NHS Fife Research Study Guide: [17] How to Write an Abstract.

Study Guide 17: How to Write an Abstract

Dr David Chinn,
Research & Development Office,
Queen Margaret Hospital, Dunfermline, Fife.
[email protected] 01383 623623 (ext 20943)
Alternative contact: Dr Amanda Wood [email protected] 01383 623623 (ext 20941)

Contents Page
1 Overview and learning outcomes 1
2 Types of reports that need an abstract 1
3 First thoughts 2
4 Example of a poor abstract 3
5 Content and structure of an abstract 3
6 An exercise in writing an abstract 6

(1) Overview and learning outcomes

Whether you are preparing a poster or oral presentation for a conference or writing a
paper for submission to a journal it is very important to prepare a good abstract. The
abstract is there to attract the reader, is often the only part that is read, or published,
and is critical in encouraging a conference organiser or editor to accept your poster,
presentation or paper in the first place. Preparing the abstract requires skill and
attention to detail. After reading this guide you should be able to:

Identify which reports need an abstract

State the components of a good abstract
Recognise the difference between a good and poor abstract
Describe the different styles of abstracts (structured versus unstructured)
Have gained practical experience of preparing a good abstract

Associated NHS Fife study guides:

18 How to write up and get your work published

(2) Types of reports that need an abstract

Abstracts are required for the following reports:

1. Conference presentations (whether oral or poster)

2. Original scientific studies, whether for a full paper (2000-4000 words) or
for a short report (800 – 1000 words)
3. Internal reports
4. Case reports or case series
5. Student texts / assignments

Abstracts are seldom required, if ever, when writing a review article (book or subject),
an invited editorial or a letter to a journal editor.

An abstract submitted for presentation at a conference will be scored by independent

reviewers for its content, clarity, originality, scientific merit, and clinical relevance.
Competition may be high so it is important that the abstract attains a high score.
Poorly written abstracts will not achieve this and the work will be declined by the
conference organisers. Furthermore, it is important that the work has been
Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 1
NHS Fife Research Study Guide: [17] How to Write an Abstract.

completed. Inclusion of a statement such as ‘collection of further data is ongoing and

full results will be presented’ is an invitation to dismiss the abstract out of hand.

A paper submitted for publication in a journal will first have its abstract reviewed by
the editor who will decide whether it should be sent for full peer review. If acceptable,
the editor will then send the abstract to one or more reviewers (referees) with an
invitation to review the full paper. Referees are not usually paid and editors will wish
to protect their cohort of willing referees by not asking them to waste their time
reviewing poorly reported studies. Hence, again, the abstract must be informative
and well written to overcome these initial hurdles.

Internal reports should have an abstract, sometimes referred to as an ‘Executive

Summary’. The usual limitations on word count will not apply to such statements but
the abstract or executive summary should not be overly long. An internal report may
also require a lay summary for the general readership.

(3) First thoughts

First read the conference guidance or journal instructions to authors regarding the
style (structured or unstructured) and word count for an abstract.

Second, decide what the key messages are. Try writing two or three sentences on:
‘What do we know already about this subject?’ and ‘What does this study add?’ This
exercise will help focus the content.

Third, devise an informative but attractive title. This is critical in attracting the casual
reader. The title may be a description of what was studied (Maternal obesity and the
risk of stillbirth: a population based case control study) or a statement of the findings
(Maternal obesity is an independent risk factor for stillbirth in nulliparous, Caucasian
women).

Write the first draft in good English with a style that makes it readable but, at this
stage do not limit the word count as the text can be précised later to fit with the
requirements. Write with an international audience in mind as your work may be read
by individuals who are not native English speakers. If space allows write as if your
audience is ignorant of the subject to improve the clarity of the content. Tips for
writing an abstract include:
1. Use short, concise sentences where possible.
2. Use active not passive tense. For example, ‘cats eat fish’ (active tense)
rather than ‘fish are eaten by cats’ (passive tense).
3. Use positive rather than negative statements. For example, ‘90% of
students passed’ rather than 10% of students failed’.
4. Use simple words. For example, ‘the study shows...’ rather than ‘the study
demonstrates...’
5. Use abbreviations and acronyms where possible (to reduce word count)
but spell them in full on their first appearance.
6. Cite numbers in numerical form rather than as words (to reduce word
count).
7. Avoid needless words.
8. Avoid imprecision and irrelevance.
9. Avoid double negatives. For example, ‘diabetes is common’ rather than
‘diabetes is not uncommon’.
Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 2
NHS Fife Research Study Guide: [17] How to Write an Abstract.

(4) Example of a poor abstract

Here is an example by Sir James Howie in ‘How to do it’, a collection of articles

published in the British Medical Journal, BMJ Publications, 1979 (see the chapter
‘How to attract the reader’, page 140).

‘An extensive survey of chickens in various situations has been made to ascertain
the incidence and points of origin of salmonellas. The study identified where infection
had been acquired. The implications of our findings are discussed and point to the
need for further research.’ [44 words]

The content of this abstract does not encourage the reader to investigate further; it is
imprecise on settings and methods and leaves the reader uninformed regarding the
‘implications’ of the study. A better effort would be:

Five thousand chickens were examined for salmonellas in 20 farms, three processing
plants and 100 shops. Infected feed containing fish meal on one farm was found to
result in widespread contamination of birds from that farm and, through them of birds
as they passed through one processing plant. We conclude that efforts to produce
clean poultry feed and to improve the hygiene of farms and processing plants will do
more to control food poisoning than the numerous, currently popular but futile
searches for human so-called carriers in shops, restaurants and homes. [91 words]

(5) The content and structure of an abstract

The abstract should ‘stand alone’ as a comprehensive description of the study, its
findings and conclusions. Abstracts may be published in conference proceedings and
those from papers will be published in databases such as MEDLINE and PubMed.
The abstract should answer five key questions:

1 Why did you start? Introduction / Background

2 What did you try to do? Aims / Objectives
3 What did you do? Methods / study design / setting / participants / main
outcome measures
4 What did you find? Results
5 What does it mean? Conclusions / Recommendations / the ‘so what’

The introduction, background, conclusions and any recommendations should consist

of one or two sentences only. The results section should comprise about half the
word count and should not include any tables or figures. The first draft need not be
constrained by the word count. Later the text will need to be trimmed but without
sacrificing the important content. Many drafts may be necessary but do not discard
earlier ones. If your paper is rejected by one journal a resubmission to an alternative
journal may require a different structure, or word count.

Proof reading is important. Ask a colleague to check it for content, accuracy and
comprehension. Spelling mistakes are easily missed. Try reading this text:

Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 3
NHS Fife Research Study Guide: [17] How to Write an Abstract.

I cdnuolt blveiee that I cluod aulaclty uesdnatnrd what I was rdanieg. The
phaonmneal pweor of the hmuan mnid, aoccdrnig to a rscheearch sutdy at
Cmabrigde Uinervtisy, it dseno't mtaetr in what oerdr the ltteres in a word are, the
olny iproamtnt tihng is that the frsit and last ltteer be in the rghit pclae. The rset can
be a taotl mses and you can still raed it whotuit a pboerlm. This is bcuseae the
huamn mnid deos not raed ervey lteter by istlef, but the word as a wlohe. Azanmig. I
awlyas tghuhot slpeling was ipmorantt!

Apparently, only 55% of us can read this easily. How did you get on?

Abstracts can be structured or unstructured. In a structured abstract the text is

broken up into sections each with a separate heading. The headings will be specified
in the instructions to authors. Here is an example from one journal:

1. Background
2. Objective(s)
3. Study design
4. Setting
5. Participants
6. Main outcome measures
7. Principal findings
8. Conclusions / Recommendations

An unstructured abstract may follow the same approach but not retain the headings.
In practice, it helps to adopt the structured approach even if the instructions do not
request it. To illustrate this see the examples below which give two versions of a
structured abstract and one version of an unstructured abstract from a study
investigating the reasons why patients chose not to take part in a physical activity
promotion trial being conducted in primary care.

TITLE: FACTORS ASSOCIATED WITH NON-PARTICIPATION IN A

PHYSICAL ACTIVITY PROMOTION TRIAL

Example 1 (structured abstract) (271 words)

Background. Non-participation can bias outcome in intervention studies of physical

activity.

Objectives. To compare characteristics, knowledge and attitudes to physical activity

in participants and non-participants of a physical activity intervention trial in primary
care.

Study design. Cross-sectional survey.

Setting. An inner city general practice, Newcastle-upon-Tyne.

Participants. Patients aged 40-64 years recruited opportunistically during surgery

visits.

Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 4
NHS Fife Research Study Guide: [17] How to Write an Abstract.

Methods. Attitudes to physical activity, views of its health benefits and barriers to
participation were elicited in interviews with participants, and by postal questionnaire
from non-participants. GP held data were used to compare anthropometry and
lifestyle between groups.

Results. Of 842 eligible patients, 276 (33%) refused outright (non-volunteers) and
566 volunteered for the intervention study, of which 353 (42%) attended a baseline
assessment and 213 (25%) subsequently defaulted. The initial refusal rate was
relatively higher amongst men, smokers and those with addresses in more deprived
areas. Response rate to the postal survey of non-volunteers was 45%. Compared
with participants the non-volunteers were more likely to be an adult carer and to
report poorer health, and less likely to have had higher education or have children
living at home. Far more non-volunteers considered they already did enough
exercise to maintain health. Non-volunteers had less knowledge of the benefits of
physical activity, attached far less importance to it in maintaining health and were
more likely to cite ‘internal’, non-modifiable barriers.

Conclusion. Recruitment of ‘hard to engage’ individuals requires careful phrasing of

the message to focus on their personal goals and to address gaps in their knowledge
about physical activity and the principal barriers they perceive. Differential uptake
across population subgroups could lead to a widening of health inequalities.

Example 2 (structured abstract, for a different journal) (337 words)

Background: Non-participation can bias outcome in primary-care based intervention

studies of physical activity (PA). Knowledge of the characteristics and attitudes to PA
of those who decline to enrol in a program are needed to inform recruitment
strategies to improve uptake.

Methods: Attitudes to PA, views of its health benefits and barriers to participation
were elicited in interviews with participants, and by postal questionnaire from non-
participants in the Newcastle Exercise Project (NEP). Patients aged 40-64 years
were recruited opportunistically during visits to an inner city general practice,
Newcastle-upon-Tyne, UK, 1993-95. GP held data were used to compare
anthropometry and lifestyle between participants and non-participants.

Results: Of 842 eligible patients, 276 (33%) refused outright (non-volunteers) and
566 volunteered for the intervention study, of which 353 (42%) attended a baseline
assessment and 213 (25%) subsequently defaulted. The initial refusal rate was
relatively higher amongst men, smokers and those with addresses in more deprived
areas. Response rate to the postal survey of non-volunteers was 45%; as a group,
those who replied under-represented smokers and those living in more deprived
areas. Compared with participants the non-volunteers were more likely to be an adult
carer and less likely to be a home owner, or to have had higher education. Non-
volunteers reported poorer health, had less knowledge of the benefits of PA and
attached less importance to it in maintaining health compared with participants. 62%
of non-volunteers considered they already did enough exercise to maintain health,
compared to 28% of participants. Non-volunteers were more likely to cite ‘internal’
barriers and cited dislike of exercise, poor health and fear of leaving their home
unattended more frequently compared with participants.

Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 5
NHS Fife Research Study Guide: [17] How to Write an Abstract.

Conclusions: Primary care patients most likely to benefit from a PA intervention are
least likely to join it. Recruitment of these ‘hard to engage’ individuals will require
careful phrasing of the message to focus on their personal goals and to address gaps
in their knowledge about PA and the principal barriers they perceive. Differential
uptake across population subgroups could lead to a widening of health inequalities.

Example 3 (unstructured abstract derived from example 2) (333 words)

Non-participation can bias outcome in primary-care based intervention studies of

physical activity (PA). Knowledge of the characteristics and attitudes to PA of those
who decline to enrol in a program are needed to inform recruitment strategies to
improve uptake. Attitudes to PA, views of its health benefits and barriers to
participation were elicited in interviews with participants, and by postal questionnaire
from non-participants in the Newcastle Exercise Project (NEP). Patients aged 40-64
years were recruited opportunistically during visits to an inner city general practice,
Newcastle-upon-Tyne, UK, 1993-95. GP held data were used to compare
anthropometry and lifestyle between participants and non-participants. Of 842 eligible
patients, 276 (33%) refused outright (non-volunteers) and 566 volunteered for the
intervention study, of which 353 (42%) attended a baseline assessment and 213
(25%) subsequently defaulted. The initial refusal rate was relatively higher amongst
men, smokers and those with addresses in more deprived areas. Response rate to
the postal survey of non-volunteers was 45%; as a group, those who replied under-
represented smokers and those living in more deprived areas. Compared with
participants the non-volunteers were more likely to be an adult carer and less likely to
be a home owner, or to have had higher education. Non-volunteers reported poorer
health, had less knowledge of the benefits of PA and attached less importance to it in
maintaining health compared with participants. 62% of non-volunteers considered
they already did enough exercise to maintain health, compared to 28% of
participants. Non-volunteers were more likely to cite ‘internal’ barriers and cited
dislike of exercise, poor health and fear of leaving their home unattended more
frequently compared with participants. Primary care patients most likely to benefit
from a PA intervention are least likely to join it. Recruitment of these ‘hard to engage’
individuals will require careful phrasing of the message to focus on their personal
goals and to address gaps in their knowledge about PA and the principal barriers
they perceive. Differential uptake across population subgroups could lead to a
widening of health inequalities.

(6) An exercise in writing an abstract

Read the paper below and compose a structured abstract of 250 words or less. Use
the following headings:
– Background
– Objective(s)
– Study design
– Setting
– Participants
– Main outcome measures
– Principal findings
– Conclusions

Compare your abstract with the published version in:

Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 6
NHS Fife Research Study Guide: [17] How to Write an Abstract.

Masson S, Chinn DJ, Tabaqchali MA, Waddup G, Dwarakanath AD. Is anaemia

relevant in the referral and diagnosis of colorectal cancer? Colorectal Disease
2007; 9: 736-739.

TITLE: IS ANAEMIA RELEVANT IN THE REFERRAL AND DIAGNOSIS OF

COLORECTAL CANCER?

Keywords: Colorectal cancer, anaemia, haemoglobin Word Count – 1327

Introduction

Colorectal cancer (CRC) is one of the commonest cancer diagnoses in the UK with
around 30 000 new cases per year, and remains one of the commonest causes of
cancer death in the UK [1]. In an effort to improve outcome and shorten the delay
between referral and diagnosis there has been a considerable increase in investment
in cancer diagnostic services in the UK National Health Service and the
establishment of national cancer waiting time targets that guarantee urgent
assessment [2]. To implement these standards, an appropriate threshold for referral
must be set with referrals being guided by the national guidelines for suspected
cancer [3,4]. With respect to CRC, the guideline’s criteria for referral include
symptoms (alteration in bowel habit and rectal bleeding), associated physical findings
(abdominal or rectal masses) and secondary effects (iron-deficiency anaemia [IDA]
and colonic obstruction). They are designed to assist identification of those at high
risk in whom urgent investigation is warranted. However, they are not evidence-
based and each referral criterion has its own level of risk, though the threshold for
each is not always explicit.

Investigation of IDA undoubtedly has a high yield for the diagnosis of gastrointestinal
malignancy [5]. In particular, it has long been held that right-sided CRC are most
closely correlated with the incidence of anaemia [6-9]. However, it remains unclear
as to how close this association is, and it has been shown that a diagnosis of
anaemia is insufficiently sensitive to aid decision making when investigating a right-
sided CRC [10]. Additionally, the diagnostic criteria for anaemia in IDA vary widely
between published studies, and the level of anaemia that requires investigation has
not been clarified.

We aimed to explore the relationship between anaemia and CRC. In particular, we

examined the occurrence of anaemia in patients with CRC to determine the
frequency of anaemia in relation to the diagnosis and site of tumour.

Patients and methods

We reviewed the hospital laboratory database and collected data prospectively from
diagnosis on patients diagnosed with CRC between January 2003 and June 2004.
The site of the cancer was noted together with the haemoglobin (Hb) estimation of
each patient at the time of referral. Anaemia was defined according to local practice
as Hb <12g/dl in females and Hb <13 g/dl in males, representing the lower limits of
our hospital reference laboratory. This differs from the threshold used in the current
national guidelines (Hb <10g/dl in females and Hb <11g/dl in males).

Data were analysed using SPSS [11]. Comparison of proportions was made using
the Chi-squared test and the 5% level indicated statistical significance.
Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 7
NHS Fife Research Study Guide: [17] How to Write an Abstract.

Results

Over 18 months, 143 patients were diagnosed with CRC; 85 (59%) were male and
58 (41%) were female. Ages ranged from 30 to 92 years in men (mean 67.4,
standard deviation 12.6) and 38-90 years in women (mean 71.1, standard deviation
11.9). The tumours were located on the right-side in 47 (33%) and on the left-side in
96 (67%), including 68 (48%) rectal tumours. Mean Hb was 12.0g/dl (standard
deviation 1.9, range 8.7-14.8) in female patients (lab reference 12.0-16.0g/dl) and
12.7g/dl (standard deviation 2.2, range 7.0-16.8) in male patients (lab reference 13.0-
18.0g/dl). Anaemia was present in 50% of female patients and 48% of male patients.
These proportions were 15.5% in females and 23.5% in males using the national
referral guidelines. Anaemia was present more frequently in those with right-sided
compared with left-sided tumours, and in those with non-rectal compared with rectal
tumours; the difference in proportions was statistically significant for females and,
with one exception, for males (Table 1).

Overall, using the laboratory reference values for haemoglobin, the difference in
proportion with anaemia for both sexes combined was, for right-sided versus left-
sided disease 32% (95% confidence interval 15 to 49%), and for non-rectal versus
rectal disease 32% (95% confidence interval 16 to 48%). Using the national
guideline for referral, the difference in proportion with reduced haemoglobin (both
sexes combined) was, for right-sided versus left-sided disease 30% (95% confidence
interval 17 to 43%), and for non-rectal versus rectal disease 22% (95% confidence
interval 9 to 35%). In multiple regression analysis, after first adjusting for sex, the
mean difference in Hb associated with (a) right-sided compared with left-sided
cancers was -1.7g/dl (standard error of the mean, SEM, 0.3g/dl) and (b) non-rectal
compared with rectal cancers was -1.6g/dl (SEM 0.3 g/dl).

Discussion

We have shown that the diagnosis of CRC is not associated with anaemia in at least
half of all cases using our local laboratory reference values for haemoglobin levels.
Furthermore, four out of five patients subsequently diagnosed with CRC had a
haemoglobin level at referral above the national, Department of Health guideline for
investigation for CRC. On further analysis, we demonstrated that patients with right-
sided and non-rectal cancers had a significantly lower haemoglobin level at
presentation than those with left-sided or rectal cancers, respectively. However,
even in these groups, a significant proportion was not judged anaemic, with around
one in three patients having a ‘normal’ Hb level. While a relationship exists between
proximal lesions and anaemia, our results confirm those of others [10] that the
presence of anaemia in CRC is too insensitive to be useful when making decisions
about colonic imaging. Furthermore, the threshold level of anaemia at which current
national guidelines recommend referral (Hb <11g/dl [male] or Hb <10g/dl [female])
[3,4] is lower than our own definition of anaemia and, consequently, the majority of
patients, including those with right-sided and non-rectal disease, do not have a Hb
below this threshold level. Clearly, anaemia is only one of several important referral
criteria, but given that many patients with distal lesions will be symptomatic [7,9], the
earlier detection of proximal lesions might be aided by a less stringent threshold.
Indeed, there is no reason to believe that mild anaemia is less indicative of serious
pathology than severe anaemia and there appears to be no correlation between the
severity of anaemia and the presence of malignancy in asymptomatic patients with
Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 8
NHS Fife Research Study Guide: [17] How to Write an Abstract.

IDA [12]. We consider the current threshold to be unjustified, being, as it is, not
evidence based.

Current efforts to improve outcome appear to be targeted at improving the referral

pathway, particularly the ‘two week standard’ [2]. It has been argued that the
introduction of this route only serves to highlight the current mismatch between
demand and service provision and occurs at the expense of ‘routine’ work [13]. In
addition, secondary-care audit suggests that, despite the introduction of this referral
pathway, although such clinics have higher diagnostic yields, they have yet to impact
on delay to treatment or disease stage and that the majority of cases still bypass this
route [14]. Notably, around 25-35% of all CRC patients currently present as surgical
emergencies with obstruction or perforation [14,15] where optimizing outcome
depends on access to appropriate specialist surgical expertise in the acute setting.
Additionally, in future, patients will be increasingly diagnosed through screening
based on recent pilot studies using faecal occult blood screens [16]. However, for
now, the most common pathway to diagnosis will continue to be presentation to
primary care with non-urgent symptoms. It is therefore important that referral
guidelines are appropriate, supported by evidence and, where possible, based on
predictive values of symptoms and other risks in primary care. In this regard, it is
incontrovertible that IDA is an indication for gastrointestinal investigation on its own
right [12,17,18]. However, our data highlight some important points about anaemia
that are pertinent to the referral of patients with suspected CRC. While our results
concur that right-sided and non-rectal cancer patients are more likely to be anaemic
at presentation, this observation must not detract from the fact that the majority of
patients with CRC are not anaemic. Additionally, we remain unconvinced in
investigating only those with a more severe anaemia (as determined by the current
national guidelines) and would advocate investigating all patients with IDA regardless
of the degree of anaemia. In conclusion, even in the absence of anaemia, all
patients in which CRC is suspected should be referred for urgent assessment via the
‘two week standard’.
References

[1] Office of National Statistics. Cancer registration statistic, 2003. London: HMSO 2005.
[2] Department of Health. The NHS Cancer Plan: A Plan for Investment. A Plan for Reform.
London: HMSO 2000.
[3] Department of Health. Referral Guidelines for Suspected Cancer. London: Department of Health
2000.
[4] Scottish Executive. Scottish Referral Guidelines for Suspected Cancer. Scottish Executive 2002.
[5] Goddard AF, McIntyre AS, Scott BB. Guidelines for the management of iron deficiency anaemia.
British Society of Gastroenterology. Gut 2000; 46 Suppl 3-4:IV1-IV5.
[6] Sadahiro S, Suzuki T, Tokunaga N et al. Anaemia in patients with colorectal cancer. J
Gastroenterol 1998; 33(4):488-94.
[7] Majumdar SR, Fletcher RH, Evans AT. How does colorectal cancer present? Symptoms,
duration, and clues to location. Am J Gastroenterol 1999; 94(10):3039-45.
[8] Goodman D, Irvin TT. Delay in the diagnosis and prognosis of carcinoma of the right colon. Br J
Surg 1993; 80(10):1327-9.
[9] Vanek VW, Whitt CL, Abdu RA, Kennedy WR. Comparison of right colon, left colon, and rectal
carcinoma. Am Surg 1986; 52(9):504-9.
[10] Rai S, Hemingway D. Iron deficiency anaemia--useful diagnostic tool for right sided colon
cancers? Colorectal Dis 2005; 7(6):588-90.
[11] Statistical package for the social sciences, release 10.0. Chicago:SPSS 2000.
[12] Niv E, Elis A, Zissin R et al. Iron deficiency anemia in patients without gastrointestinal
symptoms--a prospective study. Fam Pract 2005; 22(1):58-61.
[13] Jones R, Rubin G, Hungin P. Is the two week rule for cancer referrals working? BMJ 2001;
322(7302):1555-6.
Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 9
 NHS Fife Research Study Guide: [17] How to Write an Abstract.

[14] Flashman K, O'Leary DP, Senapati A, Thompson MR. The Department of Health's "two week
standard" for bowel cancer: is it working? Gut 2004; 53(3):387-91.
[15] Trickett JP, Donaldson DR, Bearn PE, Scott HJ, Hassall AC. A study on the routes of referral for
patients with colorectal cancer and its affect on the time to surgery and pathological stage.
Colorectal Dis 2004; 6(6):428-31.
[16] Results of the first round of a demonstration pilot of screening for colorectal cancer in the United
Kingdom BMJ 2004; 329(7458):133.
[17] Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in patients with iron-deficiency
anemia. N Engl J Med 1993; 329(23):1691-5.
[18] Kepczyk T, Kadakia SC. Prospective evaluation of gastrointestinal tract in patients with iron-
deficiency anemia. Dig Dis Sci 1995; 40(6):1283-9.

Table 1. Summary statistics for haemoglobin concentration in 143 patients with

colorectal cancer, by gender and site of cancer.

All Right- Left- P Non- Rectal P

cases sided sided rectal
Females (n) 58 22 36 37 21
Haemoglobin (g/dl):
Mean 12.0 10.7 12.9 11.3 13.4
Median 12.0 10.3 13.5 11.5 13.8
Standard deviation 1.9 1.6 1.5 1.7 1.3
Minimum 8.7 8.7 9.7 8.7 10.6
Maximum 14.8 13.8 14.8 14.8 14.7

% anaemic, lab reference * 50.0 77.3 33.3 0.001 67.6 19.0 <0.001
% meeting national referral 15.5 36.4 2.8 0.001 24.3 0 0.014
guideline **

Males (n) 85 25 60 38 47
Haemoglobin (g/dl):
Mean 12.7 11.8 13.0 12.0 13.2
Median 13.3 11.4 13.5 11.7 13.6
Standard deviation 2.2 2.3 2.0 2.3 1.9
Minimum 7.0 7.0 8.0 7.0 8.0
Maximum 16.8 15.8 16.8 15.8 16.8

% anaemic, lab reference * 48.2 64.0 41.7 0.060 60.5 38.3 0.041
% meeting national referral 23.5 44.0 15.0 0.004 36.8 12.8 0.009
guideline **

* Lab reference, normal range 12-16 g/dl females, 13-18 g/dl males. Anaemia defined as Hb <12g/dl
(females) and <13 g/dl (males).

** National referral guidelines, referral advised if Hb <10g/dl (females) or <11g/dl (males).

Copyright NHS Fife, Research & Development Office, Dr David Chinn, v2, 4th October 2015 Page 10