts. We also have a brief "case of the month" (~ 10 min case) that the residents prepare that are distributed via email and Twitter, again highlighting diagnostic reasoning and diagnostic error concepts. About 80-100 or so people are doing these cases each month.


    Does this make a difference? It is hard to measure but we have observed some indication that perhaps it does.







    William P. Follansbee, MD, FACC, FACP, FASNC

    The Master Clinician Professor of Cardiovascular Medicine

    Director, The UPMC Clinical Center For Medical Decision Making

    Suite A429, UPMC Presbyterian

    200 Lothrop Street

    Pittsburgh, PA 15213

    Office: 412-647-3437

    Fax: 412-647-3873


  • 14.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 17:52
    Today's discussion about the man who crawled out of the Canadian hospital brings me back to the thread of a couple weeks ago about the diagnosis of conversion disorder.

    There was pretty much unanimous agreement among those who commented that "conversion disorder" (CD) should never be diagnosed in the ED. But that misses the real point which we're actually getting to now in this thread.

    What was not mentioned in the prior discussion was that conversion disorder is not a frequent diagnosis in the ED, especially when compared with the frequency of other diagnoses with psycho-somatic connotations. I'm referring to the far more common diagnoses of "anxiety", "panic disorder", "hyperventilation" (in pregnant women for gosh sakes), "drug-seeking" or "opioid use disorder" (in sickle anemia patients!), and the diagnostic pièce de résistance, "malingering." 

    The fact is that many patients who present with vague or hard-to-characterize complaints, "palpitations", "weakness", "tiredness" or fatigue, non-pain chest symptoms, numbness or tingling of extremities, abdominal or other pains, etc., get worked up to some degree and, when initial findings don't confirm a "physical" diagnosis, are sent home with diagnoses of those noted above and others, too often having the real cause determined months to years later. 

    This is the problem of medically unexplained symptoms, and already COVID is teaching us that it too can manifest itself in myriad inexplicable ways and long after the acute illness (remember polio?). Like syphilis, SLE, TB, AIDS, Lyme disease and a host of others, we were taught that if you know that disease you know medicine. And we believed it, only to be fooled again and again. That perhaps is the Grand Illusion, that by knowing the disease we are aware of, we know medicine 

    And even though the ED is often the place where we see these problems magnified and exposed, we all need to be humble and acknowledge that this problem is not just an ED issue; It is pervasive throughout medicine and medical culture.

    As for solving the problem, an awful lot of the best of medical school and residency teaching is dashed on the rocks of real world practice. And fixing that.....I don't know what will fix that. Thanks to all those who took a stab at it though.

    David L Meyers, MD, MBe,  FACEP
    Former Chief, Emergency Medicine (ret) - Sinai Hospital of Baltimore
    Former Board member; Chair, Board Governance Committee; Board Liaison-Patient Engagement Committee - Society to Improve Diagnosis in Medicine (www.improvediagnosis.org)
    Adjunct Faculty, KSAS - Johns Hopkins University 

  • 15.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 19:04
    One more area that needs attention, the topic of my next column: the misattribution of geriatric patients' physical symptoms as anxiety or cognitive issues (dementia, Alzheimer's, memory loss), dismissal of their pain reports including chest pain and orthopedic pain, and the overdiagnosis of many other conditions solely associated with age. 

    I spoke with one geriatrician today who told me that he thinks my mother is "losing it" cognitively when the truth is she was questioning his diagnostic work up because he ignored her report of new symptoms. Yes she's 93 and has fallen twice recently with injury both times. Yes sometimes she loses words. Most people over 75 have some language retrieval issues. She's not losing it. She's just lost patience with him.

    Now he's urging me to move her from her home because of her memory. That might be a solution for her balance but not because of her memory. Were he speaking with any other adult child multiple states away from their super senior parent, they might have taken his word for it and put them into memory care. 

    Frustrated caregiver. 


  • 16.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 19:48
    Is interesting. Where are we going?

    I have always thought that the way airline pilots are trained has something for us.

    Here is a link to how aan airline pilot is trained.

  • 17.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 19:55
    This is such an intersting important discussion.

    For some reason I have thought that diagnostic training could use aspects of the way airline pilots are trained.

    Herre is a a link tow airline pilots are trained.

  • 18.  RE: 'Diagnostic overshadowing'

    Posted 01-18-2021 20:22
    This Is so interesting. Where are we going with education in diagnoses?

    I have always thought that the way airline pilots are trained has something for us. We could take some of their successes and transfer it to us.

    Here is a link to the training of an airline pilot.


    Somehow I thought that the repetitiveness of flight simulators would be good for the average resident and it would be a step forward I said to myself 

    I seem to recall some flight flight simulator research being talked about in medicine a few years ago, but I have not read anything recently. Why has the use of such help disappeared? What were the problems?

    We are told that the number of errors in diagnosis is rising with more patients dying. And in contrast, globally the number of planes crashing has dramatically decreased in the last 70 years. Why is this?

    I could be wrong, but have often felt that physicians in clinical practice do not ask enough questions of their colleagues and even themselves and also often reluctant to answer questions. Does this needs to be fully understood. Is this related to errors in diagnosis?

    Robert Bell M.D. 

  • 19.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 00:34
    Everyone agrees that the error was bad. We need to discuss: 
    (1.) how the doctors were intellectually seduced into making the error, and
    (2.) how to prevent others from being seduced into making the same mistake in the future. 
    We need to understand the seduction before we can prevent it. 

    What I have read suggests that there were three key facts in front of the doctors:
    (A.) the patient's complaint of pain, 
    (B.) normal MRI, and
    (C.) the patient's bipolar diagnosis.
    The providers were seduced into the conclusion that his mental illness explained the patient's complaint of pain. 

    Many reasoning errors could have contributed. But how were the providers intellectually seduced? 

    Here is a factor that seems to me to be "seductive" about the conclusion that the mental illness explained the patient's complaint of pain. That explanation is consistent the three key facts:
    - the normal MRI, 
    - the complaint of pain, and
    - the longstanding diagnosis of bipolar disease. 
    We have all seen other errors rely on this pattern of reasoning: a theory that offers a simple explanation of the key facts can be the cause. 

    The problem with this error is far greater than the fact that the patient's history of mental illness. Certainly, that prior diagnosis is an important factor. 

    What bothers me the most is that the medical evaluation is minimal. Only one step in the evaluation appears to have been done: an MRI was performed. We need to do better than arriving at conclusions after the one step. Did anyone consider causes of leg pain that were not visible on an MRI? Was the ER doctor aware of other causes? Was the psychiatrist aware of other causes? What other causes of leg pain did the doctors consider? The available information suggests that no other causes were considered. Did the doctors really think that all physical causes of leg pain are visible on MRIs? How could any ER doctor think that? 

    I agree that the patient's history of mental illness contributed to the diagnostic error. Several additional serious mistakes contributed.  We should be looking at all of them. We need to ask: "What is seductive (right) about the doctor's reasoning?" 

    We should be making a list of ALL the things that can be done to prevent doctors from being seduced into similar errors in the future. If we limit this to one, we may not make much progress quickly. 

    Lee Tilson

  • 20.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 13:14

    The points that have been raised here are very good. I want to add another aspect of the diagnostic reasoning process, and that is the over-reliance on advanced testing especially imaging. If a diagnosis is uncertain or the clinician is somewhat lazy or not diligent, an MRI or CAT scan is ordered and if this is negative then the patient is deemed free of disease. In my experience of review of many charts, I have seen too many situations where a patient presents with a complaint of a certain body area, an imaging study is ordered of that area, and this is relied upon to make the diagnosis. As is well known not all pathology. is evident on these scans. I have talked to radiologists and they have expressed concerns that they have become the definitive diagnostic clinician of the patient. A broad differential and use of a proper History and Physical are basic but still vital in our age of technology.

    Helmut Meisl MD FACEP

  • 21.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 13:54
    Helmut, your point is certainly important. From the video and the news story, we have no idea if the patient's history and physical examination were appropriate and careful enough to elicit overt, let alone, subtle findings that could have led to Guillain-Barré Syndrome, the diagnosis he was ultimately found to have. From a diagnosis perspective, that's where we should be starting. 

    The wrongness of allowing a patient to crawl out of the emergency room is a totally separate issue.


    David L Meyers, MD, MBE, FACEP
    Mobile: 410-952-8782

  • 22.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 13:57
    Excellent point. Is there a place for research on that point. Should we know the figures. Perhaps for some issues the data is there?


    Sent from my iPhone

  • 23.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 14:10

    Absolutely correct, Helmut!

    We have seen an exponential growth in imaging volume over the past few years, a continuing trend over decades, as unrealistic over-confidence in the power of imaging for diagnosis has arisen simultaneously with a decline in confidence in diagnostic physical exam skills (and decline in the amount of time allotted to doctors in their clinics to try to ellicit a full history).  Some of this is driven by the power and utility of the modality, but some is driven largely by financial and time concerns.  Since more and more physicians are employees now, and not allowed to have any control of their appointments and schedules, there is clearly a pressure to offload the diagnosis to the radiology department in order to keep the patient encounter short. Some of the problem is also undoubtely due to loss of good physical exam training in medical schools and residencies, and some of it is simply due to clinician anxiety, ordering a scan "just in case," even though the H&P is directing them toward a different diagnosis.  Studies presented at prior SIDM meetings have shown that the act of simply ordering a CT scan decreases clinician anxiety, long before any results are available!  This is clearly a real driver of the decision to scan, but is not entirely rational.

    We have discussed Bayes' theorum in this forum in the past and have also discussed signal-detection theory in terms of ROC curves for evaluating the effectiveness of any test--very commonly used to assess the utility of radiology studies.  I can share more about these topics if anyone is interested, but the point is really that advanced imaging doesn't always yield a definitive, final answer--even when there is no radiologist error involved--it's just another line of imperfect and incomplete evidence to add to the overall diagnostic formulation.

    One of my colleagues here at Penn State likes to sarcastically refer to the CT scanner as "the truth machine."  There is definitely room for clinicians to be quite a bit more skeptical about its outputs than most are.  So many disease states look like other disease states (which is why we rely on Baysean reasoning) and there is the very real issue of false-negatives and false-positives, which commonly occur.  As the machines have gotten technically so much better and the pictures are so anatomically beautiful, it lends itself to over-confidence in their diagnostic perfection as well.

    As a radiologist I would urge maintaining a healthy skepticism regarding what imaging can do for us diagnostically, and also to keep squarely in mind what it can't do, and remember that diagnostic uncertainties usually remain even after imaging.  CT or MRI cannot be relied upon as the sole means of establishing a diagnosis any more than any other diagnostic test can.

    I also am a strong advocate of evidence-based utilization guidelines for deciding when to use advanced imaging.  Performing a CT or MRI study which is not indicated based on pre-test clinical suspicion is tantamount to using the imaging test for screening an unselected population.  On a statistical basis, most positive results in such a scenario are likely to be false positives!  

    Thank-you Helmut for initiating this conversation.




    Michael A. Bruno, M.D., M.S., F.A.C.R.  
    Professor of Radiology & Medicine

    Vice Chair for Quality & Patient Safety

    Chief, Division of Emergency Radiology

    Penn State Milton S. Hershey Medical Center
    ( (717) 531-8703  |  6 (717) 531-5737

    * mbruno@pennstatehealth.psu.edu  


  • 24.  RE: 'Diagnostic overshadowing'

    Posted 01-19-2021 14:17