Diagnosis gone awry: A flawed drug test could be sending thousands of innocent people to jail each year

Edward Winslow edbjwinslow at GMAIL.COM
Wed Jul 13 15:25:26 UTC 2016


I have two comments:

   1. This thread is also concordant with the  series on the Theranos,
   problem. How much do we really push on ensuring quality control of any lab
   tests?
   2. I agree with Liz that  in interpreting ANY test "It pays to know
   Bayes". Pre test probability is at least as important as test
   characteristics in confirming, or making less likely any clinical
   condition. One  could posit that any screening test should be confirmed
   with a more accurate (reliable) test. The positive screen would increase
   the pre test probability of the subsequent test.


On Tue, Jul 12, 2016 at 5:33 PM, Regan, Elizabeth <ReganE at njhealth.org>
wrote:

> I agree that this story is not about medical diagnosis per se, but does
> raise some important issue around medical diagnosis.
>
>
>
> With the trend to algorithms driving medical decision making – especially
> with mid-level providers, I believe that there is an inappropriate
> assumption that a test result is “positive” or “negative” and that this in
> turn drives a message to the patient that they do or do not have “disease
> X”.  This is fostered by specialists talking to primary care physicians and
> giving them shortcuts and diagnostic absolutes.
>
>
>
> As discussed already, every test has a finite error rate at the level of
> performance (that we try to manage with QA procedures) and also has some
> reported association to the disease state – positive and negative
> predictive values.  The predictive power is derived from the sensitivity
> and specificity of the test in one or more populations.  (and it changes
> depending on the prevalence of disease in the population).  Bottom line –
> no test is perfect and every test results should be considered carefully in
> light of the situation and reliability of the test itself.  Some (actually
> most) tests cannot be used to “rule out” a diagnosis because few have
> perfect sensitivity.
>
>
>
> I frequently hear patients and family members (and have been told myself)
> that because test A is negative – you do not have disease X.  If test A is
> positive in 70% of subject with disease X, a negative test is really not
> helpful for an individual patient.  The correct message is that we cannot
> be sure about whether the patient has Disease X based on that test – we
> only know that the test is negative and 30% of patients who truly had
> Disease X had a negative test.
>
> It is really a problem for diagnostic error to not take those basic
> epidemiology facts into account when assessing test results.  In the NY
> Times article the consequences were wrongful imprisonment and major losses
> of life stability.  In our patients it may mean that they are sent out of
> the ER – told that they didn’t have a heart attack/strept
> infection/fracture – and may have a bad outcome, because the judgement was
> made on the negative test result – not the whole picture.  The doctor who
> doesn’t see the potential for a test to be wrong – doesn’t hedge his/her
> discussion with the patient and emphasize the need to watch for symptoms
> and return.
>
>
>
> I would argue that there is far too much reliance on test results as an
> absolute.  I think it is done in part because providers don’t understand
> the science of testing and partly to fit into algorithms.  I also think
> that just like for the police – it meets a need.   That need is to move
> patients through the system at the cost of diagnostic accuracy.
>
>
>
> Liz
>
>
>
> *From: *HM Epstein <hmepstein at GMAIL.COM>
> *Reply-To: *"IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG" <
> IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG>, HM Epstein <hmepstein at GMAIL.COM>
> *Date: *Tuesday, July 12, 2016 at 1:53 PM
> *To: *"IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG" <
> IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG>
> *Subject: *Re: [IMPROVEDX] Diagnosis gone awry: A flawed drug test could
> be sending thousands of innocent people to jail each year
>
>
>
> Thanks for sharing that, Mark.
>
>
>
> This is a hot button for me, since before I focused on diagnostic error, I
> wrote about the issues and risks families face when children are accused of
> breaking alcohol consumption laws.
>
>
>
> These often faulty field tests go hand in glove with the tests done on
> potentially impaired drivers which don't test for current intoxication but
> test instead for signs of drug use that could be as old as two weeks or
> more. The statistics around alcohol and drug use and driving are often
> based on outmoded and coerced urine and blood tests.
>
>
>
> Once in the system, the arrested person rarely has access to knowledgeable
> counsel, just like the woman in the article. Pro-publica also did an
> exhaustive investigative review of how often people are arrested and
> falsely accused who then lose their automobiles - even if they are found
> not guilty - which are sold by the arresting police precinct or its
> municipality. Those proceeds go for new gear for the precinct or for
> overtime pay so the police officers can then stop more drivers and accuse
> them of driving under the influence. There's a lot more but it is off topic
> for this thread.
>
>
>
> Best,
>
> Helene
>
>
>
>
>
>
>
> *--  *
>
> *hmepstein.com <http://hmepstein.com> *
>
> *@hmepstein*
>
> *Mobile: 914-522-2116 <914-522-2116>*
>
>
>
> *Sent from my iPhone*
>
>
>
>
>
>
> On Jul 12, 2016, at 2:09 PM, Mark Graber <graber.mark at GMAIL.COM
> <graber.mark at gmail.com>> wrote:
>
> This is a disturbing article in this week’s NY Times magazine on the
> roadside tests police use to detect narcotics.  People are being jailed for
> narcotics possession based on tests that have high, and unknown, rates of
> false positive results.
>
>
>
> *http://tinyurl.com/zhqm44o <http://tinyurl.com/zhqm44o>*
>
>
> ------------------------------
>
>
> <IMPROVEDX at list.improvediagnosis.org>
>
> To unsubscribe from IMPROVEDX: click the following link:
>
> http://list.improvediagnosis.org/scripts/wa-IMPDIAG.exe?SUBED1=IMPROVEDX&A=1
>
> or send email to: IMPROVEDX-SIGNOFF-REQUEST at LIST.IMPROVEDIAGNOSIS.ORG
> <IMPROVEDX-SIGNOFF-REQUEST at list.improvediagnosis.org>
>
>
>
> Moderator:David Meyers, Board Member, Society for Improving Diagnosis in
> Medicine
>
> To learn more about SIDM visit:
> http://www.improvediagnosis.org/
>
>
> ------------------------------
>
>
>
> To unsubscribe from IMPROVEDX: click the following link:
>
> http://list.improvediagnosis.org/scripts/wa-IMPDIAG.exe?SUBED1=IMPROVEDX&A=1
>
> or send email to: IMPROVEDX-SIGNOFF-REQUEST at LIST.IMPROVEDIAGNOSIS.ORG
>
>
>
> Moderator:David Meyers, Board Member, Society for Improving Diagnosis in
> Medicine
>
> To learn more about SIDM visit:
> http://www.improvediagnosis.org/
>
> NOTICE: This email message is for the sole use of the intended
> recipient(s) and may contain confidential and privileged information. Any
> unauthorized review, use, disclosure or distribution is prohibited. If you
> are not the intended recipient, please contact the sender by reply email
> and destroy all copies of the original message.  ­­
>
> ------------------------------
>
>
> To unsubscribe from IMPROVEDX: click the following link:
>
> http://list.improvediagnosis.org/scripts/wa-IMPDIAG.exe?SUBED1=IMPROVEDX&A=1
> or send email to: IMPROVEDX-SIGNOFF-REQUEST at LIST.IMPROVEDIAGNOSIS.ORG
>
> Moderator:David Meyers, Board Member, Society for Improving Diagnosis in
> Medicine
>
> To learn more about SIDM visit:
> http://www.improvediagnosis.org/
>



-- 
*Edward B, J. Winslow, MD, MBA*
Home 847 256-2475; Mobile 847 508-1442
edbjwinslow at gmail.com
winslowmedical.com

"The only thing new in the world is the history that you don't know"
       Harry S. Truman, 33rd President of US (1945-1953)


"... it can be argued that underinvestment in assessing the past is likely
to
lead to faulty estimates and erroneous prescriptions for future action."
        Eli Ginzberg, 1997






Moderator: David Meyers, Board Member, Society to Improve Diagnosis in Medicine


HTML Version:
URL: <../attachments/20160713/47f6705f/attachment.html>


More information about the Test mailing list