Another look at Questions regarding Radiologists and Imaging

Bruno, Michael mbruno at HMC.PSU.EDU
Wed Mar 23 13:54:22 UTC 2016

Hi Tom,

Thanks for this post—it is a very nice example of what Bob, Peter, Lenny and I were saying earlier (see below) on this thread—about how even a small amount clinical guidance (i.e., what is suspected clinically, and, more importantly, why) can be a huge help with regard to radiologic interpretation—and the more such information and communication, the better.  Having an office assistant check the “Abdominal pain 789.00” box on the request form before they fax it over to the radiology dept. just isn’t going to produce the same results!

It is noteworthy that, in this case, the clinical suspicion was not exactly right, the diagnosis wasn’t a classic case of vascular disease causing mesenteric ischemia, but rather a variant on that theme.  Nonetheless, the extra bit of clinical input provided your radiologist with a valuable clue to guide their search, and a potential diagnostic error was avoided.

Perhaps part of the problem is our colloquial tendency to refer to radiological interpretation as a “read,” as if the scan were analogous to a book, and the radiologist just has to open it up and read it.  The implication is that the answers are all simply right there, plainly written in black-and-white.  In reality, the task is a lot more like solving a complex puzzle—one in which all of the pieces are in 256 subtle shades of gray—an environment of extremely high uncertainty.  With the radiologist thus in puzzle-solving mode, every clue they receive will help them a great deal.  Of course, some puzzles will be solved anyway, without any good clues, perpetuating the misconception that such guidance isn’t needed, and some won’t ever be solved.  But isn’t it better to take matters into our own hands as much as possible, and do everything in our power to improve our odds of success?

All the best,


Michael A. Bruno, M.S., M.D., F.A.C.R.
Professor of Radiology & Medicine
Vice Chair for Quality and Patient Safety
Chief, Division of Emergency Radiology

From: Tom Benzoni [mailto:benzonit at]
Sent: Wednesday, March 23, 2016 8:56 AM
To: Society to Improve Diagnosis in Medicine; Bruno, Michael
Subject: Re: [IMPROVEDX] Another look at Questions regarding Radiologists and Imaging

A note from yesterday on how this idea is used.
This a is a case report only, n=1.
Older female presented with abdominal pain, epigastric, crampy/burning, maybe left>right.
Occurs after eating, getting worse the last few weeks.
Attendance to ED/A&E prompted by a particularly bad episode that left her faint, diaphoretic, nauseated without vomiting.
Basically benign prior history for 70+.
Labs normal.
Exam non-concerning fro peritonitis.
Requested CTA; ? was mesenteric ischemia.
(Usually, docs just click the button that says Abdominal Pain for Reason; that's sufficient to satisfy the payers.)
I didn't train that way.
I wrote the rads a note that I was concerned about mesenteric ischemia but was flummoxed by her lack of vascular risk factors.
Rads called back that, due to that note, he looked for and found high grade obstruction from arcuate ligament syndrome with downstream hyperemia. No necrosis.
Basically, simple communication made a difference.
This is not astounding, but the very ordinariness  was striking.

On Tue, Mar 8, 2016 at 12:10 PM, Bruno, Michael <mbruno at<mailto:mbruno at>> wrote:
Hi all,
Great discussion!
As a practicing academic radiologist focused on quality and process improvement and a person actively engaged research directed toward error reduction in radiology, I also have some thoughts to share on the topic of Radiologist errors.  Our interdisciplinary, multi-institution research group based here at Hershey is very focused on understanding the roots of radiologist errors, especially perception errors (the most common type).  This has become a major focus of our work here at Penn State Hershey, and has grown directly out of our process improvement efforts.  See the attached review our group published late last year (note to David Myers: as 1st author I give SIDM permission to share this PDF on the listserv).  As Lenny noted, our specialty is committed to reducing our collective error rates, which have been a very tough problem—a needle that hasn’t been moved since the 1930’s when Garland published his first papers on the topic.  (And no, it didn’t make Dr. Garland popular among his fellow Radiologists at the time!)
Germane to this discussion: one obvious strategy for error reduction would be for radiologists to receive rapid feedback when error occur, which of course requires those errors to be discovered.  In this regard, the loss of frequent person-to-person interaction between Radiologists and our clinical colleagues is a particularly painful one for those of us who have practiced Radiology long enough to remember how valuable it was to us.  One thing that has clearly been lost is the inherently error-correcting mechanism of rapid, “formative” feedback from our clinical colleagues, which is something we don’t really get from peer review.  That has been a major topic of discussion in the American College of Radiology forums in which I participate as well, and the ACR has committed resources to explore research to detect errors, reduce them, and also reduce patient harms that result from our errors, so I can assure everyone on this list that these issues are in the forefront for organized Radiology.
It has long been a core tenet of Radiology education – which we drill into our residents – that the more clinical information we can receive, the better.  Since we don’t have enough contact with our clinical colleagues, our residents commonly troll the EMR for more clinical/historical details, a process which can be tedious and time-consuming.  While there is some inherent risk of bias in knowing clinical history that is potentially misleading or incorrect (and an argument can be made for reviewing the images once first, prior to reading the clinical information, then looking at them again after reading it) on the whole, there is little question in the mind of any practicing Radiologist that the clinical guidance we receive from our colleagues is invaluable, and the more the better.
As I mentioned above, the high rate of radiologist errors in practice has been a very difficult and seemingly intractable problem for decades.  We’re attacking this problem aggressively here in the Radiology Department at Penn State Hershey from a variety of standpoints: (1) improving the quality and clarity of our written and verbal reporting, so that actionable information isn’t lost in the noise, (2) improving our dialogue/liaison with our clinical colleagues, especially in the Emergency Department, (3) taking a hard look at workload and fatigue as a driver of error.  This involves mathematically modeling our daily workflows, in cooperation with Ph.D. engineering faculty from other Penn State campuses, to hopefully be able to adjust our radiologist staffing as nimbly as possible to allow radiologists sufficient time to get through the worklist without needing to rush on any case to satisfy our growing service obligations, and (4) designing basic, hypothesis-driven research to understand the underlying drivers of perceptual error and to design and rigorously test the effectiveness of strategies to address those vulnerabilities.
As results accrue on these ongoing efforts, I will report them to this group!
All the best,
[Description: Description: Description: \\\files\Staff\M\mbruno\Signature2.gif.gif]
Michael A. Bruno, M.S., M.D., F.A.C.R.
Professor of Radiology & Medicine
Vice Chair for Quality and Patient Safety
Chief, Division of Emergency Radiology
The Milton S. Hershey Medical Center
Penn State College of Medicine
500 University Drive, Mail Code H-066
Hershey, PA  17033

Phone: (717) 531-8703<tel:%28717%29%20531-8703>
Fax:      (717) 531-5737<tel:%28717%29%20531-5737>

e-mail: mbruno at<mailto:mbruno at>
[cid:image002.png at 01CF4E6D.0D6442B0]

From: Elias Peter [mailto:pheski69 at GMAIL.COM<mailto:pheski69 at GMAIL.COM>]
Sent: Sunday, March 06, 2016 11:26 PM
Subject: Re: [IMPROVEDX] Another look at Questions

I think, on balance, the studies show more benefit than harm from providing clinical information. Specifically, as a PCP, I find that providing clinical information to the radiologist has the following benefits:

  *   The radiologist has the opportunity to check with me and discuss the study before it is done, often to suggest an alternative - and better - approach.
  *   If the radiologist knows what questions I am asking, it is easier for her to answer them.
 Ideally, the radiologist evaluates the technical adequacy, the radiologic findings and the clinical interpretation as separate and distinct processes.
One could worry about the same phenomenon of premature closure or framing bias in my work as a PCP when I take a history and then do an exam and get lab.  The history may lead me to examine some areas more carefully, not examine some areas at all, and weight different findings differently. the same is true when I order lab or other studies.  Using the already acquired information to help assign prior probabilities is anecessary, a feature rather than a bug.
It’s a difficult task and a source of error, though.

On 2016.03.06, at 9:13 PM, Peggy Zuckerman <peggyzuckerman at GMAIL.COM<mailto:peggyzuckerman at GMAIL.COM>> wrote:

Can there not be a creation of bias or limited analysis of the patient's images getting the clinical information.  Does the radiologist just focus on the ribs when the referring physician suspects a break, and ignore the less obvious spot on the pancreas?  In the interest of time, does the radiologist find himself confirming or eliminating that presumptive diagnoses and missing the 'incidental' finding?
Peggy of the kidney cancer world, with 30+% with such findings
Peggy Zuckerman<>

On Sun, Mar 6, 2016 at 2:36 PM, robert bell <0000000296e45ec4-dmarc-request at<mailto:0000000296e45ec4-dmarc-request at>> wrote:
Well said Tom,
 Another nonsense in medicine where errors abound mainly because of non-communication.
 There is a clear literature base in the assertion that readings improve with receipt of clinical information.
 What are we supposed to do with evidence based medicine?
 Ignore it?
 Rob Bell MD

On Mar 6, 2016, at 11:26 AM, Tom Benzoni <benzonit at GMAIL.COM<mailto:benzonit at GMAIL.COM>> wrote:

1980-2 I rotated with radiology.
I was taught to read my own films.
The rads wouldn't give us rad info without clinical info.
I still remember him saying "This is not a game. I'm a clinician. You use a stethoscope, I use an X Ray. We're both here with the same aim, to help the patient. You wouldn't send the patient to cardiology without telling them what the question is. Then don't do differently here."
I don't think there was a loss of productivity with these rounds; there is an efficiency in direct communication.
There is a clear literature base in the assertion that readings improve with receipt of clinical information.
tom benzoni

On Sun, Mar 6, 2016 at 9:35 AM, Elias Peter <pheski69 at<mailto:pheski69 at>> wrote:
I would echo that.
 In medical school (early 1970s, University of Rochester, NY) it was standard to make rounds that included the radiology department where medical students, residents, fellows and attendings would review the films and discuss the case with the radiologists. These were real-time mini-CPCs and very useful.
 In residency, a radiologist (usually Dr. John Juhl) came to our FP residency clinic once a week with the films that had been done on our out-patients during the prior week. We reviewed those films with him, our fellow residents, and the clinical faculty and discussed both the radiologic and clinical aspects of the cases.
 Collaborative learning.
 Peter Elias, MD

On 2016.03.06, at 9:58 AM, Koppel, Ross J <rkoppel at SAS.UPENN.EDU<mailto:rkoppel at SAS.UPENN.EDU>> wrote:

Bob Wachter says the time with the radiologist's was one of the best and the most informative of his medical school experience.

Ross Koppel, PhD, FACMI
Sociology Dept;  LDI Senior Fellow, Wharton; &
Affil Fac. Sch. of Medicine.  Chair, AMIA Clinical Information Systems Working Group.  Ph: 215 576 8221<tel:215%20576%208221>; Cell 215 518 0134<tel:215%20518%200134>

From: HM Epstein <hmepstein at GMAIL.COM<mailto:hmepstein at GMAIL.COM>>
Sent: Saturday, March 5, 2016 7:48 PM
Subject: Re: [IMPROVEDX] Another look at Questions

Hi Alan:
What were the Dean of Yale Medical School's objections when you approached him? What about other medical schools in the US? Does anyone understand  - or can you enumerate - the reasons for the resistance to improving the curriculum?


H.M. Epstein - Writing Portfolio :: Bio<><>
Bio. H.M. Epstein writes about parenting issues, politics and policy, physical and emotional healthcare, and passions both personal and professional.

Mobile: 914-522-2116<tel:914-522-2116>

Sent from my iPhone

 On Mar 5, 2016, at 8:14 AM, Alan Morris <Alan.Morris at IMAIL.ORG<mailto:Alan.Morris at>> wrote:
We, in the past, made daily (or more frequent) rounds with the radiologist.  We interpreted the images together.  I also attended post mortem examinations, and sometimes did dissection myself.
 We had much more time for thought and scholarly activities than is now allowed by the “business” of medicine.
 I was not able to convince the Dean of Yale Medical School to consider decision-support strategies to complement clinician cognitive limitations, when I attended my 50th Yale reunion.
Best wishes.

Alan H. Morris, M.D.
Professor of Medicine
Adjunct Prof. of Medical Informatics
University of Utah

Medical Director, Urban Central Region Pulmonary Function Laboratories
Pulmonary/Critical Care Division
Sorenson Heart & Lung Center - 6th Floor
Intermountain Medical Center
5121 South Cottonwood Street
Murray, Utah  84157-7000, USA

Office Phone: 801-507-4603<tel:801-507-4603>
Mobile Phone: 801-718-1283<tel:801-718-1283>

From: "Sanders, Lisa" <lisa.sanders at YALE.EDU<mailto:lisa.sanders at YALE.EDU>>
Reply-To: Society to Improve Diagnosis in Medicine <IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG<mailto:IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG>>, "Sanders, Lisa" <lisa.sanders at YALE.EDU<mailto:lisa.sanders at YALE.EDU>>
Date: Saturday, January 9, 2016 at 11:21
Subject: Re: [IMPROVEDX] Another look at Questions

I too lament the loss of collaboration between the internist or doctors of any stripe who is caring for the patient and the radiologist. In my most recent blog post: the patient went to three doctors who could not tell him what was wrong, despite the fact that he had the right tests ordered and that the radiologist had listed the diagnosis on his very short differential. It was an unusual disorder and he was an unusual subject so I can imagine that it was easy to discount such a diagnosis. But would you trust yourself to make a different diagnosis without at least consulting your own favorite radiologist? The last physician not only did not recognize the disorder demonstrated by the CTA but told the patient not to worry about it and to come back in a year to follow up. Fortunately he didn’t do that and got the intervention that was needed.
Why don’t we talk to the radiologists anymore?  Paying for that time – as suggested in the NAM report - might make a difference. I certainly hope so.

Lisa Sanders M.D.
Associate Professor
Yale School of Medicine

Clinician Educator
Yale Internal Medicine Primary Care Residency
Yale New Haven Hospital, St. Raphael’s Campus
1250 Chapel St.
New Haven, CT 06511
Office: 203-867-8117<tel:203-867-8117>
From: Leonard Berlin [mailto:lberlin at LIVE.COM]
Sent: Thursday, January 07, 2016 3:11 PM
Subject: Re: [IMPROVEDX] Another look at Questions

As a radiologist, I'd like to comment briefly on radiologic errors, which admittedly can be problematic.  It is the job of our radiology societies  to  actively work to reduce error rates, and they are trying.

But I'd like add the following: the radiologist has the images, and the clinician has the patient; establishing a diagnosis requires collaboration of both (not to mention results of lab and other tests).  If a clinician receive a radiology report that somehow doesn't make sense or fit into the clinician's pre-imaging DD, the clinician should not automatically accept the report.  Call the radiologist and say that somehow the imaging interpretation doesn't seem to fit: ask the rad to take another look at the images keeping in mind  the clinician's doubt. And/or, ask the radiologist to show it to one of his colleagues for a second opinion.   This is certainly not a solution to the overall problem of radiologic errors, but on the other hand every now and then such collaboration will mitigate an error before it harms the patient  and result in a correct diagnosis.

Sadly, in this day and age, radiologists don't communicate with ordering physicians as much as they did in the past, and as much as they should today.  That is unfortunate and a shame, but it is a fact.  But in the interest of the patient, there's no doubt that two (or sometimes even three) heads are better than one.

Lenny Berlin, MD, FACR
Skokie, IL

Date: Thu, 7 Jan 2016 08:53:37 -0800
From: Michael.H.Kanter at KP.ORG<mailto:Michael.H.Kanter at KP.ORG>
Subject: Re: [IMPROVEDX] Another look at Questions

In terms of a list of Do's and Dont's this could help.  I think the challenge will be to find Do's and Dont's that are simple enough to describe but at the same time can meaningfully reduce errors.    It is worth exploring though.

Michael Kanter, M.D., CPPS
Regional Medical Director of Quality & Clinical Analysis
Southern California Permanente Medical Group
(626) 405-5722<tel:%28626%29%20405-5722> (tie line 8+335)
THRIVE By Getting Regular Exercise

NOTICE TO RECIPIENT:  If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents.  If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them.  Thank you.

From:        robert bell <0000000296e45ec4-dmarc-request at LIST.IMPROVEDIAGNOSIS.ORG<mailto:0000000296e45ec4-dmarc-request at LIST.IMPROVEDIAGNOSIS.ORG>>
Date:        01/07/2016 07:21 AM
Subject:        Re: [IMPROVEDX] Another look at Questions

Dear Michael,

Some good points.

One of my thoughts is that we have the cart before the horse in first focussing on Errors in Diagnosis when something like 60% of errors are run of the mill standard errors. These such as communication problems that are associated with simple things like orders, wrong medications, computer problems, fatigue, education lack, and language problems. If we are truly interested in reducing the death toll and morbidity of Errors in Medicine as a whole should we not go for the jugular, identify the biggest problems, and tackle these collaboratively so that we get a significant reduction in the figures.

And that brings up a big issue that first we need to have figures to be able to say if we are making progress. And think of the revolution that will create when we get into litigation and confidentiality discussions. That big question seems to be, can we do anything meaningful in either Standard and Diagnostic Errors without good data?

I take your point that a large percentage of lab and x-ray reporting is related to diagnosis, but it is also related to monitoring over time - INRs, mammograms, and bone density come to mind. I do not know the monitoring percentage but would guess it is 10 - 20% of the total.  As these two specialty areas are so important to the whole of medicine it would seem that the consortium of specialty societies (that incidentally should have patient representation) could triage the main problems in these areas and collaboratively work on improving them. From what I hear the electronic patient record would be something that should be looked at immediately. If you do not look after the important basic foundations of medicine and the standard 60% of all errors that accounts for the biggest mortality I believe that you are never going to effectively make inroads into the Diagnostic Error piece of the pie.

I am sure that with the benefits of technology there are many opportunities to reduce/improve the number of errors in both Radiology and Laboratory pathology.  And yes, you are correct radiology is likely to help make accurate diagnoses sooner than an impression made with a stethoscope (incidentally I have often wondered how accurate, in certain hands, the stethoscope is and have questioned what is its contribution to errors - perhaps it should be triaged close to the top!). But the technological advances in radiology would place it in the forefront for improvement of error rates.

So this does not mean that we should not be focussing on Diagnostic Errors as we move forward, but strongly believe we should first be making sure that the main foundations to diagnosis are as solid as they can be.

Yes, specialists with telemedicine links to help in the review process would be a wonderful idea to help with error rates.  And also a wonderful teaching tool to both specialist and radiologist. Could well be tried out with a small study if it is not already being done.

One of my thoughts is that in medicine that is structured hierarchically we may not be asking enough questions. Particularly as technology and collaboration is slowly eroding that hierarchy. Questions are often less intrusive and more than anything are creative in group settings. Questions should be encouraged in all meetings if we wish to make quick progress.

Do you think that Do and Don’t lists are of value?

Thank you Michael,

Rob Bell

On Jan 6, 2016, at 4:54 PM, Michael.H.Kanter at KP.ORG<mailto:Michael.H.Kanter at KP.ORG> wrote:

interesting thoughts that you wrote below that are very thought provoking as usual.  Maybe I misunderstood but  you make the statement that " I felt that if there is a significant error rate in radiology reports and laboratory tests that this should be addressed first."
I personally consider error rates in radiology reports and lab tests within the scope of diagnostic errors and not sure why we would not address this quickly.   I do not believe that anywhere in the IOM report did they suggest otherwise.    In fact, radiology report errors really fall into the class of diagnostic errors related to visual errors and also occur in pathology reports and other areas of diagnosis where the diagnosis is made by visual interpretation of an image.  We recently published a method on how to decrease these for interpretation of retinal photos
Visual errors that are based on stored images (as opposed to looking at a skin lesion of a patient) have some advantages in terms of study and performance improvement including
1) they are often stored in data systems for easy retrieval and identification
2)  they can be retrospectively reviewed
3)  one can determine the incidence of diagnosing a disease if the image or image results are stored in a data system and use the variation in diagnosis as a measure of diagnostic error/variability as was done in the attached paper.
4)  Diagnosis based on visual images can be more easily studied because  the variable of data acquisition can be controlled better (that is everyone can look at the same image/data) which is very different than a physical exam or patient history.

My point is that it may be simpler to study errors due to visual interpretation of images than other types of diagnostic errors and this is something to consider addressing sooner rather than later.

I really like the idea of specialists societies getting involved in diagnostic errors.    Specialists  are likely to see such errors in their practice when referals are sent too late or with the wrong diagnosis.

Michael Kanter, M.D., CPPS
Regional Medical Director of Quality & Clinical Analysis
Southern California Permanente Medical Group
(626) 405-5722<tel:%28626%29%20405-5722> (tie line 8+335)
THRIVE By Getting Regular Exercise

NOTICE TO RECIPIENT:  If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents.  If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them.  Thank you.

From:        robert bell <0000000296e45ec4-dmarc-request at LIST.IMPROVEDIAGNOSIS.ORG<mailto:0000000296e45ec4-dmarc-request at LIST.IMPROVEDIAGNOSIS.ORG>>
Date:        01/02/2016 04:03 PM
Subject:        [IMPROVEDX] Another look at Questions

Dear All,

I would like to thank all the people on this list who kindly responded to my questions a few weeks back.

Since asking those questions I became aware that questions may be creative.

Does it follow then that statements are less creative. I then thought how does that come about?

A question I thought, stimulates listeners and readers to think of various solutions. So from one question you might get many solutions. With statements you would be more relying on the creativity of the writer presenting his or her various ideas and solutions. And these all together may not be great in number.

With regard to the medical profession as a whole and people in positions of power in the healthcare industry who can effect change, one can ask do they ask fewer questions for fear of being considered less knowledgeable? So, if this is true, almost by definition, if questions are creative, such decision makers are less likely to be broadly creative, at least initially? There are obvious exceptions. So a  question might be how to better take advantage of the crowd think?

This all makes an assumption that most things in medicine are complex and that many solutions to an issue may be necessary to arrive at a good solution.

The IOM has issued its statement/report on diagnostic errors and there are some good suggestions. I was pleased to see teamwork mentioned first in their solution. Which essentially is an extension of crowd think.

But the big question is where do we go from here? As you all know I would have preferred standard medical errors to be approached first before diagnostic errors. For example, I felt that if there is a significant error rate in radiology reports and laboratory tests that this should be addressed first. But assuming that we may have the cart before the horse, and I am not completely sure about that, what questions could be asked of the established consortium of speciality societies?

Over and above the big structural solutions, what comes to mind for me is that some of these societies have issued Do and Don’t lists for their members.

So my new question is, could those lists be looked at and the idea extended to all the specialty societies with the emphasis on reducing diagnostic error?

Also, with any statement(s) would it be good to consider asking some questions at the end to stimulate thought?

The best for 2016,

Robert M. Bell, M.D., Ph.C.

Robert M. Bell, M.D., Ph.C.
P.O. Box 3668
West Sedona, AZ  86340-3668




Visit the searchable archives or adjust your subscription at:
Moderator:David Meyers, Board Member, Society for Improving Diagnosis in Medicine

To learn more about SIDM visit:

To unsubscribe from the IMPROVEDX:
or click the following link: IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG


For additional information and subscription commands, visit:

http://LIST.IMPROVEDIAGNOSIS.ORG/ (with your password)

Visit the searchable archives or adjust your subscription at:

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

To unsubscribe from the IMPROVEDX list, click the following link:<br>
<a href="" target="_blank"></a>

HTML Version:
URL: <../attachments/20160323/257a26ef/attachment.html> ATTACHMENT:
Name: image001.png Type: image/png Size: 2137 bytes Desc: image001.png URL: <../attachments/20160323/257a26ef/attachment.png> ATTACHMENT:
Name: image002.png Type: image/png Size: 13674 bytes Desc: image002.png URL: <../attachments/20160323/257a26ef/attachment-0001.png>

More information about the Test mailing list