Discussion Board

Expand all | Collapse all

Peirce and diagnosis

  • 1.  Peirce and diagnosis

    Posted 07-04-2021 11:46
      |   view attached
    In the attached paper, I argue that the method of scientific investigation consisting of abduction, deduction and induction developed by Charles Sanders Peirce, America's greatest philosopher of all time, is precisely similar to the method of diagnosis employed by experienced physicians in practice. 
    Peirce is now well recognized to be a precursor of Neyman in development of the frequentist confidence method of statistical inference.
    Please review and comment.

    Bimal

    Bimal Jain MD
    Mass General Brigham/Salem Hospital
    Salem MA 01970.
    The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at http://www.massgeneralbrigham.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

    Please note that this e-mail is not secure (encrypted).  If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately.  Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail. 

    Attachment(s)

    pdf
    Peirce and diagnosis.pdf   134 KB 1 version


  • 2.  RE: Peirce and diagnosis

    Posted 07-05-2021 08:24
      |   view attached

    Dear Bimal,

     

    Thank you for this delightful short paper reviewing the life, work and thoughts of Charles Sanders Pierce.  I really enjoyed it and learned a great deal.

     

    As you know, I have strong Baysean leanings, as this is the primary reasoning that is applied in Radiology toward the interpretation of test results.  As I've said before on this forum, I believe that this is because many widely different diseases strongly overlap in their appearance on radiographs and imaging scans, and therefore having a prior probability (pretest probability of disease) is essential to rank-ordering the probabilities of diagnoses in the final radiology report.  In this way, imaging differs from chemical or other types of testing in that answers are very rarely binary (e.g., a fracture or no fracture) but are often only probabilistic (e.g., a case where the appearance could be due to either ischemic, inflammatory or infectious colitis, and the pretest probability becomes the deciding factor).  While arriving at the final diagnosis may be a frequentist exercise, as Pierce demonstrates, at least one of the most common intermediate steps-medical imaging-is very much locked into a Baysean model of reasoning.

     

    All the best,

     

    image004.png@01D112FF.F77F98B0

    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  |  
    image001.jpg@01D04A9B.917CDCD0

     

    *****E-Mail Confidentiality Notice*****
    This message (including any attachments) contains information intended for a specific individual(s) and purpose that may be privileged, confidential or otherwise protected from disclosure pursuant to applicable law.  Any inappropriate use, distribution or copying of the message is strictly prohibited and may subject you to criminal or civil penalty.  If you have received this transmission in error, please reply to the sender indicating this error and delete the transmission from your system immediately.

     

     




    Attachment(s)

    pdf
    Peirce and diagnosis.pdf   134 KB 1 version


  • 3.  RE: Peirce and diagnosis

    Posted 07-06-2021 08:59
    One should keep in mind that Reverend Bayes’ paper (1763) was a reaction to David Hume’s argument in "Of Miracles” (1748). Hume said that the prior probability of two events must be assessed in order to determine which is most likely to be true. Hume used this agreement to show that miracles were not proof of the existence of God. Rev. Bayes thought he could use the same argument to prove the existence of God. Both Hume's and Bayes/Price’s arguments are directly derived from the axioms of probability.




  • 4.  RE: Peirce and diagnosis

    Posted 07-06-2021 09:52
    I was not aware of relationship between Hume and Bayes. Thanks for informing us. I will look into it.
    Bimal





  • 5.  RE: Peirce and diagnosis

    Posted 07-06-2021 14:32

    Hi all,

     

    Great bits of history here – I appreciate everyone's contribution to this discussion.  Thanks especially to Bimal for starting it and for all of the information about Pierce.  Yes, Bayes developed the theory in order to help prove the existence of God.  There are some very nice historical papers available on that topic.  I hadn't heard of Pierce before and certainly didn't know his history with Gilman at the time of the founding of Johns Hopkins University (my alma mater).  His scandalous behavior would not seem so by today's standards; quite often we've seen the converse, that widely accepted behavior in the past may seem scandalous now.

     

    Bayesian reasoning is very much part of the human thought armamentarium, and I would argue that it is important to diagnosis, although Bimal makes the case eloquently about it (perhaps never) being the final step.  In my field, diagnostic radiology, I am almost never the final word in the diagnosis, and if I'm honest with myself I will have to admit that I don't generally establish the final diagnosis with imaging.  My role is probably just to reduce diagnostic uncertainty enough so that a clinician can confidently go ahead with, or withhold, a therapeutic action.  So I'm in the business of giving PROVISIONAL diagnoses, and for those I must rely very heavily on Bayesian reasoning.  Quite a bit of my knowledge base is the information used to generate the pretest (posterior) probability of disease.  Otherwise I might be just another guy writing captions for pictures!

     

    Here's a nice paper that shows how Bayesian reasoning is used clinically – helping to provide stewardship for the use of imaging test for pulmonary embolism.  The premise is that we shouldn't just scan everyone.  Selection of patients based on the pretest probability of disease minimizes false positives and optimally utilizes limited resources. 

          DOI: 10.1001/jamacardio.2021.0064

    Roy, Pierre-Marie, Friiou, Emilie, Germeau, Boris, et al.  "Derivation and Validation of a 4-Level Clinical Pretest Probability Score for Suspected Pulmonary Embolism to Safely Decrease Imaging Testing."

     JAMA Cardiol 2021 June 1;6(6):669-677

    All the best,

     

    Mike

     

     

     

    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  

    1571679014277

     






  • 6.  RE: Peirce and diagnosis

    Posted 07-06-2021 14:40
    I'd echo Mike's comments, and suggest Andy Clark's article stimulating Volume 36 Issue 3 of Behavioral and Brain Sciences <https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/issue/012119F37B6516E2BF351B7A4C92B2D3> Whatever next? Predictive brains, situated agents, and the future of cognitive science, in particular section 2.2 and the notion of the "Bayesian Brain" as part of the larger emphasis on current thinking of neural activity/networks.
    Cheers,

    David





  • 7.  RE: Peirce and diagnosis

    Posted 07-06-2021 14:43

    Thanks, David – that's a good one!

     

     






  • 8.  RE: Peirce and diagnosis

    Posted 07-06-2021 09:38
    Dear Michael,
    Thanks for your comments. Your rank ordering of several diseases which can cause the same radiologic appearance seems to me to  due to interpreting prior probability as chance of a disease being present. This does not seem like Bayesian reasoning to me in which a prior probability is interpreted as a prior degree of belief. Such Bayesian reasoning is not done in practice, I believe, as it raises the risk of not suspecting or testing a disease with a low prior probability.
    A disease is inferred (diagnosed) in practice with a high degree of confidence from a procedure which has a high probability in the form of a high frequency of inferring the disease accurately from a highly informative test result. An example is diagnosis of pulmonary embolism from positive chest CT angiogram.
    Currently, I am looking at why the Bayesian method was prescribed for diagnosis in the early 1960s. So far, it seems to be due to a resurgence of interest in this method primarily due to work of Leonard Savage and overthrow of behavioral thinking by the cognitive revolution around this time. Once I have a clearer picture, I shall post a paper about it in the near future.
    Regards,

    Bimal





  • 9.  RE: Peirce and diagnosis

    Posted 07-06-2021 08:51
    C.S. Peirce was a well known philosopher who could not get a job at Harvard, nor because he was not brilliant, but because the dean did not like him. He is best known for his theory of signs. I would recommend reading his Collected Works (Belknap Press of Harvard University Press), especially volumes V and VI.




  • 10.  RE: Peirce and diagnosis

    Posted 07-06-2021 09:46
    For folks who are interested in work that has been done to use Peircian semiotics and apply it to medicine, I'd highly recommend the work of John Nessa at Bergen.
    In particular, the two following papers are of note:

    1. Nessa, John. "About Signs and Symptoms: Can Semiotics Expand the View of Clinical Medicine?" Theoretical Medicine 17, no. 4 (1996): 363–77.
    2. Nessa, John, and Iona Heath. "About Meaning More than Fact: The Semiotics of Clinical Medicine." In Philosophical Studies in Medicine and Health, edited by Stephen M. Modell, 78–94. URAM Monographs 3. URAM, University of Toronto Press, 2015.

    I would also note that the Peircian semiotic triad is not the only model out there, and other efforts (such as Greimas's semiotic square) might prove more fruitful, per Broden, Thomas F. "Greimas between France and Peirce." The American Journal of Semiotics 15, no. 1/4 (2008): 27-89.
    I've pondered quite a while ago the folks at the Society for Complexity in Acute Illness, so would be happy to chat off-list if anyone is interested.
    Cheers,

    David





  • 11.  RE: Peirce and diagnosis

    Posted 07-06-2021 11:38
    Thanks for the information
    Bimal





  • 12.  RE: Peirce and diagnosis

    Posted 07-06-2021 09:49
    Charles Eliot was an instructor at Harvard, who formed an unfavorable opinion of Peirce while he was an undergraduate there. Charles Eliot was president of Harvard from 1869 to 1910, practically all of Peirce's professional life, and blocked all efforts for Peirce to get aa academic appointment at Harvard.
    Gilman, who was president of Johns Hopkins did not like Peirce  as he found his personal life to be scandalous. Peirce lived with his second wife to be, Juliet Proissy, before he was divorced from his first wife, Melusina Fay. Gilman too blocked attempts by Peirce to get aa academic job anywhere. Sad!

    Bimal