DICOM Structured Reporting book reprinted and is now back in stock
David Clunie
The printed edition of my DICOM Structured Reporting book
is now back in stock after having been reprinted.
It is available to order on-line using a credit card via
a secure connection at the publisher's web site:
http://www.pixelmed.com/srbook.html
david
Dee Csipo
That is great news. So people are actually implementing this! keep it
coming.
dee
;-D
William Horton, M.D.
I looked through your online version of the book, and it seem to be pretty
complete, and well written, but as a practicing radiologist I am curious.
Other than attaching significant images to a report, what is the practical
use of SR in actual clinical practice?
WH
Dee Csipo
Well it depends a lot on the practice itself, but structured reports
allow uniformity in reporting therefore making the report analysis in a
large practice, mostly in a hospital, easily automated. There are
several ongoing efforts in turning unstructured reports into structured
reports. The most significant advance achieved by the structured report
ing work is the encoding of the relationships of findings and evidence
in the reports, and the establishemnt of the observation contexts. SR
does not encode a presentation format it is more akin to a dabase schema
that reports need to fit into. SR can be as simple as the codification
of a current paper form or as complex as a completely automated OBGYN
CAD report of a fetal development study including measurements, images,
notes references to previous studies and reports associated with the
mother of the fetus, identifying her medical condition and the effects
of those findings on the fetuses current condition.
A structured report may conatain data in a computable format that lets
say a lab system may extract and use it to calibrate equipement used to
further analyze data.
The simple report indicating the significant images is an awfully
simplified use of SR. Sr is an integration vehicle that can be used to
exchange composite information in a comprehensive format, adhering to
institution established SR templates. The possibilities are endless.
William Horton, M.D.
Thanks for your comments.
SR sounds good on paper, but the benefits you cite sound more theoretical
than real. I think SR will end up being more of a "talking dog". Sure,
there may be some benefit to having a database that encodes every reported
finding, but only in an academic setting when doing a retrospective review.
I am sure that a large number of retrospective review type papers could be
generated in a short time this way (publish or perish, get it?)
As a clinically oriented radiologist i dont feel the need to overly concern
myself with the internal representation of report data. There are three
issues that stand out as being important however.
1. The report creation interface. Sure you can probably generate a few
sample reports in the lab, but imagine what it will be like to encode 18,000
of these puppies every year (figure 1600 hrs/man*year). SR would have to be
at worst no more time consuming than dictating a simple report.
2. The final textual representation of the report. The printed report is
all that stands as the official medical record.
3. The stability of the medical record. There exists the possibility that
if changes are made to the textual representations of codes or to the text
generation software that the printed form of a SR could change after it was
approved by the radiologist. This would obviously be unacceptable.
As far as relating findings and evidence to the report and to a presentation
context, this is probably not as useful as it sounds. Personally, I make it
a point not to mark findings on an examination. If a referring clinician
needs help in interpreting the images, he can confer with me, or one of my
collegues. This maintains personal communication, which is in danger of
being lost in the era of electronic distribution of images.
In short, for SR to be useful to a clinically oriented radiologist, it would
have to offer some concrete benefit. What I would like to know is what
these benefits are. (hint, saving me time or money is good)
"Dee Csipo" <dcs...@charm.net> wrote in message
news:3BF9DCFB...@charm.net...
Dee Csipo
Just some notes...
William Horton, M.D. wrote:
> Hi, Dee
>
> Thanks for your comments.
>
> SR sounds good on paper, but the benefits you cite sound more theoretical
> than real. I think SR will end up being more of a "talking dog". Sure,
> there may be some benefit to having a database that encodes every reported
> finding, but only in an academic setting when doing a retrospective review.
> I am sure that a large number of retrospective review type papers could be
> generated in a short time this way (publish or perish, get it?)
Yeap it would sure help there. It also helps a lot when you are dealing
with diagnostic information a very large hospital system like the
military or the VA.
>
> As a clinically oriented radiologist i dont feel the need to overly concern
> myself with the internal representation of report data. There are three
> issues that stand out as being important however.
>
> 1. The report creation interface. Sure you can probably generate a few
> sample reports in the lab, but imagine what it will be like to encode 18,000
> of these puppies every year (figure 1600 hrs/man*year). SR would have to be
> at worst no more time consuming than dictating a simple report.
SR will not fly unless it is automatically generated, or by a
transcriptionist or by a voice recognition system in conjunction with
the recording of the actions taken on the diagnostic workstation and
encoding the presentation state used during diagnosis and attaching it
to the SR document generated. Then the only step you the radiologist
have to take is to review it and and put your eJohn Hancock on it :-)
>
> 2. The final textual representation of the report. The printed report is
> all that stands as the official medical record.
Yes it is true today, but are you prepared to make the same statement
five years from now? The federal government has a very large ongoing
EMR effort involving several branches of the government. Once the
project is deployed the only document in existence will be electronic.
Once the government establishes the electronic report standards the
private sector is forced to follow suit due to the large scale migration
of patients in and out of the federal system.
>
> 3. The stability of the medical record. There exists the possibility that
> if changes are made to the textual representations of codes or to the text
> generation software that the printed form of a SR could change after it was
> approved by the radiologist. This would obviously be unacceptable.
No SR is a composite object as such can not change. Amendments create
new instances of SR IODs which reference the original amended SR object.
The IOD is identified by a 64 character Global UID. Once it is
created it is a non revocable contract just like a paper document.
DICOM also has security features that can make the modifications of the
IOD detectable.
>
> As far as relating findings and evidence to the report and to a presentation
> context, this is probably not as useful as it sounds. Personally, I make it
> a point not to mark findings on an examination. If a referring clinician
> needs help in interpreting the images, he can confer with me, or one of my
> collegues. This maintains personal communication, which is in danger of
> being lost in the era of electronic distribution of images.
It is not a replacement of the good doctor's personal touch ;-) But as
we move along the face to face engineering meetings are being replaced
by video conferencing augmented by the use of tools like netmeeting,
etc. Think of an SR document as a useful tool to enhance the
communication with the referring physicians, sort of the difference
between the sketch on a napkin vs. a powerpoint presentation.
>
> In short, for SR to be useful to a clinically oriented radiologist, it would
> have to offer some concrete benefit. What I would like to know is what
> these benefits are. (hint, saving me time or money is good)
A "normal" exhaustive SR report could be generated with a single mouse
click without the reading radiologist uttering a single world. For a
more complicated use case consider the generation of a finding of a
malignant tumor using a diagnostic Ws. The radiologist opens the study.
Selects the image where the tumor is most visible draws a circle
around the tumor. Clicks the mouse a menu pops up, he selects the
option of malignant tumor. closes the study.
Based on those actions the WS can create a report that identifies the
location of the malignant tumor, the size of the malignant tumor, if it
is a CT study the average density of the tumor, its extent in 3D space,
It can look up any related lab reports and verify from a DB if there
are any indications in the patients blood work for a malignant tumor,
and i could continue on. In seconds or worst case in a minute, all
these findings are presented to the reading radiologist. The headings
clearly identify all the findings, conclusions and the authors of all
those concepts, so you the reading radiologist can make a snap decision
whether you need to take a look at them or let them stand and stay in
the report. then the report is signed and verified by the reading
physician. At the same moment it is made available in an un-alterable
form to the rest of the practice. Since everything is concise,
everything about the patient's condition is presented the possibility of
missing an important fact, therefore making a misdiagnosis, is reduced,
therefore the number of litigations is reduced = you save money. The
large hospital systems regularly settle lawsuits if some of the original
diagnostic information is not available. They can justify putting in an
EMR system based on a single lawsuit that was avoided.
dee
;-D
Doug Sluis
I agree with your response.
I would emphasize that SR is all about clinical productivity
though until the applications that use SR are on the market,
SR may look like a "talking dog."
Dr. Horton does bring up a potential issue that I don't think you addressed.
> > 3. The stability of the medical record. There exists the possibility
that
> > if changes are made to the textual representations of codes or to the
text
> > generation software that the printed form of a SR could change after it
was
> > approved by the radiologist. This would obviously be unacceptable.
Your reply about the stability of the SR object is valid but
Dr. Horton's raises the issue of the rendering of the report.
Indeed, differing SR applications and application versions will not
identically display this information.
DICOM does not provide a way to convey the report presentation.
It seems that this is still an open question about whether this
is a significant issue and whether it should be in the scope of DICOM.
Any thoughts?
- Doug
"Dee Csipo" <dcs...@charm.net> wrote in message
news:3BFDE404...@charm.net...
William Horton, M.D.
I guess my question could be summed up as "what is the problem that
Structured reporting is designed to solve?" I viewed some of the works in
progress at the IHE display at the RSNA meeting this week. It seems that the
focus there is interoperability, and the structured reports are limited to
'simple image' reports.
I am not sure if this is a good thing, but certanly illustrating a report
with some significant images has a High Tech and Modern feel, and may have
some Marketing Potential. It is nice to see that different vendors could
understand and render each others reports, but the utility is limited.
As for your other comments
"Dee Csipo" <dcs...@charm.net> wrote in message
news:3BFDE404...@charm.net...
> Hi Dr. Horton,
>
> Just some notes...
>
> William Horton, M.D. wrote:
>
> > Hi, Dee
> >
> > Thanks for your comments.
> >
> > SR sounds good on paper, but the benefits you cite sound more
theoretical
> > than real. I think SR will end up being more of a "talking dog". Sure,
> > there may be some benefit to having a database that encodes every
reported
> > finding, but only in an academic setting when doing a retrospective
review.
> > I am sure that a large number of retrospective review type papers could
be
> > generated in a short time this way (publish or perish, get it?)
>
>
> Yeap it would sure help there. It also helps a lot when you are dealing
> with diagnostic information a very large hospital system like the
> military or the VA.
>
??? How does this help in a large institution any more than in a small one.
Big or small, care is delivered by individual professionals to individual
patients. you are still dealing with one patient at a time.
>
> >
> > As a clinically oriented radiologist i dont feel the need to overly
concern
> > myself with the internal representation of report data. There are three
> > issues that stand out as being important however.
> >
> > 1. The report creation interface. Sure you can probably generate a few
> > sample reports in the lab, but imagine what it will be like to encode
18,000
> > of these puppies every year (figure 1600 hrs/man*year). SR would have to
be
> > at worst no more time consuming than dictating a simple report.
>
>
> SR will not fly unless it is automatically generated, or by a
> transcriptionist or by a voice recognition system in conjunction with
> the recording of the actions taken on the diagnostic workstation and
> encoding the presentation state used during diagnosis and attaching it
> to the SR document generated. Then the only step you the radiologist
> have to take is to review it and and put your eJohn Hancock on it :-)
why routinely record the steps on the workstation, or the presentation state
used during diagnosis? This information is useful to nobody and may be a
potential source of liability.
>
>
> >
> > 2. The final textual representation of the report. The printed report
is
> > all that stands as the official medical record.
>
>
> Yes it is true today, but are you prepared to make the same statement
> five years from now? The federal government has a very large ongoing
> EMR effort involving several branches of the government. Once the
> project is deployed the only document in existence will be electronic.
> Once the government establishes the electronic report standards the
> private sector is forced to follow suit due to the large scale migration
> of patients in and out of the federal system.
>
lets keep this out of the hands of the federal government. Government
involvement is not a Good Thing. (look at the results of MQSA). Regardles of
the form of the report, printed or electronic, it is important that the TEXT
be preserved. The underlying symantic network, and IOD's and all that are
not important, its what I SAID, not what I MEANT.
>
> >
> > 3. The stability of the medical record. There exists the possibility
that
> > if changes are made to the textual representations of codes or to the
text
> > generation software that the printed form of a SR could change after it
was
> > approved by the radiologist. This would obviously be unacceptable.
>
>
> No SR is a composite object as such can not change. Amendments create
> new instances of SR IODs which reference the original amended SR object.
> The IOD is identified by a 64 character Global UID. Once it is
> created it is a non revocable contract just like a paper document.
> DICOM also has security features that can make the modifications of the
> IOD detectable.
Whats wrong with a good old ASCII text file?
Nice scenario, but beyond the one click normal report it only works like
this in PR videos.
Tim Davies
Tim Davies
Zion Wiseman
news:LF%J7.30729$gQ1.12...@news1.elmhst1.il.home.com...
William,
Like _any_ part of DICOM standard SR itself has no value for doctors :-)
It gains its value just when it used for integration of equipment from
different manufactures.
I cannot give an example form radiology but here is one from the cath lab
world:
- during the procedure some data (blood pressure, medications, etc.) is
recorded
by hemodynamic system;
- later on cardiologist may do some analysis on the images (stenosis, left
valve) on
cath worksation.
If these two pieces of equipment talking different languages it could be a
challenge
to merge this data automatically.
As a result final report generation becomes time consuming.
Obviously, to make things work some time-saving applications should
be developed.
IMHO doctors need applications utilising SR, not SR definitions themselves.
Like with cars - drivers don't need standard bolts and screwdrivers - their
mechanics do. And just when you have standard bolts and screws you
can go with your car to virtually any mechanic, not sticking with your
dealship.
Dee Csipo
No disagreement on my part. As both you and i know SR is an exchange
and not a presentation standard. I am with you on not knowing if DICOM
should be in the business of defining how reports should be presnted, I
think it is more in the courtyards of IHE.
I think the presentation issues should be left to the EMR systems and
that should be the competitve advantage that should make one buy one
system over another.
As for report management, David's book has a pretty good chapter on
managing reports using SR, and he ventures out into some of the issues
of HIS/EMR consistency.
dee
;-D
Dee Csipo
I think it is more appropriate if i do not go into some of the gory
details, but in the spirit of a good argument.... I think the emerging
applications that will utilize SR as report exchange format will prove
the usefulness of SR. Until then there is nothing wrong with the OBX
segment of an HL7 ORU message.
Too bad we had not had the chance to discuss this in person during the show.
;-D
William Horton, M.D.
Perhaps I misunderstand the meaning of 'reporting' in SR. As I see it
reporting is a human activity that involves human judgement in interpreting
[test results, waveforms, images, etc], and generating a concise written
summary of the findings. In a practical sense all that is required in a
'structured report' is 1. the patient demographics 2. a reference to the
'Current Requested Procedure' 3. The text of the physicians report. I see a
lot of extra stuff in the standard that could potentially go into a SR, and
frankly some of it is scary.
As far as whether doctors need to understand the underlying mechanism of SR:
IMHO there should be some practical imput from physicians into the
development of SR applications. It is also important for physicians to
understand the implications of using the various systems. For example a
Radiologist should be aware of the recording of an evidence tree, and its
relationship to the official medical record (i.e. what will the record show
if this were to end up in a court of law), and the side effects of various
actions made during reading an exam. (Just because I measure something
doesn't mean I want that measurment recorded in the SR).
There are vast differences between different medical practices, especially
academic vs private practice, and large vs small groups. The developers of
SR systems should solicit imput from all user groups. True, not every
radiologist is capable of understanding the standards underlying PACS, but
those of us that are need to keep the developers honest.
"Zion Wiseman" <zwis...@sinderion.org> wrote in message
news:JGXN7.138102$WW.87...@bgtnsc05-news.ops.worldnet.att.net...
William Horton, M.D.
I do appreciate a good argument, that is true. For now I think the simpler
implimentations will win out. I think that we will see SR appear in much the
way as DICOM has. The underlying solutions are generally proprietary, but
if required for exchange purposes, a DICOM wrapper can be applied.
Too bad we didnt meet at the show. Maybe next year. If you are going to be
in the Chicago area sometime let me know, it would be fun to discuss this
face to face.
WH
"Dee Csipo" <dcs...@charm.net> wrote in message
news:3C091AEC...@charm.net...