ORIGINAL ARTICLE
Optimization of Radiology Reports for Intensive Care Unit
Portable Chest Radiographs
Perceptions and Preferences of Radiologists and ICU Practitioners
Eduardo J. Mortani Barbosa, Jr, MD,* Marsha C. Lynch, MD,*
Curtis P. Langlotz, MD, PhD,w and Warren B. Gefter, MD*
Purpose: The aim of the study was to evaluate opinions and per-
ceptions of radiologists and referring practitioners regarding
reports of portable chest radiography (pCXR) obtained in the
intensive care unit (ICU).
Materials and Methods: A total of 1265 referring practitioners and
76 radiologists were invited to participate in 2 internet-based sur-
veys, containing 15 and 17 multiple choice questions, respectively,
similarly presented to both groups, utilizing a Likert scale or
multiple choices. Results were compared using the Fisher exact test
or w2 test.
Results: One hundred ninety-two referring practitioners and 63
radiologists answered the surveys, resulting in response rates of
15% and 83%. The majority of radiologists and referring practi-
tioners are satisfied with the quality of the reports; however, radi-
ologists and referring practitioners disagree about the reports’
clinical value and impact, the referring practitioners having a more
positive view. Both groups overwhelmingly agree that pertinent
clinical information is crucial for optimal image interpretation. The
2 groups differ in their preferences regarding report style and
information content, with radiologists strongly supporting concise
reports emphasizing temporal changes and major findings, whereas
referring practitioners prefer more complete, itemized structured
reports describing support devices in detail.
Conclusions: The results substantiate the perceived clinical value of
radiologist reports for pCXR, from the perspective of referring
practitioners. Nonetheless, there is disagreement regarding report
structure and content. Several issues were raised, offering oppor-
tunities for improvement, which may increase referring practi-
tioners’ satisfaction and positively impact patient outcomes. Any
strategy to implement standardized structured reports for pCXR
will have to satisfy referring practitioners’ needs while optimizing
radiologists’ efficiency, will have to be widely accepted, and will
have to fulfill the overarching goal of maximizing the value of
pCXR reports.
Key Words: intensive care unit, portable chest radiography report,
structured report template, clinical value, clinical utilization
(J Thorac Imaging 2016;31:43–48)
From the *Department of Radiology, University of Pennsylvania,
Philadelphia, PA; and wDepartment of Radiology, Stanford Uni-
versity, Stanford, CA.
The authors declare no conflicts of interest.
Correspondence to: Eduardo J. Mortani Barbosa, Jr, MD, Department
of Radiology of the University of Pennsylvania, 3400 Spruce Street,
Philadelphia, PA 19104 (e-mail: eduardo.barbosa@uphs.upenn.
edu).
Supplemental Digital Content is available for this article. Direct URL
citations appear in the printed text and are provided in the HTML
and PDF versions of this article on the journal’s Website, www.
thoracicimaging.com.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/RTI.0000000000000165
J Thorac Imaging  Volume 31, Number 1, January 2016
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
T he radiology report is the primary means of communi-
cation between the radiologists and the referring
practitioner, documenting the radiologist analysis, with
medical-legal and quality-of-care implications. In the
intensive care unit (ICU), the portable chest x-ray (pCXR)
is a major diagnostic tool.
The interpretation of pCXR is challenging because of
frequently suboptimal image quality, large number of
support devices, and multiple comorbidities. Moreover, the
radiologist is often pressed to provide a fast turnaround
time, increasing the likelihood of dictation errors and
reduced information content. Finally, pCXRs are often
interpreted with minimal or inadequate clinical informa-
tion. Most institutions have not adopted structured report
templates for pCXR, and unstructured narrative reports
predominate, with highly variable information content and
consequent implications in quality.1
Prior surveys have elicited the perspectives of referring
practitioners and radiologists regarding radiology report
quality and relevance. None, however, specifically targeted
ICU pCXR reporting. One of the first surveys of referring
practitioners’ needs and preferences2 proposed improve-
ments in report structure and communication of urgent
findings. A study of >800 CXR reports3 of inpatients
found high variability with respect to terminology and
substantial uncertainty regarding the diagnosis. Another
study4 found a strong preference for itemized, structured
reports, citing improved completeness and consistency as
major benefits. A large survey in the Netherlands5 found
that most referring practitioners and radiologists thought
that itemized structured reports are better suited to complex
examinations.
A pCXR report that does not address the ordering
referring practitioners’ needs, or is not available within a
reasonable timeframe, can lead to poor outcomes in crit-
ically ill ICU patients. We elicited both referring practi-
tioners’ and radiologists’ perspectives regarding the overall
clinical value and utilization of pCXR reports, in order to
better understand perceptions about quality and prefer-
ences regarding report format and content and identify
opportunities for improvement.
MATERIALS AND METHODS
The research proposal detailing the study objectives
and design was approved by the hospital Institutional
Review Board, which waived the requirement for informed
consent (HIPAA waiver) of the research subjects.
Two similar surveys were designed for ICU referring
practitioners and radiologists at 2 tertiary care hospitals in a
major academic health system in the Northeast United States.
www.thoracicimaging.com | 43
Mortani Barbosa et al J Thorac Imaging  Volume 31, Number 1, January 2016

ICU referring practitioners were defined as attending prac- RESULTS

titioners, housestaff, and midlevel providers caring for A total of 255 questionnaires were answered, either
patients in the ICU. An ICU attending was defined as any partially or completely, and utilized for analysis. Response
board-certified intensivist, internist, anesthesiologist, or other rates were computed for the entire group and also for each
specialist working in the ICU. ICU housestaff was defined as category. The range of response for each specific question
trainees (residents or fellows) rotating in the ICU. Midlevel was calculated. Table 1 provides details.
provider was defined as any nurse practitioner or physician Tables 2 and 3 depict all answers and statistical anal-
assistant. Radiologists were defined as attendings and yses, respectively, for Likert-type and non–Likert-type
housestaff. A radiology attending was defined as a board- questions, comparing referring practitioners (clinicians =
certified radiologist, including thoracic imaging subspecialists C) and radiologists (R). Please refer to Appendixes 1 and 2
and nonthoracic imagers who read pCXR as part of call for each question (Supplemental Digital Contents 1 and
duties. Radiology housestaff was defined as trainees (resi- 2, http://links.lww.com/JTI/A58 and http://links.lww.com/
dents or fellows) in the radiology department. JTI/A59).
Each survey was designed such that most questions Efficiency in pCXR reporting is a paramount concern
were identically or very similarly worded (semantically for radiologists. Thirty-one of 59 (52.5%) radiologists were
identical) for both groups. The referring practitioner survey satisfied with their efficiency in creating pCXR using the
contained 15 multiple-choice questions and 1 free text current unstructured format, and 25 of 58 (43.1%) thought
comment box. The radiologist survey contained 17 multi- structured reports would increase their efficiency.
ple-choice questions and 1 free text comment box. Appen- One hundred thirty-three of 165 (80.6%) referring
dixes 1 and 2 detail all questions (Supplemental Digital practitioners and 47 of 59 (79.7%) radiologists were sat-
Contents 1 and 2, http://links.lww.com/JTI/A58 and http:// isfied with the quality of pCXR reports. The difference
links.lww.com/JTI/A59). between the 2 groups was not statistically significant
The research was publicized through direct contact with (P = 0.213). However, 138 of 165 (83.6%) referring prac-
ICU referring practitioners and radiologists in leadership titioners thought that the pCXR report positively impacts
positions and through invitation e-mails, utilizing internal management and outcomes of most ICU patients, whereas
lists. Participation was voluntary, and confidentiality was only 25 of 59 (42.4%) radiologists concurred. The differ-
assured, given that all answers were anonymous and data ence was statistically significant (P < 0.0001). Both groups
would only be analyzed in aggregate. No patient-identifiable [105/165 (63.6%) referring practitioners and 41/59 (69.5%)
information was utilized at any point. The surveys were radiologists] agreed that the pCXR report often mentions
distributed through an online platform (Surveymonkey, important findings that the referring practitioners would
Portland, OR). The URL to the survey was included in the not have noted themselves. The difference was not statisti-
invitation e-mail. Two invitations were sent, approximately 3 cally significant (P = 1.000).
months apart, with 1 e-mail reminder in between. The majority of referring practitioners [105/165
The results were downloaded in spreadsheet format (63.6%)] answered that they always read the entire pCXR
(Microsoft Excel, Redmond, WA). For the Likert-type report, whereas 51/165 (30.9%) answered that they only
questions (Appendix 1, Supplemental Digital Content 1, read the report if they have a clinical question. Only 9/165
http://links.lww.com/JTI/A58), total agreement (agree (5.5%) referring practitioners answered that they never
completely + agree somewhat) and total disagreement read the report and rely on their own interpretation. When
(disagree somewhat + disagree completely) were computed. asked how much of the report they read, referring practi-
If >50% of the answers were in the total agreement or total tioners were nearly evenly split between answering that they
disagreement columns, then the expression on the statement read the entire report [69/165 (41.8%)] or that they read the
was considered “yes” or “no,” respectively. If >50% of the Impression section, and if they still have questions, then
answers were in the neutral column, the expression was
considered “neutral.” Lastly, if total agreement, total dis-
agreement, and neutral were all <50%, the expression on
the statement was considered “undecided.” TABLE 1. Response Rates
For the 5 questions that did not use Likert scales
(Appendix 2, Supplemental Digital Content 2, http://links. Total Total Response
lww.com/JTI/A59), a table was constructed to allow direct Sent Responses Rate (%)
comparison of the referring practitioners’ and radiologists’ Referring providers
answers, and majority answers (ie, any alternative comprising Attending physicians 73 37 50.7
>50% of responses) were considered the preferred answer. Nurse practitioners/ 25 13 52.0
Whenever there were >2 possible choices, if none of the physician assistants
Trainees—fellows/ 1167 142 12.2
alternatives reached the 50% threshold, then the expression
residents
on the statement was considered “undecided.” Free text Total for all referring 1265 192 15.2
comments were analyzed using qualitative techniques. providers
Statistical Analysis was performed using SAS 9 (SAS, Radiologists
Cary, NC). Questions not utilizing Likert scales were ana- Attending physicians 32 28 87.5
lyzed using w2 statistics. Questions utilizing Likert scales were Trainees—fellows/ 44 35 79.5
initially dichotomized by grouping the answers into 2 cate- residents
gories (total agreement = agree completely + agree some- Total for all radiologists 76 63 82.9
what; total disagreement = disagree completely + disagree Overall response rate 1341 255 19.0
Response rate per question, for answered questionnaires
somewhat) and thereafter analyzed using the w2 test or Fisher
Maximum 93.7
exact test for small sample sizes. P values <0.05 were con- Minimum 76.9
sidered to denote statistically significant differences.

44 | www.thoracicimaging.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Thorac Imaging  Volume 31, Number 1, January 2016 Optimization of Portable Chest Radiography Reports

TABLE 2. Results for Likert Type Questions

N (%)
Likert Agree Agree Disagree Disagree
Questions Completely Somewhat Neutral Somewhat Completely Result Concordance
Ra 13 (22.0) 18 (30.5) 13 (22.0) 12 (20.3) 3 (5.1) Yes Not
applicable
Rb 8 (13.8) 17 (29.3) 13 (22.4) 13 (22.4) 7 (12.1) Undecided Not
applicable
C1 57 (34.5) 76 (46.1) 24 (14.5) 6 (3.6) 2 (1.2) Yes Yes
R1 19 (32.2) 28 (47.5) 6 (10.2) 5 (8.5) 1 (1.7) Yes P = 0.213
C2 67 (40.6) 71 (43.0) 17 (10.3) 9 (5.5) 1 (0.6) Yes No
R2 10 (16.9) 15 (25.4) 17 (28.8) 11 (18.6) 6 (10.2) Undecided P < 0.0001
C3 24 (14.5) 81 (49.1) 39 (23.6) 18 (10.9) 3 (1.8) Yes Yes
R3 10 (16.9) 31 (52.5) 10 (16.9) 8 (13.6) 0 (0.0) Yes P = 1.000
C4 30 (19.2) 60 (38.5) 14 (9.0) 34 (21.8) 18 (11.5) Yes Yes
R4 20 (34.5) 15 (25.9) 5 (8.6) 14 (24.1) 4 (6.9) Yes P = 0.867
C5 29 (18.6) 60 (38.5) 37 (23.7) 24 (15.4) 6 (3.8) Yes Yes
R5 12 (20.7) 19 (32.8) 14 (24.1) 7 (12.1) 6 (10.3) Yes P = 0.689
C6 50 (32.1) 85 (54.5) 17 (10.9) 4 (2.6) 0 (0.0) Yes No
R6 4 (6.9) 22 (37.9) 17 (29.3) 12 (20.7) 3 (5.2) Undecided P < 0.0001
C7 28 (17.9) 57 (36.5) 32 (20.5) 34 (21.8) 5 (3.2) Yes No
R7 4 (6.9) 21 (36.2) 8 (13.8) 15 (25.9) 10 (17.2) Undecided P = 0.025
C8 98 (64.1) 37 (24.2) 9 (5.9) 8 (5.2) 1 (0.7) Yes Yes
R8 37 (64.9) 13 (22.8) 3 (5.3) 3 (5.3) 1 (1.8) Yes P = 0.753
C9 100 (65.4) 38 (24.8) 7 (4.6) 7 (4.6) 1 (0.7) Yes Yes
R9 44 (77.2) 9 (15.8) 2 (3.5) 2 (3.5) 0 (0.0) Yes P = 0.731
C 10 5 (3.3) 14 (9.2) 24 (15.7) 60 (39.2) 50 (32.7) No Yes
R 10 0 (0.0) 1 (1.8) 8 (14.0) 9 (15.8) 39 (68.4) No P = 0.016

they look for answers in the body of the report [75/165
(45.5%)].
Most referring practitioners [75/156 (48.1%)] preferred a
structured, itemized, complete report, organized by system;
57/156 (36.5%) preferred a concise report that only men-
tioned major findings and temporal changes. The majority of
radiologists [33/58 (56.9%)], in contrast, preferred a concise
report that only mentioned major findings and temporal
changes. Only 5/156 (3.2%) referring practitioners preferred
free text, unstructured reports compared with 11/58 (19.0%)
radiologists. The difference between the 2 groups, however,
did not reach statistical significance (P = 0.146).
Both groups [91/156 (58.3%) referring practitioners
and 44/58 (79.3%) radiologists] preferred a concise
description of temporal changes and major significant
findings instead of a detailed description of all findings
organized by system for follow-up pCXR reports. The
difference between the 2 groups was statistically significant
(P = 0.005), even though there was concordance, because
of the overwhelming support by radiologists of concise
reports. Nonetheless, when asked about a specific example
(clinical support devices), most referring practitioners [101/
156 (64.7%)] expressed their preference for a detailed
description of type and position of every line/tube/device in
every report, in contradistinction to most radiologists [44/
58 (75.9%)], who preferred that only new, changed, or
malpositioned lines/tubes/devices be mentioned. The dif-
ference was statistically significant (P < 0.0001).
Report length is not perceived to correlate with thor-
oughness of the radiologists’ assessment, as the vast majority of
referring practitioners [135/153 (88.2%)] and radiologists [50/57
(87.7%)] agreed that even if a report is short, the assumption
was that the radiologist reviewed the image thoroughly. The
difference was not statistically significant (P = 0.753).
We asked whether a global assessment score indicating
the need for intervention and the time frame for such (eg,
level of urgency) would provide additional clinical value,
without elaborating the details. Eighty-nine of 156 (57.1%)
referring practitioners and 31/58 (53.4%) radiologists sup-
ported this idea, whereas 30/156 (19.2%) referring practi-
tioners and 13/58 (22.4%) radiologists were against it. The
difference was not statistically significant (P = 0.689).
The vast majority of referring practitioners [135/156
(86.5%)] agreed that the language and style of pCXR
reports were clear, with only 4/156 (2.6%) disagreeing, but
radiologists were far more critical, with 26/58 (44.8%)
agreeing and 15/58 (25.9%) disagreeing. The difference was
statistically significant (P < 0.0001).
Both groups [138/153 (90.2%) referring practitioners
and 53/57 (93.0%) radiologists] demonstrated over-
whelming agreement with the concept that clinical infor-
mation must be provided in order for the radiologist to
provide an optimal interpretation of pCXR. The difference
was not statistically significant (P = 0.731).
We allowed free text comments to gain further insight.
Thirty-one of 192 (16.1%) of the referring practitioners and

Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. www.thoracicimaging.com | 45
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Mortani Barbosa et al J Thorac Imaging  Volume 31, Number 1, January 2016

question, inconsistency in format and information content,

TABLE 3. Results for Non-Likert Type Questions and laterality errors or failure to mention new lines/tubes/
Non-Likert Questions Clinicians Radiologists Concordance devices. Most timeliness concerns related to the need for a fast
C11, R11 production of pCXR reports, in the absence of which, in many
a 105 (63.6) 9 (15.3) No circumstances, referring practitioners would be forced to make
b 51 (30.9) 37 (62.7) P < 0.0001 clinical decisions on the basis of their own interpretation.
c 9 (5.5) 13 (22.0)
C12, R12
a 69 (41.8) 3 (5.1) No DISCUSSION
b 13 (7.9) 3 (5.1) P < 0.0001 Our results substantiate the perceived clinical value of
c 4 (2.4) 16 (27.1) radiologist reports for pCXR, from the perspective of
d 75 (45.5) 34 (57.6)
e 4 (2.4) 3 (5.1)
referring practitioners. It offers insight into how to struc-
C13, R13 ture the report and what kind of information to include, in
a 75 (48.1) 13 (22.4) Yes order to maximize referring practitioners’ satisfaction and
b 57 (36.5) 33 (56.9) P = 0.146 potentially positively impact patient outcomes. Radio-
c 5 (3.2) 11 (19.0) logists are highly concerned about efficiency, and generally
e 19 (12.2) 1 (1.7) prefer highly concise reports. Referring practitioners,
C14, R14 however, value completeness, thoroughness, and timeliness,
a 65 (41.7) 12 (20.7) No and at the same time prefer itemized structured reports. In
b 91 (58.3) 46 (79.3) P = 0.005 light of these findings, successful implementation of
C15, R15
standardized structured reports for pCXR requires aligning
a 101 (64.7) 14 (24.1) No
b 55 (35.3) 44 (75.9) P < 0.0001 the needs and expectations of all stakeholders.
Although a slim majority of radiologists are satisfied
with their efficiency in creating unstructured pCXR reports,
10/63 (15.9%) of the radiologists entered some text into this the group is undecided whether a structured report will, in
field. The comments were analyzed and classified according principle, increase their efficiency. It is important to empha-
to the main topic discussed and to determine whether the size that a specific structured report template was not pre-
tone was positive, neutral, or negative. sented; nonetheless, several questions addressed radiologists’
Among radiologists’ comments, 10% related to clin- preferences regarding report structure and information con-
ical information, 10% to clinical value, 50% to report tent. The majority of radiologists consistently prefer a report
structure, and 30% were not classifiable; 10% were pos- that is concise, focused on major findings and temporal
itive, 40% were neutral, and 50% were negative. A recur- changes, and only mentions new, changed, or malpositioned
rent issue among radiologists was that structured reports support devices instead of every support line and device, and
could hinder reporting efficiency by imposing constraints, they also prefer 1-phrase reports for follow-up pCXRs if
and that this problem was amplified by the very large there is no abnormality or change from the previous report.
number of daily pCXRs ordered. Additional aspects Referring practitioners, in contrast, tend to prefer complete
included the following: skepticism about referring practi- structured reports, itemized by system, and also prefer that all
tioners actually reading the reports; the idea that stand- support devices be described in every report. The reasons for
ardizing the way lines/tubes/devices are reported could be such discordance can be inferred from the analysis of free text
beneficial for patient care; the unmet need for better clinical comments. Radiologists, having to read a large number of
information to produce optimal reports; and the concept pCXRs within a short period of time, are very concerned
that lack of standardization in pCXR reporting may about efficiency and therefore tend to be against statements
undermine communication of important findings. that would increase the length of the reports. Referring
Among referring practitioners’ comments, 29.0% related practitioners, although generally appreciative of brevity,
to report structure, 25.8% to information technology (IT) express the need for easily accessible and complete informa-
limitations, 16.1% to report quality, 16.1% to timeliness of tion and are concerned that very concise reports may force
report creation and distribution, 6.5% to survey methodology, them to look at previous reports to obtain a more complete
3.2% to report utilization, and 3.2% to communication of representation of the findings, possibly disrupting continuity
results; 16.1% were positive, 38.7% were neutral, and 45.2% of care. Another concern is that very terse reports may create
were negative. A recurrent issue with respect to IT is that the the impression of lack of thoroughness from the radiologist.
current systems in place do not allow the referring practi- However, our results indicate that this is not true; both
tioners to optimally convey their clinical questions or concerns, groups strongly agree that there is no correlation between
and, as a result, many times these are ignored by the radiol- report length and thoroughness of the radiologist’s inter-
ogists. Opinions regarding report structure varied. Recurrent pretation. This finding differs from prior published work.5
aspects include the following: reports should be concise but The concept of a global assessment score was briefly
clinically relevant; clinical questions should be answered; a presented but not elaborated, and both groups demonstrated
differential diagnosis should be provided; and findings that support for it. The potential benefit of such a scoring system
impact management should be emphasized. A few referring would be to flag studies in which there are findings that need
practitioners expressed concern that a very concise report urgent or immediate attention by means of IT systems, facili-
stating “No change” would force them to look at prior tating communication and prompt intervention. Radiologists’
reports, and this finding could be detrimental to continuity of acceptance of such a system would be conditional on it being
care. Many referring practitioners expressed preference for a simple to use and not detrimental to dictation efficiency.
report structure containing IMPRESSION and COMMENTS There is discordance between the radiologists’ per-
rather than a single paragraph of free text. Opinions regarding ception of how referring practitioners utilize their reports
report quality focused on lack of answer to the clinical and what referring practitioners actually do. This

46 | www.thoracicimaging.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Thorac Imaging  Volume 31, Number 1, January 2016
discrepancy has relevant implications, as it suggests the
need to improve communication between the 2 groups to
achieve a better mutual understanding of each group’s
expectations, and also supports a structured report tem-
plate that separates the conclusion from the rest of the
report (eg, Impression separate from Findings), to facilitate
effective communication.6,7
We assessed report quality from multiple perspectives,
most importantly from the clinical value of pCXR, but also
from correctness and clarity of language. The majority of
both groups agree that pCXR reports provide value by
mentioning important issues the referring practitioners
might not have noticed had they analyzed the image with-
out the report. Nonetheless, when asked the more specific
question of whether the pCXR report positively impacts
management and clinical outcomes for ICU patients, there
is disagreement, with radiologists being skeptical and
undecided, whereas referring practitioners overwhelmingly
agreed. Whether the referring practitioners’ perception
translates to measurable differences in patient outcomes is a
question open to investigation, but it suggests that pCXR
reports are a fundamental component of the referring
practitioners’ arsenal to manage complex ICU patients.
In terms of report clarity and correctness, radiologists
were self-critical, being undecided regarding clarity and cor-
rectness. Referring practitioners had a more favorable opinion,
with an overwhelming majority agreeing that reports are
clearly written, and a small majority agreeing that they are
adequately proofread. This difference may arise due to the fact
that radiologists are exposed to a far greater number of reports
and therefore to a greater number of poorly written reports
with grammatical/typographical errors. In any case, the
responses indicate that there is opportunity for improvement.
Both groups demonstrate overwhelming agreement that
it is crucial for the radiologist to have access to relevant
clinical information to be able to provide an optimal report.
Nonetheless, this ideal scenario does not conform to reality.
Analysis of free text comments indicates that many referring
practitioners are frustrated with the inability to convey
relevant, tailored information to radiologists and by them
frequently ignoring clinical questions. Similarly, many radi-
ologists are frustrated with the incomplete, inaccurate, or
even incorrect patient history. Once again, communication is
the core issue. The solution may arise from better, more
flexible IT systems that allow better flow of information
without creating additional burden to either group.8,9
There are limitations to our study. Our overall response
rate was relatively low (19.0%) but not substantially different
from similar prior studies.5,10 However, it is important to
emphasize that this was skewed downward because of the
large number of trainees that we had to include in the
e-mailing list who do not rotate in the ICUs, and therefore
were much less likely to respond. We did not have access to
lists of trainees by specialty. Nonetheless, we were able to
obtain more customized lists for ICU referring practitioners
(excluding residents) and radiologists, which were totally
inclusive (all eligible practitioners at the participating hospi-
tals were invited). The response rate from ICU referring
practitioners (excluding residents) at 50.7% was higher than
most published studies, and the response rate from radiol-
ogists was very high at 87.5%, which is surprising given that
the invitation to the surveys was sent only twice. This sug-
gests that the topic is of great interest to radiologists who
report pCXR. Furthermore, the response rate from these
groups would suggest that nonrespondent bias, which is
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Optimization of Portable Chest Radiography Reports
inherent in any voluntary survey, may not play a major role
in extrapolating the responses to the overall group opinion,
particularly among radiologists. Nonetheless, we did not
have the resources to analyze nonrespondent data. Fur-
thermore, subgroup analysis comparing responses of
attending physicians and trainees was not feasible because of
the relatively small sample size. This interesting question may
be addressed in larger surveys in the future.
Another potential limitation is the generalizability of
the results, given that the all the respondents practice within
a major urban tertiary-care academic institution. It is rea-
sonable to assume that this diverse practitioner population
is representative of major urban centers in the United
States, but not necessarily of smaller community or rural
practices, or even urban centers in other countries.
An intriguing question is whether CXRs should be
reported at all. Prior studies11 have compared radiologists
and nonradiologists to assess their accuracy at interpreting
a variety of findings on chest radiography. The radiologists’
performance was substantially superior to other specialists.
Our data strongly corroborate that referring practitioners
perceive value in having a radiologist interpreting a pCXR,
and it suggests that the radiologist report may potentially
positively impact patient management and outcomes.
There is a common misconception that the most per-
formed imaging study worldwide, the chest radiograph, is a
“simple” examination, regarding imaging acquisition and
interpretation, which we believe is far from the truth.
Because all the information is present in a single image, its
interpretation may be more challenging than “complex”
cross-sectional imaging modalities. Moreover, this is
amplified in the ICU setting, wherein multiple comorbid-
ities and diagnoses are the rule, vast numbers of support
devices are commonly used, and imaging quality is fre-
quently suboptimal. Therefore, even though ICU pCXR
constitutes a small portion of the field of radiology, many
of the issues addressed in the surveys are applicable to other
modalities and clinical settings, although they are more
acutely relevant in the realm of pCXR reporting.
A radiology report is primarily a communication tool.
It is only relevant if it is able to convey the thought process
and concerns of the radiologist clearly, accurately, and in a
timely manner. Moreover, referring practitioners must be
able to act upon the report.
Our results allow better delineation of the challenges
and inform a prescription for the future.12–14 The first chal-
lenge is improving IT infrastructure to foster better com-
munication between referring practitioners and radiologists
and ensuring that both sides have the ability to seamlessly
enter and access relevant and accurate clinical information.
The second challenge is to create a structured report template
that is organized in such way as to optimize efficiency, sim-
plify report generation, guarantee the presence of crucial
information, and maximize clinical relevance and value.15–18
Our proposed template for pCXR, based on our survey
results, is a structured report with 3 sections: Impression,
Additional Comments, and Technique. The Impression
appears at the top, contains a fill-in field that can be dictated
as free text, should be highly concise, and should contain only
answers to clinical questions and/or clinically relevant find-
ings and diagnoses. Additional Comments contains all the
other findings under subheaders such as Lungs/Pleural
Spaces, Cardiac/Mediastinum, Skeleton, and Support Devi-
ces. It should not be redundant with the Impression section
and would contain fill-in fields that can be set to a default
www.thoracicimaging.com | 47
Mortani Barbosa et al
value, filled with free text or with standard macros. Ade-
quately positioned support devices, in particular, can be
described using standard macros, at the same time satisfying
referring practitioners’ needs and allowing radiologists to be
most efficient. Finally, Technique would contain a descrip-
tion of the radiographic technique, prior comparisons, and
clinical indication, elements that are necessary for billing
compliance. We advise against implementing strict
standardized report templates using pick-lists for the fields, as
this generally has an adverse effect on efficiency, whenever the
number of choices for the pick lists is large, as it often has to
be for proper pCXR interpretation.
We believe that in this era of diminishing reimburse-
ments streamlining the structure and content of ICU pCXR
reports on the basis of feedback from the referring practi-
tioner, can better serve ICU clinical needs and improve
referring practitioner satisfaction without compromising
radiologist efficiency in report creation, all desirable out-
comes with respect to the overarching goal of providing
high-quality patient care.
ACKNOWLEDGMENT
The authors would like to thank Yanly Wang, MS, for
her contributions on statistical analysis of the data.
REFERENCES
1. Hirvonen-Kari M, Sormaala MJ, Luoma K, et al. Quality of
chest radiograph reports. Acta Radiologica (Acta Radiol).
2014;55:926–931.
2. Clinger N, Hunter T, Hillman B. Radiology reporting: attitudes
of referring practitioners. Radiology. 1988;169:825–826.
3. Sobel J, Pearson M, Gross K, et al. Information content and
clarity of radiologists’ reports for chest radiography. Acad
Radiol. 1996;3:709–717.
4. Naik SS, Hanbidge A, Wilson SR. Radiology reports: examining
radiologist and referring practitioner preferences regarding style
and content. Am J Roentgenol. 2001;176:591–598.
48 | www.thoracicimaging.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Thorac Imaging  Volume 31, Number 1, January 2016
5. Bosmans JML, Weyler JJ, De Schepper AM, et al. The
radiology report as seen by radiologists and referring referring
practitioners: results of the COVER and ROVER surveys.
Radiology. 2011;259:184–195.
6. Dunnick NR, Langlotz CP. The radiology report of the future:
a summary of the 2007 Intersociety Conference. J Am Coll
Radiol. 2008;5:626–629.
7. Cohen MD. The radiology report of the future: the ignored
impression. J Am Coll Radiol. 2008;5:1017.
8. Sistrom CL, Langlotz CP. A framework for improving
radiology reporting. J Am Coll Radiol. 2005;2:159–167.
9. Garvey CJ, Connolly S. Radiology reporting—where does the
radiologist’s duty end? Lancet. 2006;367:443–445.
10. Johnson AJ, Ying J, Swan JS, et al. Improving the quality of
radiology reporting: a practitioner survey to define the target.
J Am Coll Radiol. 2004;1:497–505.
11. Mehrotra P, Bosemani V, Cox J. Do radiologists still need to
report chest x rays? Postgrad Med J. 2009;85:339–341.
12. Hall F. The radiology report of the future. Radiology.
2009;251:313–316.
13. Reiner B. Radiology report innovation: the antidote to medical
imaging commoditization. J Am Coll Radiol. 2012;9:455–457.
14. Reiner BI, Knight N, Siegel EL. Radiology reporting, past,
present, and future: the radiologist’s perspective. J Am Coll
Radiol. 2007;4:313–319.
15. Johnson AJ, Chen MYM, Zapadka ME, et al. Radiology
report clarity: a cohort study of structured reporting compared
with conventional dictation. J Am Coll Radiol. 2010;7:
501–506.
16. Hasegawa Y, Matsumura Y, Mihara N, et al. Development of
a system that generates structured reports for chest x-ray
radiography. Methods Inf Med. 2010;49:360–370.
17. Amorosa JK, Bramwit MP, Mohammed TL, et al. ACR
Appropriateness Criteria routine chest radiographs in intensive
care unit patients. J Am Coll Radiol. 2013;10:170–174.
18. Marcovici PA, Taylor GA. Journal club: structured radiology
reports are more complete and more effective than unstruc-
tured reports. Am J Roentgenol. 2014;203:1265–1271.