See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8267044

How to Write an Effective Discussion

Article in Respiratory Care · November 2004

DOI: 10.4187/respcare.11435 · Source: PubMed

CITATIONS READS

119 79,629

1 author:

Dean Hess
Massachusetts General Hospital
356 PUBLICATIONS 17,533 CITATIONS

SEE PROFILE

All content following this page was uploaded by Dean Hess on 13 December 2013.

The user has requested enhancement of the downloaded file.

 How to Write an Effective Discussion
Dean R Hess PhD RRT FAARC

Introduction
Elements to Include in the Discussion
State the Major Findings of the Study
Explain the Meaning of the Findings and Why the Findings Are
Important
Relate the Findings to Those of Similar Studies
Consider Alternative Explanations of the Findings
State the Clinical Relevance of the Findings
Acknowledge the Study’s Limitations
Make Suggestions for Further Research
Give the “Take-Home Message” in the Form of a Conclusion
Things to Avoid When Writing the Discussion
Overinterpretation of the Results
Unwarranted Speculation
Inflating the Importance of the Findings
Tangential Issues
The “Bully Pulpit”
Conclusions That Are Not Supported by the Data
Summary

Explaining the meaning of the results to the reader is the purpose of the discussion section of a research
paper. There are elements of the discussion that should be included and other things that should be
avoided. Always write the discussion for the reader; remember that the focus should be to help the
reader understand the study and that the highlight should be on the study data. Key words: publishing;
writing; manuscripts, medical; communication. [Respir Care 2004;49(10):1238 –1241. © 2004 Daedalus
Enterprises]

Introduction ducted the statistical analysis and grouped the summary

results into table and graphs. But what does it mean? Ex-
You have carefully written the hypothesis. You have plaining the meaning of the results to the reader is the
designed the study and collected the data. You have con- purpose of the discussion section. Although the discussion
comes at the end of the paper, you should be thinking
about what you will write in the discussion section from
the moment that the study is conceived. Questions that you
Dean R Hess PhD RRT FAARC is affiliated with the Department of
Respiratory Care, Massachusetts General Hospital, and Harvard Medical
will develop in the discussion should be considered from
School, Boston, Massachusetts.

Dean R Hess PhD RRT FAARC presented a version of this article at the
RESPIRATORY CARE Journal symposium, “Anatomy of a Research Paper: Correspondence: Dean R Hess PhD RRT FAARC, Respiratory Care,
Science Writing 101,” at the 48th International Respiratory Congress, Ellison 401, Massachusetts General Hospital, 55 Fruit Street, Boston MA
held October 5–8, 2002, in Tampa, Florida. 02114. E-mail: [email protected].

1238 RESPIRATORY CARE • OCTOBER 2004 VOL 49 NO 10

 HOW TO WRITE AN EFFECTIVE DISCUSSION

Table 1. Elements to Include in the Discussion Explain the Meaning of the Findings and Why the
Findings Are Important
State the study’s major findings
Explain the meaning and importance of the findings
Relate the findings to those of similar studies No one has thought as long and as hard about your study
Consider alternative explanations of the findings as you have. As the person who conceived, designed, and
State the clinical relevance of the findings conducted the study, the meaning of the results and their
Acknowledge the study’s limitations importance seem obvious to you. However, they might not
Make suggestions for further research be so clear for the person reading your paper for the first
time. One of the purposes of the discussion is to explain
the meaning of the findings and why they are important,
Table 2. Things to Avoid in the Discussion without appearing arrogant, condescending, or patroniz-
ing. After reading the discussion section, you want the
Overpresentation of the results
reader to think, “That makes perfect sense. Why hadn’t I
Unwarranted speculation
thought of that?” Even if your study findings are provoc-
Inflation of the importance of the findings
Tangential issues
ative, you do not want to force the reader to go through the
The “bully pulpit” paper multiple times to figure out what it means; most
Conclusions that are not supported by the data readers will not go to that effort and your findings will be
Inclusion of the “take-home message”; save this for the conclusions overlooked, disregarded, and forgotten.
section
Relate the Findings to Those of Similar Studies

the study’s outset. Why is the study important? How does No study is so novel and with such a restricted focus
this study relate to previous studies? What are the limita- that it has no relation to other previously published papers.
tions of the study design? There are elements of the dis- The discussion section should relate your study findings to
cussion that should be included and other things that should those of other studies. Questions raised by previous studies
be avoided (Tables 1 and 2). Most important, always write may have served as the motivation for your study. The
the discussion for the reader; the discussion is not a forum findings of other studies may support your findings, which
for you to impress others with your knowledge of the strengthens the importance of your study results. Stoller et
subject. You should be trying to convince the reader of the al3 discussed their study results in the context of a previous
merits of the study results. study by others: “Our finding that changing in-line suction
catheters less frequently is associated with lower cost and
no higher incidence of VAP replicates the findings of a
Elements to Include in the Discussion randomized controlled trial conducted by Kollef et al, upon
which our amended policy was based.” It is also important
State the Major Findings of the Study to point out how your study differs from other similar
studies. An example can again be drawn from Stoller et
The discussion should begin with a statement of the al:3 “Certainly, differences in the specific criteria used to
major findings of the study. This should be the very first define VAP could contribute to the rate differences be-
paragraph in the discussion. It should be a direct, declar- tween the present study and that of Kollef et al. For ex-
ative, and succinct proclamation of the study results. How- ample, comparison of the criteria for nosocomial pneumo-
ever, it should not include data or reference to the study nia in our study with the criteria used by Kollef et al shows
design. Several examples illustrate the point. In a paper by similar component features but different rating schemes to
Anton et al1 the discussion begins with the sentence, “Our establish the diagnosis.”
results confirm that these nasal and full-face masks are
similarly efficient over 15 min of NPPV with COPD pa- Consider Alternative Explanations of the Findings
tients recovering from acute hypercapnic respiratory fail-
ure.” This clearly states the most important finding of that Despite efforts to remain objective and to maintain equi-
study. Fluck et al2 began the discussion section of their poise, it is easy to consider only those explanations that fit
paper with the sentence, “Our findings suggest that ambi- your bias. It is important to remember that the purpose of
ent light has no statistically significant effect on SpO2 read- research is to discover and not to prove. It is easy to fall
ings and that ambient light’s effect on SpO2 is clinically into the trap of designing the study to prove your bias
unimportant.” That is a good example of a direct, declar- rather than to discover the truth. When writing the discus-
ative, and succinct proclamation of the study results. sion section, it is important to carefully consider all pos-

RESPIRATORY CARE • OCTOBER 2004 VOL 49 NO 10 1239

 HOW TO WRITE AN EFFECTIVE DISCUSSION

sible explanations for the study results, rather than just message should be the first sentence of your conclusions
those that fit your biases. section. In some journals the conclusions section is a para-
graph or subsection at the end of the discussion, whereas
State the Clinical Relevance of the Findings other journals (RESPIRATORY CARE, for instance) require a
separate conclusions section. The conclusions section may
The reason we conduct studies is usually to improve the
also provide suggestions for practice change, if appro-
care of our patients. Thus it is important to cast the find-
priate. An example of a well-written conclusion comes
ings of your study in the context of clinical practice. For
from a study by Apostolopoulou et al,6 who wrote: “VAP
which patients do the results apply and for which do they
is a common infection and certain interventions might af-
not apply? Experimental studies conducted in the labora-
fect the incidence of VAP. ICU clinicians should be aware
tory usually do not involve human subjects, but the results
of the risk factors for VAP, which could prove useful in
may have clinical implications, which should be stated. A
identifying patients at high risk for VAP and modifying
paper by Swart et al4 gives an example of a laboratory
patient care to minimize the risk of VAP, such as avoiding
study, the clinical relevance of which is overtly stated:
unnecessary bronchoscopy or modulating enteral feeding.”
“The clinically important measurements, for both screen-
ing and monitoring, are predominantly FEV1 and FVC,
and the Spirospec and Masterlab 4.0 showed excellent Things to Avoid When Writing the Discussion
correlation (r ⫽ 0.99) and very good limits of agreement
for FEV1 and FVC. For FEV1 and FVC the Spirospec and Overinterpretation of the Results
the Masterlab 4.0 could be used interchangeably.”4
It is easy to inflate the interpretation of the results. Be
Acknowledge the Study’s Limitations careful that your interpretation of the results does not go
beyond what is supported by the data. The data are the
All studies have limitations. Unfortunately, the limita- data: nothing more, nothing less.
tions of some studies are fatal flaws that preclude publi-
cation. However, even the best studies in the most prestigious Unwarranted Speculation
journals have limitations. It is far better for you to identify
and acknowledge your study’s limitations than to have them
There is little room for speculation in the discussion.
pointed out by a peer-reviewer or a reader (in a letter to the
The discussion should remain focused on the your data
editor after publication). Fluck et al2 acknowledged a limita-
and the patients and/or devices in your study. If the sub-
tion of their study and used it to make a suggestion for further
jects in your study had asthma, it is usually not appropriate
research: “We used only healthy white subjects, to minimize
to speculate about how your findings might apply to other
confounding variables. Future research should include testing
patient populations. If your study used volume-controlled
subjects with darker skin and subjects whose oxygen satura-
ventilation, it may not be appropriate to speculate about
tion is below normal (⬍ 95%).”
how the findings might apply to pressure-controlled ven-
Make Suggestions for Further Research tilation. If you feel compelled to speculate, be certain that
you clearly identify your comments as speculation: “We
Although a study may answer important questions, other speculate that. . . . ”
questions related to the subject may remain unanswered.
Moreover, some unanswered questions may become more Inflating the Importance of the Findings
focused because of your study. You should make sugges-
tions for further study in the discussion section. Labora- After all of the hard work that goes into a study, it is
tory experimental studies typically lead to suggestions for easy to attribute unwarranted importance to study findings.
follow-up clinical studies with human subjects. An exam- We all want our study to make an important contribution
ple comes from a laboratory study of oscillating positive that will be cited for generations to come. However, un-
expiratory pressure (OPEP) devices by Volsko et al,5 who warranted inflation of the importance of the study results
wrote, “One subject that remains to be explored is how to will disgust reviewers and readers. A measure of humility
determine at the bedside whether a patient can perform goes a long way.
OPEP and, if so, which device to select.”

Give the “Take-Home Message” in the Form of a Tangential Issues

Conclusion
It is important to remain focused on the hypothesis and
What is the “take-home message”? What do you want study results. Injecting tangential issues into the discussion
the reader to remember from your study? The take-home section distracts and confuses the reader. Tangential issues

1240 RESPIRATORY CARE • OCTOBER 2004 VOL 49 NO 10

 HOW TO WRITE AN EFFECTIVE DISCUSSION

run the risk of diluting and confounding the real message reader understand the study and that the highlight should
of the study. be on the study data.

The “Bully Pulpit” REFERENCES

1. Anton A, Tarrega J, Giner J, Guell R, Sanchis J. Acute physio-
Do not use the discussion section to criticize other stud- logic effects of nasal and full-face masks during noninvasive
positive-pressure ventilation in patients with acute exacerbations
ies. Although you should contrast your findings to other
of chronic obstructive pulmonary disease. Respir Care 2003;
published studies, this should be done professionally. Do 48(10):922–925.
not use the discussion to attack other investigators. More- 2. Fluck RR Jr, Schroeder C, Frani G, Kropf B, Engbretson B. Does
over, never preach to the reader. ambient light affect the accuracy of pulse oximetry? Respir Care
2003;48(7):677–680.
3. Stoller JK, Orens DK, Fatica C, Elliott M, Kester L, Woods J, et al.
Conclusions That Are Not Supported by the Data Weekly versus daily changes of in-line suction catheters: impact on
rates of ventilator-associated pneumonia and associated costs. Respir
The hypothesis 3 study 3 data 3 conclusions should Care 2003;48(5):494–499.
be a tight package. Avoid the temptation to allow your 4. Swart F, Schuurmans MM, Heydenreich JC, Pieper CH, Bolliger CT. Com-
parison of a new desktop spirometer (Spirospec) with a laboratory spirom-
biases to enter into the conclusions. eter in a respiratory out-patient clinic. Respir Care 2003;48(6):591–595.
5. Volsko TA, DiFiore JM, Chatburn RL. Performance comparison of
Summary two oscillating positive expiratory pressure devices: Acapella versus
Flutter. Respir Care 2003;48(2):124–130.
6. Apostolopoulou E, Bakakos P, Katostaras T, Gregorakos L. Inci-
The discussion section gives you an opportunity to ex- dence and risk factors for ventilator-associated pneumonia in 4 mul-
plain the meaning of your results. When writing the dis- tidisciplinary intensive care units in Athens, Greece. Respir Care
cussion, remember that the focus should be to help the 2003;48(7):681–688.

RESPIRATORY CARE • OCTOBER 2004 VOL 49 NO 10 1241

View publication stats