| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Medical Imaging, Toronto General Hospital, University Health Network, 200 Elizabeth St., Toronto, Ontario M5G 2C4, Canada.
Received April 5, 2000;
accepted after revision August 29, 2000.
Address correspondence to S. R. Wilson.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. The study was conducted in two parts. The first part was a retrospective audit of existing medical imaging prose reports to determine their content. The second part comprised a questionnaire containing three mock clinical scenarios. Three pairs of reports were provided for each scenario, with only essential information in the first pair, some optional information in the second, and the most complete report in the third. Each pair consisted of a prose and itemized report with identical content. Participants ranked reports by preference and were asked specific questions regarding report content. The questionnaires were mailed to referring clinicians and administered during an interactive forum to staff radiologists, radiology fellows, and radiology residents.
RESULTS. The audit of existing reports showed a wide variation in all fields with consistency limited to a given radiologist. Responses to the questionnaire showed that, in general, a majority of radiologists and referring clinicians preferred itemized reports. The itemized report with the most detailed content was ranked highest of all three scenarios.
CONCLUSION. Prose reports foster a lack of standardization of content among different radiologists. Itemized reports facilitate complete documentation of information and measurements and are more popular with both radiologists and referring clinicians.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The advent of new reporting technologies such as computer-generated itemized and voice-recognition reporting systems [1] has led us to reassess the merits of traditional prose dictated reports. The advantages of these new technologies are manifold and include the following: rapid report generation, production of an organized and legible report, attractive presentation, bypass of transcription, immediate verification, and potential immediate delivery of results via fax or electronic mail. Potential disadvantages of a computer-generated reporting system include the requirement of a rudimentary level of computer literacy. Typing skills, for reports of cases of complex abnormalities, are of benefit to augment the basic report. Interpretation errors and the need for a quiet working environment are drawbacks to the use of voice recognition. Nonetheless, the potential decrease in report completion time necessitates more than a cursory evaluation of these technologies.
With a recent hospital merger of multiple sites including our hospital, a move toward an efficient and consistent reporting system has become essential. Critical to this decision-making process, but unknown, were the preferences regarding radiology reports of both the providers (radiologists) and the receivers (referring clinicians). We hypothesized that computer-generated itemized reports would be acceptable to both groups in lieu of prose dictated reports. With this hypothesis in mind, we designed a study to determine their preferences and opinions specifically regarding report style and content.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Part I
The first part was a retrospective audit of 272 randomly chosen reports,
which were dictated by staff radiologists, to evaluate content. The reports
were from sonographic examinations, CT studies, mammography, barium studies,
and unenhanced radiography performed during the preceding year. Each report
was evaluated for the presence of the clinical indication for the study and a
mention of a comparison study (or lack thereof), pertinent negative findings,
and the radiologist's opinion and recommendations regarding further studies or
course of action.
Part II
The second part consisted of a questionnaire administered in two forms: by
mail to 184 referring clinicians to the Section of Ultrasound at The Toronto
General Hospital and in an interactive forum to staff radiologists, radiology
fellows, and radiology residents. The questionnaire comprised three mock
clinical scenarios of various complexities often encountered in a busy
sonography department. Each scenario contained three pairs of sonography
reports, with each pair describing the same patient (Tables
1,2,3,4,5,6).
The two reports in each pair were identical in content but differed in style
(either prose or itemized). The content of the reports varied between the
pairs: the first pair contained only essential information, the second pair
contained additional optional information, and the third pair had the most
detailed content. Participants were asked to select the preferred style of
report for each pair, answer specific questions regarding the content of the
reports, and rank all six reports in each scenario in order of preference.
|
|
|
|
|
|
The first clinical scenario described a 43-year-old man complaining of right upper quadrant pain; no abnormalities were shown by sonography (Tables 1 and 2). The second scenario was a 61-year-old man undergoing routine sonographic follow-up of an abdominal aneurysm; the study revealed a new unsuspected renal mass (Tables 3 and 4). The third scenario consisted of a 27-year-old woman with a painless pelvic mass. Sonographic examination showed a large cystic mass with benign features; the morphology of the mass was consistent with an endometrioma (Tables 5 and 6).
The first section of the clinician questionnaire addressed demographics: clinical specialty, number of years in practice, and affiliation with a teaching hospital. Questions regarding patient referral practices and report turnover time were posed. After reading the clinical scenarios, participants were asked which style of report they preferred in general. They were also asked when they would prefer prose reports (four options were provided: always, for normal examinations only, for abnormal examinations only, or never), when they would prefer itemized reports, and whether they would prefer itemized reports with the abnormal findings grouped separately from the normal findings.
At the interactive forum, participants were first asked to provide demographic information (number of years in practice, areas of interest in radiology) and rate their degree of computer literacy and typing skills on a scale from 1 (none) to 5 (excellent). After a brief introduction of the study, participants were shown slides of the sonography reports from the three clinical scenarios and posed the same related questions. They were then asked specific questions about both the style and content of all types of radiology reports. Regarding itemized reports, participants were asked for which imaging modalities, if any, these reports are appropriate. Regarding report style, they were asked which style of report was preferred and why. Participants were then asked if an itemized reporting system, a voice-recognition reporting system, or a combination of both was available would they use it. Regarding the report content, participants were asked if it was ever appropriate to have "normal examination" or "no significant interval change" as the complete report. They were asked if it is necessary to describe the technique of an examination and, if so, under what circumstances. The necessity of mentioning the use of a contrast agent and the route of administration was also queried. If informed consent was obtained for a procedure, participants were asked whether this information should be mentioned in the report. It was also asked whether it is necessary to document in the report the risks and complications discussed with the patient before informed consent was obtained and, if so, when. At the end of the slide presentation, time was allotted for questions and concerns otherwise not addressed.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Part II: Questionnaire
Clinician demographics and general information.—Of the 184
questionnaires, 95 were returned (52% response rate) from clinicians
representing a variety of specialties (42 internists, 26 surgeons, 10
emergency physicians, seven clinicians practicing obstetrics-gynecology, six
family practitioners, two critical care specialists, one medical geneticist,
and one neonatologist). Among the internists, seven practiced general internal
medicine, whereas the remaining 35 subspecialized in one of the following:
cardiology, endocrinology, gastroenterology, infectious diseases, geriatrics,
nephrology, neurology, or respirology. The surgeons who responded practiced
general surgery, transplantation, cardiovascular surgery, otolaryngology,
urology, orthopedics, or thoracic or vascular surgery. The number of years in
practice ranged between 1 and 40 years, with a mean of 14.1 years.
Seventy-five physicians had urban-based practices, one had a community-based
practice, and 17 had both types. Two clinicians did not specify. All
respondents worked in some capacity at a teaching hospital.
Referral practices.—With regard to patient referral for abdominal, pelvic, or abdominal-pelvic sonographic examinations, 53 respondents (56%) referred between one and five patients per week for such examinations. Eighteen clinicians (19%) referred none, 13 (14%) referred from five to 10 patients, and seven (7%) referred 10-20 patients. Only four clinicians (4%) referred more than 20 patients per week.
A majority of respondents (56 clinicians [59%]) received between one and five sonography reports per week. Fourteen (15%) received none, 12 (13%) received from five to 10 reports, and eight (8%) received 10-20 reports. Five clinicians (5%) received more than 20 reports per week. More than half of the respondents (54 clinicians [57%]) had previously received a computer-generated itemized report from our department.
Report completion time.—Twelve participants (13%) received a verified report in their office in 1-2 days; all were satisfied with this completion time. Many clinicians (41 respondents [43%]) received their reports in 3 days to 1 week. Of these clinicians, only 24 (59%) thought that this was a satisfactory completion time. Sixteen physicians (17%) received reports in 1-2 weeks, and 10 (11%) stated it took more than 2 weeks. Physicians' satisfaction with these longer completion times was, not surprisingly, low (three clinicians [19%] and none satisfied, respectively). The remaining clinicians (16 [17%]) had no response, and of these, five clinicians commented on their satisfaction with preliminary reports being provided the same day of the examination.
Although 43 respondents (45%) were satisfied with their respective completion times, 35 physicians (37%) were not (17 clinicians [18%] had no response). Some (40 clinicians [42%]) considered a 1- to 2-day completion time to be reasonable, 18 (19%) preferred to receive a report on the same day as the examination, and 31 (33%) thought 3 days to 1 week was reasonable. However, when specifically asked how valuable immediate receipt of a typed report was, a majority of clinicians (52 participants [55%]) stated it was "extremely valuable." Only five (5%) thought it was not valuable at all.
Participants were also asked about ways to improve report completion time. Suggestions included delivering results immediately via fax or electronic mail, providing a report for the patient to bring to the clinician's office, and generating reports faster.
Radiologist demographics and general information.—The forum was attended by a total of 25 staff radiologists, radiology fellows, and radiology residents. Sonography, body imaging, chest radiology, genitourinary radiology, musculoskeletal radiology, neuroradiology, gastrointestinal radiology, interventional radiology, oncologic radiology, and nuclear medicine were the areas of interest of participating staff radiologists. The number of years in practice (of staff radiologists) ranged from 2 to 30 years with a mean of 14.5 years. There was marked variability in the degree of computer literacy and the level of typing skills of the participants. Most had an average to above-average degree of computer literacy. Only five radiologists admitted to poor typing skills; the remaining radiologists rated their typing skills as average, above average, or excellent.
Results from the Clinical Scenarios
Report style.—The computer-generated itemized style was the
preferred style of report for most clinicians
(Table 7). This preference was
seen for each pair in all three scenarios regardless of report content. Unlike
the options given on the clinicians' questionnaire, the option of "no
preference" (in addition to "prose" and
"itemized") was given to the participants of the radiologist forum
when asked to choose their preferred style of report for each pair. Although a
majority of radiologists chose the itemized style, a significant number liked
the prose format. This preference was seen for all three scenarios. As the
content of the reports increased, there was a tendency toward choosing the
itemized report style (Table
7).
|
Report content.—Both clinicians and radiologists were asked
specific questions about each scenario regarding report content. All
respondents did not answer all questions. No statistically significant
differences were evident between the responses of the two groups (chisquare
test of independence, 1 degree of freedom, 
=0.01). Most radiologists
and clinicians believed that the clinical indication, scan quality,
descriptive details, measurements, the radiologist's opinion and
recommendations for further studies, pertinent negative findings, and the date
of comparison study should be included in the reports
(Table 8). A point of
contention with both clinicians and radiologists dealt with the issue of a
radiologist's recommendations regarding patient referral. Fourteen
radiologists (56%) and 47 clinicians (51%) believed that this practice was
appropriate.
|
Approximately a quarter of clinicians and radiologists (25% and 24%, respectively) valued measurements of normal organs. Fifty-five clinicians (60%) and 19 radiologists (76%) wanted measurements of organs in the area of interest. Most wanted measurements of abnormal organs (78 clinicians [85%] and 23 radiologists [92%]). The responses of the two groups regarding measurements of incidental findings were also comparable. Thirty-nine clinicians and 11 radiologists (44% in both instances) desired the measurements of incidental insignificant findings. More than half the clinicians (54 clinicians [61%]) and 16 radiologists (64%) wanted incidental insignificant findings in the organ of interest measured. An overwhelming 86 clinicians (97%) and 24 radiologists (96%) valued measurements of incidental but significant findings.
Ranking of reports.—When clinicians were asked to rank the six reports from each scenario in order of preference, the itemized report with detailed content was the most preferred (75 clinicians [81%] in the first scenario, 68 [76%] in the second scenario, and 78 [86%] in the third). When radiologists were asked the same questions, a smaller majority also chose the itemized report with detailed content as the most preferred (13 radiologists [52%] in the first scenario, 17 [68%] in the second scenario, and 14 [56%] in the third scenario). The prose report with limited content was least preferred by both clinicians and radiologists for all the scenarios.
Style preference overall.—After the clinical scenarios,
participants were asked which style of report they preferred overall. Most
clinicians (73 respondents [86%]) chose the itemized style, whereas only 10
(12%) selected the prose format. Ten clinicians (11%) had no response (all,
however, had chosen the itemized format as their preferred style in all three
mock scenarios), and two clinicians stated no preference (although this option
was not provided). Overall 16 radiologists (64%) preferred the itemized report
style. Only five radiologists (2%) preferred the prose style, and four had no
preference. There was no statistical difference between radiologists and
clinicians with regard to their preference for either the prose or itemized
style (chi-squared test of independence, 1 degree of freedom, =
0.01).
Reasons cited for choosing an itemized report included the ease of both reading the report and extracting information from it, its focused and organized structure, and the ability to readily discern what was specifically examined. "Maintenance of context" and "narrative flow" were among the reasons stated for choosing the prose format. The preference for a particular report style was independent of the examination results (i.e., a study with normal findings versus a study with abnormal findings). Interestingly, 22 radiologists (88%) stated they would use an itemized reporting system if it were available, 23 (24%) would use a voice-recognition system if available, and 21 (84%) would use a combination of the two.
Structure of the itemized report.—The itemized reports in the questionnaire were constructed as a list of two columns, with the examined organs on the left and the sonographic findings on the right. Abnormal findings were not highlighted in any way. The last question of our survey addressed whether a change in this report structure would affect clinician preference. Specifically, we asked whether clinicians would prefer itemized reports with the abnormal findings grouped separately from the normal findings; 31 clinicians (37%) answered yes. However, a larger percentage (36 clinicians [43%]) preferred that the reports be structured like they were in the scenarios.
Additional Questions Posed to Radiologists
A number of questions regarding report content were posed to the
radiologists. When asked if it was ever appropriate to have "normal
examination" as the complete report, only nine radiologists (36%)
agreed. Sixteen radiologists (64%) thought it appropriate to have "no
significant interval change" as the complete report. Sonography was not
considered to be the only imaging technique appropriate for itemized reports;
CT and MR imaging were commonly mentioned as other techniques for which such
reports could be incorporated.
The majority of radiologists thought it was necessary to describe the technique of an examination, whether contrast material was used, and the route of contrast administration (24 radiologists [96%], 24 [96%], and 23 [92%], respectively). Nineteen participants (76%) believed that informed consent should be mentioned in the report if it was obtained. Only seven (28%) thought it necessary to document in the report the risks and complications discussed before obtaining informed consent.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Gagliardi [10] stressed the value of the standardized report. Our results show a strong preference for computer-generated itemized reports by both the referring clinicians from our institution and our radiology colleagues, including those both in established practice and in training. Report appearance, completeness, legibility, and the structured format are among the most important cited advantages of this method. Incorporation of fields to cover not only the relevant anatomy but also the indication for the study, the scan quality, dates of comparison studies, and the radiologist's opinion can result in more consistent inclusion of all relevant data. Our audit of existing reports showed the inconsistencies that appear with the use of traditional prose reports. In spite of the disclosed preference of our radiology colleagues to include information such as the clinical indication, use of a comparison study, and radiologist's opinion in the report, these data were frequently missing in their own prose reports. We recognize that the audit is small and includes a small number of reports from many radiologists. Nonetheless, we believe that in dictated prose reports pertinent data are frequently omitted. Furthermore, if specific organs are not mentioned in the report, then the clinician is unsure whether the organs were examined. This would explain the preference by clinicians in all scenarios for reports with more detail, regardless of whether there was complex disease. McLoughlin et al. [6] made similar observations.
Itemized report generation is structured in our system using a decision tree, with a variable number of optional fields on each branch. For example, in an abdominal sonography report, the selection of the "kidney" field would result in a subdivision into two branches—"left" and "right"—that then would further subdivide into "size," followed by fields for additional comments. If a field was selected and filled in, the findings were included in the report. Each field had certain standard options that could be set to the user's preference; if a field was not chosen, it could default to "normal," disappear entirely, or require the input of information (either typed using a keyboard or, for those with poor typing skills, by voice recognition). For certain fields, the user may be provided with standard options. Under scan quality, for example, possible options would be a choice of "excellent," "good," "satisfactory," "poor," or "nondiagnostic." Alternatively, the system may be structured to have an automatic default to the most likely choice.
As with any new technology we have found that with experience a radiologist soon becomes adept at using the itemized reporting system, which is user-friendly. A concern raised during the interactive forum dealt with the potential of decreased efficiency because of poor typing skills resulting in longer reporting times. In reality, we have discovered that typing information in the chosen fields is not difficult regardless of the typing skills of the user. We believe that any initial increased reporting time during the learning period is more than amply compensated for by the advantages of an immediately verified printed report. In a large department like ours, the use of a report template aids a busy radiologist and facilitates a prompt but complete report, satisfying both the patient and the referring clinician.
Although initially introduced on a trial basis, the itemized reporting system has found a growing number of users in our department. It has long been our practice that every patient leaving our department takes a preliminary report (usually handwritten) to give to their referring clinician. These preliminary reports have now been replaced with verified itemized reports, which are greatly appreciated by our clinical colleagues. These reports are part of the patient's permanent hospital record and are printed either to be placed in an inpatient's chart before he or she leaves the department or to accompany outpatients who are scheduled to see their referring clinician after their sonographic examination.
The results of our study alone justify the continued use of the itemized system. However, we are discovering other advantages with this type of reporting package. Our current picture archiving and communications system (PACS) network enables both radiologists and clinicians to view current radiologic examinations (sonography, unenhanced radiography, CT, MR imaging) from a variety of departments in our hospital. Our next step is to incorporate the radiology report with the corresponding examination so that the clinician can view not only the images but also the radiologist's interpretation at the same time. The use of the itemized reporting system, for which the report is created and stored on a computer database, lends itself to this scenario becoming reality in the near future.
With this study, we have attempted to document not only clinician preferences regarding the style and content of radiology reports but also those of our radiology colleagues. In the process, we have tried to show the merits of itemized reporting systems in accommodating these preferences. With the merger of multiple institutions including our own, we believe that a computer-generated itemized reporting system has the potential not only to provide the efficiency and consistency required by such restructuring but also to satisfy an expanding pool of referring clinicians. We anticipate that any future audits with this type of system in place would, without question, reveal improved and more complete documentation of information.
The current state of medical imaging facilitates the transmission of images via a computer network over a widespread distribution not only to areas in radiology but also to hospital wards, emergency departments, operating rooms, clinicians' offices, and even physicians' homes. Radiology reports remain the major means of communicating the trained and professional opinion of radiologists who interpret an examination to the referring clinician. Our intent to incorporate a computer-generated itemized and complete report to accompany the images would meet the objectives of both our referring clinicians and radiologists.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Reiner and E. Siegel Radiology reporting: returning to our image-centric roots. Am. J. Roentgenol., November 1, 2006; 187(5): 1151 - 1155. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Berlin Replacing traditional text radiology reports with image-centric reports: a shift from epiphany to enigma? Am. J. Roentgenol., November 1, 2006; 187(5): 1156 - 1159. [Full Text] [PDF]
|
|
|
|
|
|
S. Ondategui-Parra, S. M. Erturk, and P. R. Ros Survey of the use of quality indicators in academic radiology departments. Am. J. Roentgenol., November 1, 2006; 187(5): W451 - W455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. L. Sistrom and J. Honeyman-Buck Free Text Versus Structured Format: Information Transfer Efficiency of Radiology Reports Am. J. Roentgenol., September 1, 2005; 185(3): 804 - 812. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. P. Langlotz Automatic Structuring of Radiology Reports: Harbinger of a Second Information Revolution in Radiology Radiology, July 1, 2002; 224(1): 5 - 7. [Full Text]
|
|
|
|
|
|
L. Berlin Relying on the Radiologist Am. J. Roentgenol., July 1, 2002; 179(1): 43 - 46. [Full Text] [PDF]
|
|
|
|
|
|
L. F. Rogers Information Transfer: Radiology Reports Am. J. Roentgenol., March 1, 2001; 176(3): 573 - 573. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
