Writing The Biomedical Manuscript:
A Systematic Approach
Christopher Dant
Stanford Medical School
Manuscript Writing


Part I

About writing and what makes a good paper

Parts of a manuscript

Figures and tables

Writing strategies
Part II

Essentials of good writing

Sentences-Paragraphs

Common writing mistakes
Medical Communications Today

Medical science becoming increasing specialized

1976: ~ 5,000 biomedical journals, only in libraries

2006: >17,000 biomedical journals, electronic on internet

Subspecialties and new vocabularies has increased dramatically over
past 20 years

Biomedical research moving to interdisciplinary initiatives—
The NIH Roadmap

Thus, readership increasingly specialized yet interdisciplinary,
requiring an approach that is common, clear, simplified.
“We must strive to make our articles increasingly ‘reader friendly’ and cross-discipline in
language…”
Dr. Jerome Kassirer, Former Editor, NEJM
A Recognized Problem
“There is no form of prose
more difficult to
understand and more
tedious to read than the
average scientific paper!”
-Dr. Francis Crick, 1994
The Astonishing Hypothesis
The Avoidable Downfall
Your research

Carefully planned

Novel

Flawlessly designed and executed
Your paper

Poorly designed and written—rejected or delayed

The loss or delay of disseminating important
critical information to the science community
Journal Editors Agree

Good writing signals clear thinking and an
organized approach

Clear direct English and logical, cohesive,
organized writing are key to acceptance

Even the most novel and well-constructed
study will be rejected if the writing is
flawed
Writing is an Essential Skill

The ability to communicate clearly and precisely
through the written word is an essential skill for
medical researchers

Delayed publications and denial of funding
because of poorly written manuscripts and grants
continues to plague researchers

The career of a researcher can depend heavily on
this skill
Key Difficulties


Many papers are poorly constructed and written

Some scientists have not learned good manuscript
writing techniques

Others do not enjoy writing, and do not take the time
or effort to ensure that the prose is clear and logical.
Authors are typically so familiar with the material
that it is difficult to step back and view it from the
point of view of a reader not familiar with the
science

Peer review is therefore critical
Manuscript Deficiencies
57 articles evaluated to Emerg Med—28 accepted, 29 rejected/pending
Of these 29:
Ambiguous methods
77%
Ambiguous results
68%
Conclusions not warranted by data
72%
Poor referencing
56%
Inadequate study design description
51%
Unclear tables
49%
Overly long discussion
49%
Inadequate definition of terms
49%
“Deficiencies in manuscript preparation are more frequent than mistakes in
study design and execution. Specific training…in manuscript preparation is
indicated.”
Taylor and Brown, Emerg Med 13(4):444-50, 2001
Top 10
Reasons Manuscripts Rejected
1.
Wrong journal, format, preparation
2.
Disorganized study design
3.
Defective tables, figures
4.
Poor organization throughout, writing, spelling
5.
No hypothesis or problem statement
6.
No or insufficient conclusion
7.
Overinterpretation of results
8.
Article unfocused, too verbose and long
9.
Inappropriate statistical methods; methods not sufficient to repeat
study
10. Poorly written abstract/title
Pierson DJ, Respiratory Care 49(10), 2004
Byrne DW, Publishing Medical Research Papers, Williams and Wilkins, 1998
The Paper

Writing and editing the paper is the last step in
the research process

The paper tells the story from study inception,
through data collection, statistical analysis,
findings and and discussion

The process of writing the paper should be
analogous to the research process—it requires
attention to detail, time, and revision
Manuscript Reviews
Peer Reviewers
No - DOA
Manuscript
Yes
Masked review
Journal
Decision Editor
Rejection - Outright
Revise-Acceptance ?
Receipt of
manuscript by
editorial asst
Editor
Title & Abstract
Headings
References
Tables/Figures
Read Through
Appropriate to
journal?
Conform to
guidelines?
Editor Reports
Summary of peer reviews
Summary of editor’s review
Revise-Accepted
Acceptance - Outright
Start with Outline

Outline each segment of the paper using
traditional outline: I, II, III, A, B, 1, 2, a

Forces logical thought and order

Eliminates unorganized thinking and writing

Uncovers flaws in arguments

Reduces wordiness

Makes writing easier

Include your draft figures, tables
Outline
I.
Introduction
A.
Zinc plays a critical role in biochemical functions in cells
1.
Mitochondrial function (Billings)
2.
Cell motility (Jones, Smith)
B.
Zn concentrations affected by physiological changes in pregnancy
(Billings)
C.
Zn deficiency increases spontaneous abortions and pregnancy
complications
1.
Rhesus monkeys (Putter)
2.
White rats (Michaels, Reiss)
D.
In humans, the role of Zn deficiency in pregnancy outcome is unclear
(Brown)
E.
Objective: we evaluated whether Zn supplementation during
pregnancy is associated with changes in birth outcomes.
Journal Editor:
What’s A Good Manuscript?
Title descriptive and specific
Abstract descriptive, specific, and correct length
Introduction and background short and strong
Research question clearly stated
Literature cited is comprehensive and relevant
Methods descriptive enough to be replicated; appropriate statistical analyses
Figures and Tables stand on their own, support conclusions, well constructed
Citations relevant to topic
Discussion within boundaries of findings; demonstrate how findings have helped
resolve stated problem; implications and future work addressed
Writing clear, terse, logical
Manuscript follows journal guidelines
The Title
The Title

First reviewed by Journal Editors before
abstract

Short

Specific, Relevant, Descriptive

Write last—your findings and conclusions may
alter your title
Title: Ask Yourself

What is the single most important point of this
study?

How would I tell my colleague, in one short
descriptive sentence:
what’s this study about?

A descriptive, specific title perfectly framing
your study will be apparent only after you’ve
written the paper and abstract.

Start with a short descriptive working title
Unnecessary Title Phrases

A Study of… A Study to Determine Results of…

An Innovative Method…

Contributions to (of)…

Investigations on (concerning, about)…

Observations on…

A Trial Comparing…
Title—Specific & Descriptive

A Study Involving Medical Imaging with Genetic Patients
and Turner’s Syndrome
MRI Brain Imaging in Children With Turner’s Syndrome and
Other X Chromosome Abnormalities

Nerve Growth Factors and Sodium Channels in Pancreatic
Cells
Nerve Growth Factor Increases Sodium Channel Expression
in Pancreatic (Beta) Cells: Implications for Insulin
Secretion
Title—Specific & Descriptive
Down Syndrome—Where we are today: A Review
Down Syndrome: Genetic, Behavior, and Functional Neuroimaging Research
2000-2006

Aldosterone and Heart Failure
Aldosterone Plasma Concentrations Increase with Severity of Congestive
Heart Failure

A study of MI in older Americans 1994-1999
Epidemiological survey of MI in Community-Dwelling American Males Over
65 years

Lazarus arise! Life and Death Issues in Intensive Care
End-of-Life Care Issues for Critically Ill Patients in Intensive Care Hospitals

Title—Specific & Descriptive
Hepatitis C virus associated membranoproliferative
glomerulonephritis: A tale of Mice and Men
Membranoproliferative Glomerulonephritis Associated with Hepatitis C
Virus in F39(b) Nude Mice: Applicability to Humans

Isolated unilateral tubular sclerosis-associated severe late-stage renal
cystic disease in neonates
Severe, Late-Stage Renal Cystic Disease in Neonates Associated with
Isolated Unilateral Tubular Sclerosis

Drug trial comparing systemic beta blocker with calcium-channel
blocker in CHF
Open-label Comparison of Altenolol and Propranolol versus Nifedipine in
Patients with CHF: Beta Blocker and Calcium-Channel Blocker
Mechanisms

Don’t Stack Adjectives

Female but not male serotonin reuptake
transporter (5-HTT) model knockout mice exhibit
bladder instability: Implications
5-HTT female (not male) knockout mice have
unstable bladders: Implications for Stress Urinary
Incontinence Treatment
Good Titles—Sentences


Intellectual impairments occur in children with blood lead
concentrations above 10 mg per deciliter
Increased 17b-estradiol suppresses PTHrP gene expression
in breast cancer cell lines

Spinal cord stimulation attenuates visceromotor reflexes in
a rat model of post-inflammatory colonic hypersensitivity

Rhinovirus challenge decreases antioxidant enzymes in
respiratory epithelial cells
Not Sentences But Good Titles

Comparison of MRI and CT for Detection of Acute
Intracerebral Hemorrhage

Extracranial Thrombotically Active Carotid Plaque as a Risk
for Ischemic Stroke

Annual Revaccination Against Influenza and Mortality Risk in
Community-Dwelling Elderly Persons

Effect of Antihypertensive Agents on Cardiovascular Events
in Patients With Coronary Disease and Normal Blood
Pressure. The CAMELOT Study: A Randomized Controlled
Trial
The Abstract
The Abstract

1st Impression to journal editor and the reader!

Follow the Journal’s Guidelines

Most abstracts are often too long: ≤250 words: Cannot
upload your paper!

Structure it (outline it)
“The abstract is the single most important part of a manuscript, yet the
most often poorly written”
-JAMA Editor
The Abstract

First looked at by editors/sometimes only thing read by
readers

Sometimes only part available electronically—KEY words!

Summarizes the main points succinctly:






Background/Significance
Objective
Study design, method
Primary germane results
Principal conclusions, implications
Do NOT be vague—be substantive and brief


NOT “The implications are summarized”
INSTEAD Summarize the implications!
Abstract

Emphasize methods, main results, and conclusion

Introduction/purpose: 1 short sentence

Put objective as imperative style:

Objective: To evaluate whether zinc supplementation
during pregnancy affects infant birth measures.

Methods, Results: 2-4 sentences

Conclusion: 1-2 sentences
Structured Abstract
Context—Summarize the study rationale and provide clinical (or
other) reason for the study question.
Objective—State the purpose or question asked. If more than one
objective, state primary objective and key secondary
objectives.
Design—Describe basic design, including relevant details.
Setting—General community, primary care, hospital, etc.
Patient or other population—describe demographics, disorders,
inclusion/exclusion criteria, etc.
Interventions—name, dose, dosage
Main outcome measure(s)
Results
Conclusions
The Effect of Zinc Supplementation on
Pregnancy Outcome
Objective—To evaluate whether zinc supplementation during pregnancy affects infant
birth measures.
Design—Randomized, double-blind, placebo-controlled trial.
Setting—Outpatient clinic at University of Alabama at Birmingham.
Patients—580 healthy African-American pregnant women with plasma zinc levels below
normal levels, randomized at 19 weeks’ gestational age and divided by median body
mass of 26 kg/m2 into placebo and zinc supplement groups.
Intervention—Women receiving a non-zinc-containing prenatal vitamin tablet were
randomized to 25 mg/day zinc or placebo.
Outcome Measures—Birth weight, gestational age at birth, head circumference at birth.
Results—Infants from zinc supplement group had greater birth weight (p<0.01) and head
circumference (p=0.02) than those in placebo group. Women with body mass ≤ 26
kg/m2 had infants with significantly higher birth weights (median 245 g, p<0.001) and
larger head circumference (median 0.7 cm, p=0.003).
Conclusions—Daily zinc supplementation in women with low plasma zinc concentrations in
early pregnancy is associated with greater birth weights and head circumferences,
with the effect occurring in women with body mass index ≤26 kg/m2. The specific
effects of zinc on the fetus are unknown, and future work is focusing on zinc effects
on embryonic cells in vitro.
192 words
Abstract: Be Specific

The principles of reconstruction of the traumatic losses of the external
ear are presented, with emphasis on effective treatment of the acutely
injured ear. The steps necessary for obtaining satisfactory reconstruction
are discussed, including two new techniques.

In the past decade, advances in soft tissue surgical techniques have
allowed surgeons to successfully reconstruct detached ears. We present
two new surgical reconstruction techniques of the acutely injured
detached ear. These include the use of local and distant tissues to obtain
soft-tissue coverage, and the recent use of Silastic cartilage for structural
support…
The Introduction
Introduction

Why did you carry out this research? State the specific purpose or
rationale for the study.

What is the existing state of knowledge of this topic? Synthesize
information tracing the development of the problem and summarize
its current state…ie, the background. You ask (with citations):



What’s known?
What’s unknown?
What are the gaps in knowledge this study will fill?

What are you going to do and what do you expect to find?
State your hypothesis or question clearly (Objectives, Aims)

Give only strictly pertinent references.
Introduction

This is a vital part of your paper—it convinces (or not) the reader
whether your study:
 Has merit and asks important research questions
 Is focused and supported by relevant recent citations
 Is ultimately important to human health and human disease

Reviewers and editors will judge the paper’s importance in the
introduction.

You will better focus your introduction AFTER you construct your
findings (results) and consider them (discussion).

Your research question is the most important part—in your discussion,
you will address whether the question or hypothesis was answered
based on your data.
Introduction Structure
1.
What is the general problem or current situation?
Zinc plays a critical role in many biochemical functions, including nucleic acid
metabolism and is critical in early development.
2.
What is the specific problem or controversy? Its significance?
Zinc deficiency is associated with increase metabolic problems in fetuses.
Studies evaluating relationship between zinc intake and pregnancy
outcomes have produced conflicting results for many reasons…
3.
What are our hypotheses/questions, and how will we answer them?
To clarify the relationship between zinc levels in the mother’s diet and
pregnancy outcomes, we undertook a randomized placebo-controlled
trial of zinc supplementation.
Our objective was to determine if zinc supplementation was associated with
higher birth weight.
Our findings will help to provide continuing nutritional guidelines in
pregnancy.
The Methods
Methods are Critical: Editors’ Responses
What section contains the most flaws?
0
What section responsible for outright rejection?
Discussion
Discussion
Results
Methods
Introduction
Results
Methods
Introduction
5
10
15
20
% Responses
0
5
10
15
20
% Responses
How frequently do Editors encounter manuscript problems?
Poorly written, excessive jargon
Inadequate/inappropriate presentation
Poor description of design
Excessive zeal and self promotion
Rationale confused, contradictory
Essential data omitted, ignored
Boring
Important work of others ignored
Seldom
Byrne DW, Publishing Medical Research Papers, Williams and Wilkins, 1998
Occasionally
Frequently
Methods

Editors judge the study on whether your methods are
adequate to answer your specific aim or hypothesis
Rationale for choosing procedures/tests
 The pivotal point to judge whether the results are valid


Don’t suggest a method you have no expertise with
Your peer reviewer may uncover this
Use consultants for methods you have no experience with, stating this in
paper


Methods usually the weakest section
Often deficient in detail, not providing enough information to replicate
the study
 Statistical shortcomings

Methods

Study design or analysis type and period of study

Condition or disease studied

Human subjects approval

Details of sample (number, recruiting methods of study subjects,
patients, how organized)

Interventions, outcome measures, statistical analyses

Include the locations and times that data were collected

Give enough information to replicate the study; don’t assume only the
specialist in your field will read it
Methods

Balance between brevity and completeness

Sometimes reference an often-used method

Use figures and tables (eg, flow diagram)

Naming things—be consistent


Acronyms—spell out first time, use consistently throughout
Specialized tests, terms—use identical name in text, figs, tables

Develop list of frequently used terms

Present in logical order and your subsequent results
should follow that same order
Method—Procedures
Method diagrams communicate schedule of procedures, enrollment, study design,
mechanisms of action, guidelines, algorithms to reduce text and increase
comprehension.
The Results
Results—The Beginning


The heart of your paper
Write after figures and tables are constructed





Consider your data critically
Construct tables, figures and include them in outline
Write the results
Use subheadings
Results determine



Whether you’ve answered your original question(s)
Your direction for future studies
Both of which belong in the discussion
Results—The Beginning

State ALL the findings



Whether significant or not
Without bias or interpretation
Do not include weaknesses, strengths of study, ie don’t discuss results

List experiments in order listed in methods

Use logical headers and group your findings



Characteristics of study subjects
Findings in order listed in methods
General to specific

Use past tense

Results confirm or reject your hypothesis: they do not prove
anything.
Results





Short and to the point—Main or most important findings
first
Present only data directly relevant to the study—focus
Don’t repeat methods but you may remind the reader
briefly how you measured something.
Allow the data to speak for itself—use tables/figures —
construct them first and use as a basis for writing
In Tables and Figures, be descriptive, specific. Do not
repeat the obvious:




NO: Results of the kidney lead analysis are shown in Table 1.
YES: Kidney lead concentrations increased in group 1 over the first 10
study weeks (Table 1).
Present absolute numbers and percentages so reviewers
can judge the significance of the findings.
Statistical significance ≠ clinical significance
Results or Data?
Results
Mean translational movements in the X (left to right), Y
(back to front) and Z (bottom to top) head directions
were 0.10 ± 0.11 mm, 0.16 ± 0.03 mm, and 0.65 ± 0.58
mm, respectively. Mean rotational movements about
the three axes were 0.44 ± 0.42 degrees, 0.24 ± 0.26
degrees, and 0.18 ± 0.17 degrees, respectively.
Movement was not significantly correlated with age for
translation in the X (r = -0.09; p = 0.69), Y (r = 0.21; p =
0.35) or Z (r = -1.02; p = 0.64) directions. Movement
was not significantly correlated with age for rotation in
the X (r = 0.15; p = 0.51), Y (r = -0.20; p = 0.35) or Z (r
= 0.02; p = 0.94) directions.
Results!
Results—Don’t Regurgitate Data

As shown in Table 1, the mean age of participants was 20.4 ± 2 years, and
80% of patients were Caucasian. Treatment group contained 40 patients,
whereas control group contained 45 patients. Table 2 shows the
demographics of women in these groups. There were 24 women in the
control group, and 33 women in the treatment group...

There were no significant differences in treatment and control patient
intake demographics (Table 1), although a significantly greater number of
patients in the treatment group dropped from the study for a variety of
reasons, mostly relating to adverse reactions. However, analysis of
patients in this group later revealed that those dropped patients had
significant disease at intake (Table 2). In comparing the two treatment
groups (Figure 1), we found that...
Don’t State the Obvious
Figure 1 is a graph illustrating
the plasma zinc levels
(µmoL/L) over the 37 weeks
versus gestational age in both
the zinc supplement group
and placebo group. The
placebo and the zinc group
both decreased over the 37
weeks of the study, but the
differences were significant for
the zinc group.
State What’s Important
We measured mothers’ plasma
zinc levels before randomization
(week 19) and at 26, 32, and 37
weeks’ gestational age (Fig 1).
Beginning as early as 26
weeks and at each timepoint,
differences in plasma zinc levels
between placebo and zinc
supplement groups were
statistically significant (P≤0.05)
after randomization.
Results—Major Mistakes

Failure to provide all the data critical to answering
the research question

Interpreting or commenting on results


“Six of the 20 patients required intubation, illustrating the seriousness
of this problem” (belongs in Discussion)
“Over 40% of treated rats exhibited a decreased inflammatory
response, an unexpected finding” (belongs in Discussion)

Failure to adequately address statistical methods

Tables and figures inappropriate, unbalanced

Tables and figures poorly constructed
The Discussion
Discussion Construction

Summarize major findings—1st paragraph

Explain how your findings relate to those of
others—what do they mean?

Clinical relevance of the findings?

Limitations and how this influenced your study?

How will you overcome these in the next studies?

Explain the implications of findings

What future direction(s) will you take?
Discussion: Getting Carried Away

Few studies make discoveries changing the course of
scientific direction, and so authors:

Attempt to overly state or the importance of their findings

Come to erroneous or unsupported conclusions

Uncritically accept statistical results

This all distracts from work’s importance and signals to the
reviewer problems with the research

Also results in excessive length, a common problem

Authors should let the data speak for themselves
Discussion—Common Mistakes
1.
Unwarranted speculations
2.
Injecting tangential issues
3.
Conclusions not supported by the data
4.
Not suggesting future directions for research
hypothesis  study  data/results  conclusions
TIGHT PACKAGE
Sections Unbalanced
Article 3650 words
Abstract
Introduction
Original
Methods
Final
Results
Discussion
0
500
1000
1500
Tables & Figures
Tables and Figures

Critical to a paper—Editors and readers look at these before
reading the paper!

Editors judge your paper on how well these are constructed

Stand alone and tell a complete story

Unambiguous—immediately clear

Eliminate numerical data and long explanations in text

Figures display important trends, procedures, simplify
detailed data, and show basic methodologies.
Tables
This requires a table!
Tables
This result does NOT require a table!
Growth medium aeration was essential for the growth
of S. coelicolor. At room temperature (24°C), no growth
was measurable in stationary cultures, whereas in
aerated cultures, we measured substantial growth (78
Klett units).
Tables & Result
In women with BMI <26
kg/m2, zinc
supplementation was
associated with a
significant increase in
birth weight of 248 g
(P=0.005), an increase in
head circumference of
0.7 cm (P=0.005), and
increase in arm length of
0.3 cm (P=0.03). The
other outcome measures
all favored the zinc
supplement group but
the differences were not
statistically significant
(Table 2).
Table & Result
Table 3 shows the mean birth weight by the BMI categories recommended
by the NIH Institute of Medicine. The lower the BMI, the greater the
effect of zinc supplementation on birth weight.
Patient Disposition (Results)
Complex
Study Design
Simplified
Bar or Line Graphs-Colors?
This graph

will appear in the journal like this:
Journals DO NOT allow color graphs unless they are necessary for
understanding the graph
Simple Graph




Use graphing software in word/powerpoint to create: KISS
No more than 3-4 groups
Keep all lines solid, few symbols
Put in SD and P values if relevant
From The Journal
Editor’s Perspective
Prepare Your Manuscript Carefully


Incorrect style irritates reviewers and editors, and the
wrong style suggests that another journal previously
rejected the paper.
Edit carefully



Check accuracy of references with original sources


Eliminate spelling, punctuation, and grammar errors
Good writing requires rewriting
Incorrect citations inconvenience the publisher and are a
disservice to the reader
Double-check numerical data!

Numbers in abstract, text, tables, figures, ledends, and text
must be consistent and correct
Avoid Repetition

Do not disclose results in introduction

Do not repeat the Introduction in Discussion

In text



Do not repeat figure legends, table titles, or contents of the
tables themselves
Use tables sparingly

Presenting a few facts in text takes less space than a table

Do not use tables for presenting simple lists
Abbreviations, definitions, symbols in figures and tables
must be explained in legends and footnotes

Never refer a reader back to text for such information
Journal Review
 Full review and decision takes ~1 month
 Editors make decision based on arguments; they don’t
count votes from Peer Reviewers
 Most papers undergo 2 rounds before publication
 For borderline decisions, a goal is to avoid multiple
rounds of review
 Pressure to publish quickly may lead to rejection if
further experiments are needed
What Helps or Hinders the Paper?
What Helps?





New data — to a point
Referee or Editor made factual errors (easy to prove)
Careful and accurate response to criticisms (table)
Telling the editor that reviews were helpful in improving the paper
Knowing how to submit to the journal electronically — Practice!
What Doesn’t?




Referees were “unfair” and the criticisms were largely “not valid”
Guesses at referee identity followed by personal attacks
Specific evidence of bias by referee (difficult to prove)
Endorsements or (positive) statements about your standing and
reputation
The Paper Is Cohesive

Question (objective, specific aim) is posed in Introduction

Methods tell how you propose to answer these aims

Results presented answer (or not) the question

Discussion should be within the bounds of the results

Conclusions directly answer the original questions in the
Introduction

Each section should refer back to one another
Evaluate Your Paper

To understand and evaluate your paper, the
editor will ask (and so should you):
What specific questions/aims does the paper address?
 Are the methods/design adequate to answer your
questions?
 What are the main conclusions?
 What specific evidence (data) supports those
conclusions?
 What is the quality of that evidence?
 Conclusions: what is the study’s significance…what
insights or new directions are evident?

My Suggestions

Put the manuscript away for a couple of days

Read troublesome areas aloud

Don’t try to edit a mangled paragraph—delete and
rewrite it

Your colleagues reviews of writing and table/figures are
valuable—don’t be defensive about edits

Let go of “academic” writing habits and don’t imitate
others’ writing. Develop your own clear, direct style
Writing Deficiencies
Most commonly cited by journal editors

Wordiness and redundancies

Cut, condense, combine

Poor flow of ideas

Outline to catch logic problems

Poor syntax and grammar

Consult an editor

Excessive abstraction

Be specific and descriptive

Unnecessary complexity

Keep it simple and direct

Excessive compression

Do not overly compress writing

Unnecessary qualification

Qualify statements as necessary
Byrne, D. Science Editor 23:2, 2000
Summary
 Outline your paper
 Start early as your data is being analyzed
 Look at your data and decide how to organize and present
your results: tables, figures, text
 Patterns and clues will emerge to guide your argument
 Start with results then introduction and
discussion/conclusions
 Write title and abstract last
 Put it away, re-read, give to your colleagues to read
 Revise, revise, and re-revise
 Adhere to journal guidelines!
 Critically evaluate your paper with an editor’s eye
 Write clearly, logically, and simply!