| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Commentary |
1 Department of Radiology, Rush Medical College, Chicago, IL; and Department of Radiology, Rush North Shore Medical Center, 9600 Gross Point Rd., Skokie, IL 60076.
Received April 24, 2006; accepted after revision May 8, 2006.
This article is a commentary on the preceding article by Reiner and
Siegel.
Keywords: radiology education • radiology reports
"Then you should say what you mean," The March Hare went on. "I do," Alice hastily replied; "At least—at least I mean what I say— That's the same thing, you know." "Not the same thing a bit!" said the Hatter. —Lewis Carroll Alice's Adventures in Wonderland [1]The newest computer can merely compound, at speed, the oldest problem in the relations between human beings, and in the end the communicator will be confronted with the old problem of what to say and how to say it. —Edward R. Morrow [2]
Those of us who studied plane geometry during our high school education learned that the logic of geometry provided the framework for logical thinking and sound reasoning. We learned to approach a problem by beginning with a hypothesis—a statement that is known or assumed—and then proposing a conclusion, for which a proof or analysis must be presented. Reasoning is the process by which we construct the proof or formulate the analysis.
Reasoning, however, can be faulty and lead to a poorly constructed proof that results in an incorrect conclusion. For example, as illustrated in one geometry textbook [3], a person discovers that his finger hurts if he touches a hot coffee pot. If he concludes that his finger will hurt anytime he touches a coffee pot, he is wrong, because it is not the coffee pot but rather the heat in it that causes the discomfort. If the person determines that his finger will hurt if he touches any hot vessel, his conclusion is correct. A proper analysis or proof is necessary to ensure that the conclusion is a true one.
In a Commentary published in this issue of the AJR, University of Maryland researchers Bruce Reiner and Eliot Siegel [4] begin with the hypothesis that traditional prose (free text) radiology reports have become characterized by disorganization, ambiguity, and inconsistency. It is a hypothesis with which, because it accurately identifies deficiencies that occasionally mar the clinical value of radiology reports, many observers, including myself, agree. Reiner and Siegel then proceed to propose a conclusion, supported by their proof, that because text reports have lost or are losing their communication value, they should be replaced by image-centric reports. I offer an alternate conclusion: If traditional text radiology reports are indeed losing their communication value, they should not be replaced, but rather improved by educating radiologists about how to sharpen their writing skills.
Criticisms of Traditional Radiology Reports
Many of the observations presented by Reiner and Siegel [4] as part of their proof appear to be valid: "Most radiologists continue to create reports in a manner strikingly similar to that of their predecessors practicing 100 years ago"; "Members of the referring clinician community cite poor report content and organization as the greatest problems with radiology reporting [5]"; "Current reporting processes are subject to multiple sources of failure and frustration, including the errors and delays associated with dictation, transcription, and delivery"; "Improper and/or poor communication is cited as one of the principal causes of radiology lawsuits"; and "The end product for both digital dictation and speech recognition...has been criticized by the larger community for disorganization, ambiguity, and inconsistency."
Let us examine more closely some of the observations referenced by Reiner and Siegel [4] to determine the specific nature of the criticisms levied at text reports. In addition to commenting on poor report content and organization, Naik et al. [5] found well over 90% of surveyed radiologists believed it was necessary for their radiology reports to describe the technique of an examination, whether contrast material was used, and the route of contrast administration. Seventy-six percent believed that informed consent should be mentioned in the report if it was obtained. In other words, Naik et al. found that radiologists believe that their reports should be more comprehensive. Lafortune et al. [6] and Clinger et al. [7], also referred to by Reiner and Siegel as being critical of the "limitations and inefficiencies" of text reporting, actually concluded that the written radiology report should, among other things, respond directly to the clinical questions asked and contain a conclusion and a description of findings. Langlotz [8], as noted by Reiner and Siegel, did editorialize about the inherent shortcomings of narrative radiology reports, such as use of ambiguous terms, but Langlotz also pointed out that these reports sometimes do not address key clinical questions.
Observations by researchers not mentioned in Reiner and Siegel's Commentary [4] are well worth noting. Eighty-three years ago Enfield [9] criticized radiologists who issued written radiology reports that "describe in detail all that the roentgenologist sees in the film but does not tell what he thinks about it, what conclusions he draws from it, and what it means to him." Enfield then exhorted radiologists to "give not only their opinion but also their method of arriving at that opinion." Rothman [10] has written that because radiologists are paid for using both "their eyes and their brains," a complete radiology report must include both sets of evaluations—not only a complete description of the abnormalities seen with the eyes of the radiologist, but also a discussion of the findings that are important to the radiologist's brain. Ronai [11] has emphasized that radiology reports should "grab the attention of, convey a message to, and elicit a response from the referring physician." In an article about dealing with vague radiology reports [12], I suggested that radiologists should not sign off a radiology report until the following questions are answered in a meaningful manner: What did I see on the radiographic study?; By what do I believe the findings are caused?; What do I want my referring physician to conclude from my interpretation?; and What do I suggest the referring physician does next?
Florida researchers Sistrom and Honeyman-Buck [13] compared free-text reports with structured reports that are essentially a computerized reporting product that uses a software interface to allow radiologists to choose relevant findings from a large and highly structured menu of possibilities. In some such systems, the available software allows the radiologist to produce phrases and sentences that may look like a typical narrative report, whereas other systems produce reports in which the results are organized under headings, similar to a laboratory report. These researchers found that structured and narrative text reports were similar in accuracy, efficiency, and reading time. Referring physicians surveyed by these researchers preferred structured reports over free-text formats, notwithstanding the fact that Sistrom and Honeyman-Buck found that one drawback to structured formats was that the free-text reports could better convey a "sense of severity and more acuity" than structured reports.
In a classic treatise on radiology reporting, Canadian radiologist Harald O. Stolberg [14] stressed the importance of written reports: Stolberg also pointed out that radiology reports must make clear to the reader that the reporting radiologist was aware of previous pertinent imaging reports and that a comparison with previous examinations was made. Differential diagnoses, a relevant discussion of the sensitivity and specificity of the radiologic findings, and the limitations of the specific radiologic examinations are also often important components of the final radiology report, Stolberg reminded us.
Judgments of clinical colleagues about radiology are increasingly made through these documents rather than through personal interaction because the radiology report is the only contact that we may have with many referring physicians. The only way to get their respect is by composing intelligent and well-written reports.
In their General Guide to Science Writing[15], Garratt and Mattinson stressed:
Writing is one of our indispensable means of communication. It helps us to formulate ideas, and provides us with a permanent record of them... When you have thought clearly, you have never had any difficulty in saying what you thought. And when you cannot express yourself, depend on it that you have nothing precise to express, and that what incommodes you is not the vain desire to express, but the vain desire to think more clearly.... Tables, figures and diagrams, collectively referred to as illustrations, should always be used where they will save words or make your argument clearer. However, you must make sure that your reader does not need to refer to them to follow the general flow of your arguments.
A common denominator permeates all of the observations referred to thus far: namely, a plea that radiologists include in their radiologic reports meaningful commentary that extends beyond simply describing radiologic findings, such as technical details about how the radiology examination was performed; limitations of the examinations; clinical implications of the radiologic findings; and, when appropriate, differential diagnoses, comparison with previous studies, and recommendations for additional studies, follow-up, or both. Clearly, the consensus of these researchers is that radiology reports should be more inclusive, not less.
Communicating Results
Reiner and Siegel [4] appropriately express concern about breakdowns in communication that both adversely impact patient care and are causative factors in up to 80% of malpractice lawsuits. They emphasize the many difficulties encountered by radiologists in attempting to directly communicate urgent or significant unexpected findings to referring physicians and point out that an increasing number of vendors now offer Web-based systems that track, archive, and verify referring physician responses to e-mail, voice communications, faxes, and text messaging. These concerns are valid and have been expressed by others [16]. Clearly, improvements in communication of radiology results are needed, whether they are text or image-centric, and digital systems appear to be the key for a solution. As Rogers [17] has suggested, the way out of this "reporting morass" lies in the digital world, with paperless electronic reports. "Once digitized reports are formulated, they can easily be distributed to wherever they may be required," writes Rogers, who continues:
Report macros—standard report formats for varying imaging studies—can be pre-pared and individualized to suit each radiologist, much like the standard reports that are now used by some practices. The report is done, the examination is completed, and the results are immediately available. An electronic report has been available on the monitor when and wherever the images are viewed by whoever is viewing them.
Coffee Pots, Roots, and Hieroglyphics
Let us return to the coffee pot, using it as a metaphor for text reports. My colleagues Reiner and Siegel and I have discovered that our fingers hurt when we touch a hot coffee pot. My two colleagues appear to have concluded that all coffee pots are too dangerous to touch and therefore should be discarded, an action that would, by the way, sadly condemn them to forever forsake drinking brewed coffee. My conclusion, however, is that only certain coffee pots are too dangerous to touch because they are too hot, and therefore all we need do is adjust their temperature. Happily I can continue to drink brewed coffee. So much for the metaphor. My real conclusion is that we must preserve written text radiology reports but with the proviso that the skill with which radiologists write these reports must be improved.
Reiner and Siegel [4] claim that text reports will eventually become obsolete and unnecessary and want us to return to our "image-centric roots." They propose that the report of the future be composed of "key images" annotated with highlighted pertinent findings. They suggest that "a well-codified set of icons and symbols" be developed to assist in the annotation. The interpreting radiologist could draw the "universal symbols" for cardiomegaly, pulmonary edema, and pleural effusion directly over the areas in question in the chest radiographic image, for example. They continue: "These symbols then become the equivalent of the text report, assuming the reader understands and can `decode' the symbols being used."
My colleagues [4] suggest that the same icons and symbols can be used by clinicians who want to confer with the interpreting radiologist:
The referring clinician might observe a nodular density... [that] the clinician could mark with a question mark. The radiologist receiving the query, in turn, would mark the area in question with the symbol for calcification, denoting costochondral calcification as the source of the nodular density. Although the radiologist would have the ability to consult with text (in written or spoken form), such consultation would be largely unnecessary, as the entire consultation will be streamlined through the electronic transmission of annotated images using a standardized gesture/symbol language.
The use of symbols or pictures to represent words or sounds is not a novel idea, of course. Now called "hieroglyphics," they were used five thousand years ago as the system of writing in ancient Egypt. Today's dictionaries define hieroglyphics as a "picture or symbol with a hidden meaning; illegible or undecipherable writing" [18]. In fact, Egyptian hieroglyphics were totally undecipherable until 1799, when the Rosetta stone, a slab of black basalt discovered at the mouth of the Nile River, provided the key to interpretation of Egyptian hieroglyphics. The image-centric report consisting of icons and symbols envisioned by Reiner and Siegel [4] presumably would be accompanied by a modern "Rosetta stone"-type of decoding device.
Would Image-Centric Reports Be Practical?
The current version of the ACR Practice Guideline for Communication of Diagnostic Imaging Findings [19] mandates that "an official interpretation (final report) shall be generated and archived following any examination, procedure, or officially requested consultation." Included as components of the report are items such as Let us keep in mind, then, that of all these components, only the Findings section can be replaced by an image-centric report.
relevant clinical information, procedures and materials, findings, potential limitations, clinical issues, comparison studies and reports, and an Impression that includes a precise diagnosis, a differential diagnosis, and follow-up or additional diagnostic studies to clarify or confirm the Impression.
The word "interpret," is defined as "to set forth the meaning of something not plain or clear; to clarify the meaning of; elucidate; to expound the significance of or bring out the meaning in a revealing manner" [18]. Would the exclusive use of symbols and icons, without using free text, qualify as an "interpretation"?
A radiology report is an integral part of a patient's medical record and, as stated by a New Jersey Appeals Court [20], the medical community must be able to rely on those records in the continuing and future care of that patient. The medical record then, and specifically the radiology report included in it, must be understandable not only to the radiologist and the referring physician, but also to other physicians, patients and their families, attorneys, judges, and jurors, any or all of whom may have occasion to review that record at any time in the future.
Radiologists' Writing Skills Must Improve
Reiner and Siegel [4] contend that the radiology community has "not done a good job in addressing reporting and communication shortfalls. We should aggressively look to technology and innovation for assistance in bringing these processes into the twenty-first century." They suggest that "the concept of supporting imaging data with text reports may eventually become obsolete and unnecessary," and "if a radiologist can identify `key images' and annotate these with pertinent findings highlighted, then the traditional text report may no longer be necessary." Radiologists should "be freed of the text report," and "interpret, report, and communicate via images alone." This, they believe, is the correct conclusion to the hypothesis that our current method of documenting our interpretations of radiologic examinations by preparing written reports is rife with problems. My colleagues' hypothesis is correct, but their conclusion is not. The correct conclusion is that we must educate radiologists how to write better reports.
This education should ideally begin when "Johnny," the future Mr. Radiologist, and "Mary," the future Ms. Radiologist, are elementary school students. Unfortunately, that ideal may no longer exist in many localities because it seems that over the past several decades the quality of the educational product found in many of our nation's schools has declined. All too often one of the "3 Rs" of education—"wRiting"—has fallen away from the other two, "Reading," and "aRithmetic." In too many situations, the "R" signifying writing has been replaced by a "B" signifying "Blogging." In today's world of electronic communication, the art of composing and applying the written word is rapidly diminishing. With people so easily communicating in phrases, half-sentences, buzz words, incomplete paragraphs, and sound- and word-bytes by means of e-mails, faxes, and messages left on telephone voice mail, formal written communication has deteriorated and continues to do so.
Forty years ago, Canadian radiologists Dunbar and Usher [21] wrote:
Must the radiologist give the impression of being not only ungrammatical but posi-tively antipathetic to the rules of grammar and syntax? We were fascinated in the reports we collected by the verbless clauses, misplaced modifiers, dangling participles, ineffective split infinitives, amorphous sentence structure, and tortured syntax. We can do better than this, unless we accept the idea that all grammar belongs in the grammar school or that when we entered medical school our synapses became inadequate for our syntax.
In a study of radiology reporting, radiologist Richard Gunderman [22] and his associates at Indiana University made the following observation:
The radiology report represents perhaps the radiologist's most conspicuous and enduring product. A radiologist may be the world's greatest lesion detector or possess the most encyclopedic knowledge base of differential diagnoses, but if he is poor at communicating his thoughts to referring physicians, the value of his input will be less than optimal.
Reiner and Siegel [4] characterize the text report as "the albatross of radiology"—that is, an obvious handicap, constant burden, or heavy cross to bear [18]. I would not characterize text reports in such harsh terms. Text reports are admittedly weak and limited in the hands of some radiologists whose writing abilities are less than optimal. Educating radiologists how to write effectively and accurately in their early formative years—and repeatedly through refresher courses during their careers if necessary— can correct this weakness.
Summary
Reiner and Siegel [4] propose that we should "take radiology reporting to the next level" by converting from text reports to reports containing images and symbols. In his book, The Art of Readable Writing, originally published 57 years ago, Rudolf Flesch [23] specifically addressed the question of whether words are necessary and whether they can be replaced by symbols and pictures:
Easy reading is difficult to write, so difficult that most people would rather try anything else but write when they face a job of simple explanation. They escape from words into pictures, symbols, graphs, charts, diagrams. The idea that you can explain things without explaining them in words is pure superstition. A favorite proverb of picture-and-diagram lovers is "one picture is worth more than a thousand words." It simply isn't so. Try to teach people with a picture and you may find that you need a thousand words to tell them exactly what to look at and why.
Anything pictorial or graphic does help as long as there is enough text to back it up. I do not mean captions; I mean that the running text must tell the reader what the illustration means, how we should look at it, and why. Nothing is self-explanatory—it's up to the writer to explain it. And the writer has to do it in words.
If radiologists write in text form what they mean in a manner understood by referring physician readers, then the readers will experience an epiphany—that is, "a perception or intuitive grasp of the essential nature or meaning of something" [18]. Should referring physician readers fail to understand what the radiologist has written, they will be confounded by an enigma—that is, "something obscure, hard to understand or explain" [18].
Radiologists who in their written reports want to say what they mean must focus on words and sentences, not images. This does not preclude radiologists' use of images, with or without annotations, to enhance a text report to assist referring physician readers in understanding what the radiologists want them to know. In the final analysis, the job of radiologists is to impart to their referring physicians their superior knowledge about all aspects of the specific radiologic examination. There is only one way to impart that knowledge: with words.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Pochaczevsky Image-Annotated Radiographic Reports Am. J. Roentgenol., June 1, 2007; 188(6): W576 - W576. [Full Text] [PDF]
|
|
|
|
|
|
R. J. Stanley What Does "Good Medicine" Mean? Am. J. Roentgenol., November 1, 2006; 187(5): 1145 - 1145. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
