You are on page 1of 9

Support Care Cancer (2014) 22:307314

DOI 10.1007/s00520-013-1963-8

ORIGINAL ARTICLE

Predictors of anxiety and depression in people


with colorectal cancer
Nicola M. Gray & Susan J. Hall & Susan Browne &
Marie Johnston & Amanda J. Lee & Una Macleod &
Elizabeth D. Mitchell & Leslie Samuel & Neil C. Campbell

Received: 16 January 2013 / Accepted: 27 August 2013 / Published online: 1 October 2013
# Springer-Verlag Berlin Heidelberg 2013

Abstract 0.015) or diarrhoea (p =0.021), reporting a high negative life


Background People living with colorectal cancer are at risk of and emotional impact (p <0.001) and having difficulties with
anxiety and depression. We investigated what factors were finance (p =0.007). Having neo-adjuvant radiotherapy in-
most highly associated with these. creased the odds of depression caseness (p =0.007), as did
Methods Four hundred and ninety-six people with colorectal poorer physical (p =0.007), cognitive (p <0.001) and social
cancer completed the Hospital Anxiety and Depression Scale (p < 0.001) functioning, having constipation (p = 0.011),
(HADS). Data on functioning, symptoms, illness perceptions reporting a high negative life and emotional impact (p <
and social difficulties were collected by questionnaire. Case- 0.001), having difficulties with personal care (p =0.022) and
note-identified disease, treatment and co-morbidity data were communicating with others (p =0.014).
recorded. Multiple logistic regression identified factors inde- Conclusion Levels of anxiety caseness were similar to those
pendently predictive of anxiety and depression caseness. of non-clinical samples, but depression caseness was higher,
Results Self-reported history of anxiety/depression predicted particularly in those who had received neo-adjuvant radiother-
anxiety but not depression caseness. Depression caseness apy. Most factors associated with possible or probable depres-
predicted anxiety caseness (p =0.043), as did poorer self- sion may be modified with appropriate intervention.
reported cognitive functioning (p =0.001), dyspnoea (p =
Keywords Anxiety . Depression . Colorectal cancer
Electronic supplementary material The online version of this article
(doi:10.1007/s00520-013-1963-8) contains supplementary material,
which is available to authorized users. Introduction
N. M. Gray (*) : S. J. Hall : M. Johnston : A. J. Lee :
N. C. Campbell An ageing population combined with improvements in the
Division of Applied Health Sciences, University of Aberdeen,
detection of cancer has led to increasing numbers of people
Polwarth Building, Aberdeen AB25 2AY, UK
e-mail: n.gray@abdn.ac.uk being diagnosed with cancer. Better treatments mean more
people are being cured of, or are living longer with, the
S. Browne disease. Recent predictions suggest that in the next 30 years,
General Practice and Primary Care, University of Glasgow,
1 Horselethill Road, Glasgow G12 9LX, UK
almost a quarter of people aged 65 or over will be cancer
survivors, with resultant increased demands made on health
U. Macleod services [1]. Research shows that individuals diagnosed with
Hull York Medical School, University of Hull, Hertford Building, cancer experience ongoing poor health make greater use of
Hull HU6 7RX, UK
health services and report significantly impaired health-related
E. D. Mitchell quality of life [2, 3]. These people are also likely to be socially
Social Dimensions of Health Institute, Universities of Dundee and St and financially disadvantaged [4]. Colorectal cancer is the
Andrews, Airlie Place, Dundee DD1 4HJ, UK third commonest cancer in men and in women in the UK
(40,700 people were diagnosed in 2010, http://info.
L. Samuel
Department of Clinical Oncology, Aberdeen Royal Infirmary, cancerresearchuk.org/cancerstats/types/bowel/index.htm).
Foresterhill Road, Aberdeen AB25 2ZP, UK More than half of all people diagnosed with colorectal cancer
308 Support Care Cancer (2014) 22:307314

will be alive at 5 years, and changing population demographics stage and cancer treatments received) are important because
will increase the number living with this cancer, which is they may alert clinical staff to populations at particularly high
commonest in older age. Treatment for colorectal cancer risk of anxiety and depression. Other factors, which may be
almost always involves surgery, can result in stomas and is amenable to intervention, are important because they may
frequently accompanied by chemotherapy, radiotherapy or both offer opportunities for treatment. For example, it may be
[5]. The adverse effects of both the disease and its treatment can possible to influence or change peoples perceptions of their
be prolonged and can impact adversely on quality of life. illness, which could potentially alleviate anxiety or depression
People with colorectal cancer report long-lasting physical [25].
and emotional difficulties including lack of energy, bowel The overall aim of this paper is to describe the levels of
problems, dyspnoea, gastrointestinal problems, sexual dys- anxiety and depression among people with colorectal cancer
function, anxiety and depression [68]. Symptoms of anxiety and to identify factors which are associated with them. Symp-
and depression in particular are common in survivors of toms of anxiety and depression can exist along a spectrum of
cancer [9]: anxiety has been reported for 8 to 23 % of people severity, but most clinicians and researchers define people as
with colorectal cancer and depression for 16 to 39 % [1013]. cases of anxiety or depression if their symptoms exceed
Whilst other features of health-related quality of life tend to particular levels. We will define cases of anxiety and depres-
improve over time in survivors of colorectal cancer [2], de- sion, using predefined levels on the Hospital Anxiety Depres-
pression tends to be longer lasting [14] and there is some sion Scale [26]. Our specific objectives are to (1) describe the
evidence to suggest that depression, and psychological co- prevalence of anxiety and depression caseness among people
morbidity generally, is predictive of mortality in cancer pa- with a confirmed diagnosis of colorectal cancer and (2) iden-
tients [15, 16]. It is important, therefore, that anxiety and tify fixed and modifiable factors which are associated with
depression, and the factors which increase the risk of these, anxiety and depression caseness.
are identified and tackled early.
Several factors have been found to be associated with
anxiety and depression in people with cancer [17]. Women Materials and methods
and younger people experience more anxiety and depression
[18, 19]. Symptoms especially fatigue, dyspnoea, nausea and Participants
anorexia are associated with both anxiety and depression [18,
20]. Advanced disease stage and more complex treatment are This study was nested within the PICT study, full details of
associated with anxiety [12, 19, 21, 22]. These factors explain which are described elsewhere [2].
little of the variance in levels of anxiety and depression, Two groups of patients were eligible for the study. The first
however, so researchers have called for other factors predic- group consisted of patients with colorectal cancer who had
tive of anxiety and depression in both cancer and general commenced their initial treatment and were less than 26 weeks
populations to be investigated [18]. In studies assessing the from diagnosis. They were recruited in North East Scotland,
association between health-related quality of life (HRQoL) via the colorectal cancer multidisciplinary team, and in Glas-
and anxiety and depression in patients with colorectal cancer, gow, via the colorectal specialist nurses. The second group
poorer functioning in a variety of ways (physical, cognitive, consisted of patients who were being followed up for colo-
emotional and social) was associated with increased psycho- rectal cancer diagnosed at least 48 weeks but no longer than
logical distress [10, 12, 23]. There is some evidence that social 2 years previously. They were recruited in North East Scotland
difficulties such as lack of social support are associated with via colorectal oncology and surgical outpatient clinics. Pa-
negative affect whereas positive beliefs, including optimism tients were excluded if they were unable to give informed
and belief in personal control, are associated with positive consent or complete the questionnaire (e.g. due to dementia)
affect [24]. or were expected by their clinicians to have a life expectancy
Some of these factors may be responsive to intervention, so of less than 1 month.
it is important to investigate whether, and how much, they Eligible patients were first approached by a member of
contribute to anxiety and depression. Ideally, all of the factors their clinical team and then by the researchers who sought
found previously to be associated with anxiety and depression written informed consent for participation. Participants were
(summarised above) should be included in a single analysis. asked to complete a questionnaire booklet. Most were com-
This means that data are needed not only on demographic, pleted during face-to-face interviews with a researcher, but
disease and treatment factors but also on symptoms, function- some were self-completed or completed by telephone if par-
ing, social difficulties and beliefs about the illness. Statistical ticipants lived long distances away (e.g. in Orkney or Shet-
modelling should help to identify which factors are indepen- land) or stated a preference.
dently associated with anxiety and depression and, therefore, All participants were recruited between 1 June 2006 and 31
most important. Fixed factors (e.g. sex, age group, disease May 2008. This project was reviewed and fully approved by
308 Support Care Cancer (2014) 22:307314

will be alive at 5 years, and changing population demographics stage and cancer treatments received) are important because
will increase the number living with this cancer, which is they may alert clinical staff to populations at particularly high
commonest in older age. Treatment for colorectal cancer risk of anxiety and depression. Other factors, which may be
almost always involves surgery, can result in stomas and is amenable to intervention, are important because they may
frequently accompanied by chemotherapy, radiotherapy or both offer opportunities for treatment. For example, it may be
[5]. The adverse effects of both the disease and its treatment can possible to influence or change peoples perceptions of their
be prolonged and can impact adversely on quality of life. illness, which could potentially alleviate anxiety or depression
People with colorectal cancer report long-lasting physical [25].
and emotional difficulties including lack of energy, bowel The overall aim of this paper is to describe the levels of
problems, dyspnoea, gastrointestinal problems, sexual dys- anxiety and depression among people with colorectal cancer
function, anxiety and depression [68]. Symptoms of anxiety and to identify factors which are associated with them. Symp-
and depression in particular are common in survivors of toms of anxiety and depression can exist along a spectrum of
cancer [9]: anxiety has been reported for 8 to 23 % of people severity, but most clinicians and researchers define people as
with colorectal cancer and depression for 16 to 39 % [1013]. cases of anxiety or depression if their symptoms exceed
Whilst other features of health-related quality of life tend to particular levels. We will define cases of anxiety and depres-
improve over time in survivors of colorectal cancer [2], de- sion, using predefined levels on the Hospital Anxiety Depres-
pression tends to be longer lasting [14] and there is some sion Scale [26]. Our specific objectives are to (1) describe the
evidence to suggest that depression, and psychological co- prevalence of anxiety and depression caseness among people
morbidity generally, is predictive of mortality in cancer pa- with a confirmed diagnosis of colorectal cancer and (2) iden-
tients [15, 16]. It is important, therefore, that anxiety and tify fixed and modifiable factors which are associated with
depression, and the factors which increase the risk of these, anxiety and depression caseness.
are identified and tackled early.
Several factors have been found to be associated with
anxiety and depression in people with cancer [17]. Women Materials and methods
and younger people experience more anxiety and depression
[18, 19]. Symptoms especially fatigue, dyspnoea, nausea and Participants
anorexia are associated with both anxiety and depression [18,
20]. Advanced disease stage and more complex treatment are This study was nested within the PICT study, full details of
associated with anxiety [12, 19, 21, 22]. These factors explain which are described elsewhere [2].
little of the variance in levels of anxiety and depression, Two groups of patients were eligible for the study. The first
however, so researchers have called for other factors predic- group consisted of patients with colorectal cancer who had
tive of anxiety and depression in both cancer and general commenced their initial treatment and were less than 26 weeks
populations to be investigated [18]. In studies assessing the from diagnosis. They were recruited in North East Scotland,
association between health-related quality of life (HRQoL) via the colorectal cancer multidisciplinary team, and in Glas-
and anxiety and depression in patients with colorectal cancer, gow, via the colorectal specialist nurses. The second group
poorer functioning in a variety of ways (physical, cognitive, consisted of patients who were being followed up for colo-
emotional and social) was associated with increased psycho- rectal cancer diagnosed at least 48 weeks but no longer than
logical distress [10, 12, 23]. There is some evidence that social 2 years previously. They were recruited in North East Scotland
difficulties such as lack of social support are associated with via colorectal oncology and surgical outpatient clinics. Pa-
negative affect whereas positive beliefs, including optimism tients were excluded if they were unable to give informed
and belief in personal control, are associated with positive consent or complete the questionnaire (e.g. due to dementia)
affect [24]. or were expected by their clinicians to have a life expectancy
Some of these factors may be responsive to intervention, so of less than 1 month.
it is important to investigate whether, and how much, they Eligible patients were first approached by a member of
contribute to anxiety and depression. Ideally, all of the factors their clinical team and then by the researchers who sought
found previously to be associated with anxiety and depression written informed consent for participation. Participants were
(summarised above) should be included in a single analysis. asked to complete a questionnaire booklet. Most were com-
This means that data are needed not only on demographic, pleted during face-to-face interviews with a researcher, but
disease and treatment factors but also on symptoms, function- some were self-completed or completed by telephone if par-
ing, social difficulties and beliefs about the illness. Statistical ticipants lived long distances away (e.g. in Orkney or Shet-
modelling should help to identify which factors are indepen- land) or stated a preference.
dently associated with anxiety and depression and, therefore, All participants were recruited between 1 June 2006 and 31
most important. Fixed factors (e.g. sex, age group, disease May 2008. This project was reviewed and fully approved by
Support Care Cancer (2014) 22:307314 309

the Multi-Centre Research Ethics Committee for Scotland, individual items can be used for analysis too. We were inter-
Committee A. ested in the individual factors associated with anxiety and
depression, so we analysed items individually.
Materials Additional data were collected by questionnaire and by
data abstraction from hospital and general practice (GP) case
The main outcomes considered were anxiety and depression notes. The questionnaire booklet asked questions about par-
scores on the Hospital Anxiety and Depression Scale (HADS) ticipants other illnesses (past or present), smoking habits and
[26]. The HADS was designed as a self-assessment scale to whether they were currently experiencing pain as well as
detect depressed and anxious mood in general medical outpa- collecting information on date of birth, postcode, living ar-
tient settings, although it is now widely used in non-hospital rangements (i.e. living on own, living with spouse, living with
settings and in research studies. It consists of 14 questions, others), dependents, ethnicity, education, employment status,
seven measuring anxiety and seven measuring depression home ownership and annual income. Participants postcode
yielding two subscale scores. Each question is scored from 0 was used to allocate participants to an area-based measure of
to 3, with a higher score indicating a higher level of mood deprivation (Carstairs quintile) (http://www.isdscotland.org/
disorder. In a review of 747 studies which used HADS, it was Products-and-Services/Deprivation/Recommendations-and-
found to have good validity and reliability when assessing Downloads/). Participants were categorised as living in an
anxiety and depression caseness [27]. The mean Cronbachs urban or rural area, based on home postcode, using the
alpha for the anxiety subscale was 0.83 and for the depression Scottish Executive Urban Rural 6-category Classification
subscale 0.82. (20052006) (http://www.scotland.gov.uk/Publications/2006/
HRQoL was measured using the EORTC-QLQ C30 07/31114822/UR2006downloads). Travelling time from
(http://groups.eortc.be/qol/eortc-qlq-c30), a widely used participants home to their nearest cancer centre was also
instrument which provides measures of overall HRQoL, five used to provide a proxy for rural or urban residence.
aspects of functioning (physical, role, cognitive, emotional Information on disease stage, treatments undertaken and co-
and social) and eight symptoms commonly reported by morbidities was collected from participants medical case
cancer patients (fatigue, pain, nausea and vomiting, dyspnoea, notes.
loss of appetite, insomnia, constipation and diarrhoea), and
includes a question on the perceived financial impact of the Statistical analysis
disease. It has been found in numerous studies to have good
validity and reliability [28]. HADS anxiety and depression subscale scores were not nor-
Participants perceptions of their illness were measured mally distributed so they were categorised into non-case
using the Revised Illness Perception Questionnaire (IPQR) (score less than 8) and case (score of 8 or more). A score of
[29]. A high score on the timeline, timeline cyclical, conse- 8 was used as the cut-off as this has been found to optimise
quences and emotional representations subscales represents a sensitivity and specificity for the HADS as a screening instru-
negative view of how long the illness will last, how cyclical ment for anxiety disorders and depression [27]. Defined in this
the illness is and the impact of the illness on a persons life and way, anxiety and depression caseness is not synonymous
emotions. A high score on the personal control, treatment with a clinical diagnosis of anxiety or depression but provides
control and illness coherence subscales represents a positive a good indicator of its level in a population. These cut points
view of how much control a person feels they have over their are commonly used in research to allow comparison with
illness, how well the treatment they receive will control their other research in both cancer and general populations. Each
illness and how well they understand their illness. The IPQR of the eight symptom scales of the EORTC-QLQ C30 was
has been found to have good validity and reliability split into two categories (had symptom versus no symp-
(Cronbachs alpha for its different subscales 0.79 to 0.89) [29]. tom), the functioning subscale scores were kept as continuous
The Social Difficulties Inventory (SDI) was used to iden- variables. We excluded the emotional functioning scale from
tify individuals experiencing social difficulties. The SDI was analysis because it has been found to be predominantly a
originally designed to provide an assessment of social diffi- measure of anxiety [10], and we excluded the global quality-
culties experienced by cancer patients [30]. It is a 21-item of-life scale because anxiety and especially depression are
questionnaire covering a wide range of issues, including dif- considered integral to it. The scores on the Revised IPQR
ficulties with personal care, dependents, finance, communica- were kept continuous. Each of the 21 questions of the Social
tion, sexual matters and mobility. Both individual items and Difficulties Inventory was dichotomised into no difficulties
summated scales have been shown to have good validity and (response 0) or some difficulties (responses 13) because
testretest reliability [31]. Internal consistency for summated responses were skewed.
scales is good (Cronbachs alpha ranged from 0.73 to 0.86), The chi-square test was used to test whether there were
but the validation studies have shown that responses to significant associations between anxiety and depression
310 Support Care Cancer (2014) 22:307314

caseness and (1) socio-demographic and lifestyle factors, (2) or depression (currently or in the past) and had pain were most
disease and treatment factors, (3) co-morbidities (participant likely to be classed as anxiety cases (Supplementary Table 5).
reported and case note recorded) and (4) categorised question- The individual models of questionnaire responses identified
naire responses from the QLQ C30 and SDI. The independent those with poorer cognitive functioning and symptoms of
t test (or non-parametric equivalent MannWhitney U ) was nausea and/or vomiting, dyspnoea, sleep disturbance or diar-
used to compare differences in mean (or median where appro- rhoea (from the QLQ C30), those attributing a high number of
priate) scores across the QLQ C30 functioning subscales and symptoms to their illness, those with a poorer understanding of
IPQR subscale scores. Variables with a conservative p value their illness or whose illness had negative emotional conse-
of 0.1 were included in the regression model described quences (from the IPQR) and those having difficulties with
below. carrying out domestic chores, finance, communicating with
Multiple logistic regression was used to investigate which others, living arrangements and feeling isolated (from the
factors were independently associated with anxiety and de- SDI) were most at risk of being an anxiety case. Scoring 8 or
pression caseness. For both anxiety and depression, forward more on the depression subscale of the HADS also significantly
stepwise logistic regression was used to create separate increased the risk of being an anxiety case (Supplementary
models for (1) socio-demographic and lifestyle factors, (2) Table 5).
disease and treatment factors, (3) co-morbidities (participant When all of these predictive factors were considered to-
reported and case note recorded) and (4) questionnaire re- gether, the significant predictors of anxiety caseness which
sponses to the QLQ-C30, IPQR and SDI. remained statistically significant were self-reported anxiety or
Each of the significant predictors identified from the four depression (currently or previously), poorer cognitive func-
grouped models was then used as potential predictors in the tioning scores on the QLQ C30, having dyspnoea or diar-
final model. The fixed factors were entered into the models rhoea, scoring the emotional impact highly on the IPQ-R,
first; the potentially modifiable factors were then added to reporting financial difficulties on the SDI and having a HADS
assess the contribution of each of these to the overall variance depression subscale score of 8 or more (Table 1). These factors
explained by the models. A pseudo R 2 (Nagelkerke R 2) is explained almost half of the variance in anxiety caseness
reported to help clarify the contribution of each factor. The (Nagelkerke R 2 =0.489).
final models were re-run using the full dataset to minimise
cases with missing data. Depression caseness

The mean depression subscale score was 4.07 (SD 3.76).


Results Using a score of 8 or more as a cut-off, 83 of 491(17 %)
participants were classed as cases. Of those scoring more than
A total of 496 participants were included in the analysis, 187 8 on the anxiety subscale, 44 (46 %) also scored 8 or more on
from Glasgow and 309 from North East Scotland (including the depression subscale.
the remote northern islands of Orkney and Shetland). The From the initial individual models, people who did not own
mean age of participants was 66 (standard deviation (SD) their own homes were newly diagnosed and who had pre-
11.1 years); 72 % of participants were over the age of 60. More operative (neo-adjuvant) radiotherapy were at increased risk
men than women participated in the study, 55 versus 45 %. of depression caseness. The same co-morbidities which were
Further details on participant characteristics and variables con- predictive of anxiety caseness were also predictive of depres-
sidered in the analyses are provided in Supplementary Tables 1 sion caseness, namely self-reported anxiety or depression and
to 4. current pain (Supplementary Table 6).
Three participants were excluded from the anxiety analyses From the individual questionnaire models, participants
and five from the depression analyses because of missing reporting poorer physical, cognitive and social functioning
responses to questions on the HADS. and symptoms of pain and constipation (from the QLQ
C30), those who attributed a high number of symptoms to
Anxiety caseness their illness, those with a negative view of how long their
illness would last and those who reported large consequences
The mean anxiety score was 4.32 (SD 4.08); however, 97 of on their life, perceived less control over their treatment and
493 (20 %) participants were categorised as cases on the reported high emotional consequences (from the IPQR) were
anxiety subscale using a score of 8 or more as a cut-off. more likely to be depression cases. From the SDI model,
From our initial models of the socio-demographic and reporting difficulties with domestic chores, personal care,
lifestyle factors and the disease and treatment factors, we communicating with others, sexual matters, getting around
identified those who were unemployed, living in the most and isolation increased the risk of being a depression case.
deprived areas, current smokers, had reported having anxiety Scoring 8 or more on the anxiety subscale of the HADS also
Support Care Cancer (2014) 22:307314 311

Table 1 Associations between the HADS anxiety subscale and the socio-demographic, lifestyle, disease and treatment factors, co-morbidities
(participant reported and case note recorded) and questionnaire responses to the QLQ-C30, IPQR and SDI

Univariate analysis Multivariate analysis

Odds ratio (95 % p value Odds ratio (95 % p value Cumulative


confidence interval) confidence interval) Nagelkerke R 2

Fixed variables
No self-reported history of anxiety or depression 1.00 (ref) 1.00 (ref)
Self-reported history of anxiety or depression 5.27 (2.929.52) <0.001 2.36 (1.065.29) 0.036 0.094
Potentially modifiable variables
Cognitive functioning (QLQ C30) 0.96 (0.950.97) <0.001 0.98 (0.960.99) 0.001 0.271
No dyspnoea (QLQ C30) 1.00 (ref) 1.00 (ref)
Dyspnoea 4.07 (2.566.46) <0.001 2.14 (1.163.93) 0.015 0.301
No diarrhoea (QLQ C30) 1.00 (ref) 1.00 (ref)
Diarrhoea 2.51 (1.603.96) <0.001 2.01 (1.113.65) 0.021 0.317
Emotional representations (IPQR) 1.30 (1.221.38) <0.001 1.25 (1.161.34) <0.001 0.461
No difficulties with finance (SDI) 1.00 (ref) 1.00 (ref)
Difficulties with finance 4.34 (2.647.14) <0.001 2.45 (1.284.68) 0.007 0.480
No HADS depression caseness 1.00 (ref) 1.00 (ref)
HADS depression caseness 8.12 (4.8013.73) <0.001 2.07 (1.034.17) 0.043 0.489

significantly increased the risk of being a depression case Scotland (http://www.scotland.gov.uk/Publications/2011/10/


(Supplementary Table 6). 26083042/6). North East Scotland comprises many rural
When all the predictive factors from the individual models areas, some of which are very remote, as well as the Orkney
were entered into the final model, we found that having had and Shetland Islands. We made considerable efforts to be
pre-operative (neo-adjuvant) radiotherapy increased the risk inclusive and recruit a consecutive sample, but our findings
of being a depression case. Poorer scores on the physical, do not represent people with a very short life expectancy. The
cognitive and social functioning subscales of the QLQ C30 age, sex and disease staging distribution of the recruited sample
and having constipation also increased the risk of depression approximated that of the Scotland wide distribution for 2007
caseness. Participants reporting that their disease had negative (www.isdscotland.org/cancer). However, the participants in our
emotional consequences (IPQR) and those with difficulties sample were slightly younger (28 % of our sample were aged
with personal care and communicating with others were also less than 60 years compared to 17 % of those diagnosed in
more likely to be depression cases (Table 2). Almost 60 % of Scotland in 2007) with fewer very old participants (11 versus
the variance in depression caseness was explained by these 25 % were aged 80 years or more). Our main outcome measure,
factors (Nagelkerke R 2 =0.596). the Hospital Anxiety and Depression Scale, was originally
designed as a screening questionnaire [26]; diagnosis of
clinical anxiety or depression also requires clinical assessment.
Discussion The questionnaire does, however, provide a useful objective
indication of those at high risk. It has been used extensively in
We found that a fifth of participants had possible or probable both research and clinical practice, so it allows our findings to be
anxiety and nearly a fifth had possible or probable depression. compared to those of others both in cancer and non-cancer
Anxiety caseness was more likely in those who were classed as populations.
depression cases (and vice versa), but the other factors that
predicted the chances of anxiety and depression caseness differed. Our findings in the context of other studies

Strengths/limitations of our study Compared to a general UK population sample [32], our sam-
ple had fewer people scoring 8 or more on the anxiety
The study was conducted in two health board areas of Scotland, subscale (20 versus 33 %) but more people scoring 8 or above
covering diverse regions in terms of deprivation and rurality. on the depression subscale (17 versus 12 %). The levels of
Glasgow comprises some of the most deprived populations in caseness in our sample were comparable to those reported in
312 Support Care Cancer (2014) 22:307314

Table 2 Associations between the HADS depression subscale and the socio-demographic, lifestyle, disease and treatment factors, co-morbidities
(participant reported and case note recorded) and questionnaire responses to the QLQ-C30, IPQR and SDI

Univariate analysis Multivariate analysis

Odds ratio (95 % p value Odds ratio (95 % p value Cumulative


confidence interval) confidence interval) Nagelkerke R 2

Fixed variables
No pre-operative radiotherapy 1.00 (ref) 1.00 (ref)
Pre-operative radiotherapy 1.53 (0.882.66) 0.134 2.94 (1.336.48) 0.007 0.023
Potentially modifiable variables
Physical functioning (QLQ C30) 0.95 (0.940.96) <0.001 0.98 (0.960.99) 0.007 0.278
Cognitive functioning (QLQ C30) 0.95 (0.940.96) <0.001 0.97 (0.950.99) <0.001 0.406
Social functioning (QLQ C30) 0.96 (0.950.97) <0.001 0.97 (0.960.98) <0.001 0.498
No constipation (QLQ C30) 1.90 (1.143.14) <0.001 1.00 (ref)
Constipation 2.67 (1.265.66) 0.011 0.521
Emotional representations (IPQR) 1.23 (1.171.31) <0.001 1.16 (1.071.26) <0.001 0.570
No difficulties with personal care (SDI) 1.00 (ref) 1.00 (ref)
Difficulties with personal care 6.63 (3.9911.00) <0.001 2.38 (1.144.98) 0.022 0.583
No difficulties communicating with others (SDI) 1.00 (ref) 1.00 (ref)
Difficulties communicating with others 6.63 (3.9011.26) <0.001 2.60 (1.215.57) 0.014 0.596

other samples of cancer patients. Strong et al. [11] report financial difficulties. These are similar risk factors to those
anxiety caseness of 23 % in a sample of 3,071 cancer patients of the general population and the prevalence of anxiety
(of whom 15 % were bowel cancer patients) using a cut-off appeared similar. Anxiety was more common in people with
score of 9 or more (15 % of our sample scored 9 or more on symptoms of dyspnoea or diarrhoea. We cannot say from this
the anxiety subscale) [11]. In the sample of Strong et al., 16 % study whether dyspnoea and diarrhoea caused anxiety or were
scored 8 or more on the depression subscale, compared to the a response to it. Both of these symptoms may, however, be
17 % in our sample. A systematic review has also found related to colorectal cancer and may be amenable to treatment.
depressive symptoms to be common in long-term cancer Cases of depression were more frequent among those who
survivors [9]. Taken together, these findings suggest that had received pre-operative (neo-adjuvant) radiotherapy. Pa-
depression is increased by a diagnosis of colorectal cancer tients with more locally advanced rectal cancer often receive
but levels of anxiety remain similar to the general population. pre-operative radiotherapy. It is usually given with concomi-
There are several studies reporting associations of anxiety tant chemotherapy and can be associated with more complex
and depression with particular factors among people with co- surgery, stoma formation and side effects such as sexual
lorectal cancer. For example, Simon et al. [21] found signifi- dysfunction. These factors were not found to be significant
cantly higher rates of anxiety and depression in those with more individually in our analysis, but their combination, along with
advanced stage disease and Medeiros et al. [13] for those the additional length of time in treatment (typically 5 weeks of
receiving chemotherapy [13, 21]. Alacacioglu et al. [12] report- radiotherapy and a further 6 weeks before surgery), may
ed associations between nearly all EORTC-QLQ C30 symp- explain our finding. We had insufficient numbers to test
toms and functioning scales and anxiety and depression [12]. whether short-course radiotherapy had less impact on depres-
We included all these variables in our analysis and attempted to sion and this should be investigated in future research.
look as comprehensively as we could at demographic, disease, Most of the remaining factors that we found to be indepen-
treatment, social and psychological factors. By conducting dently associated with possible or probable depression have
modelling, we were able to identify which factors were inde- potential for modification by appropriate treatment. In partic-
pendently significant so most proximally associated with, and ular, physical functioning can be improved by interventions
of most importance to, anxiety and depression. targeting physical fitness and social functioning and difficul-
ties with personal care have been improved by appropriate
Meaning and implications case management [33, 34]. Constipation was common and can
be treated simply and effectively with both diet and laxatives.
Unsurprisingly, cases of anxiety were more frequent among It was perhaps unsurprising that those with heightened emo-
those with previous anxiety or depression or reporting tional representations of their illness should be more at risk of
Support Care Cancer (2014) 22:307314 313

depression (and anxiety); this was targeted effectively by the 2. Gray NM, Hall SJ, Browne S, Macleod U, Mitchell E, Lee AJ,
Johnston M, Wyke S, Samuel L, Weller D, Campbell NC (2011)
Edinburgh intervention Depression Care for People with
Modifiable and fixed factors predicting quality of life in people with
Cancer [35]. colorectal cancer. Br J Cancer 104:16971703. doi:10.1038/bjc.
The implications of our findings for clinicians are that both 2011.155
anxiety and depression are common among people with colo- 3. Elliott J, Fallows A, Staetsky L, Smith PW, Foster CL, Maher EJ,
Corner J (2011) The health and well-being of cancer survivors in the
rectal cancer, but it is depression whose prevalence is partic-
UK: findings from a population-based survey. Br J Cancer 105(Suppl
ularly increased. Nearly all associated factors are modifiable 1):1120. doi:10.1038/bjc.2011.418
with treatment, so early identification and effective treatment 4. Department of Health (2012) Quality of life of cancer survivors in
has potential to improve patient care. Clinicians should pay England. Report on a pilot survey using Patient Reported Outcome
Measures (PROMS). National Institute for Health Research, London
particular attention to patients with previous anxiety, receiving
5. Scottish Intercollegiate Guideline Network (SIGN) (2011) Diagnosis
neo-adjuvant radiotherapy, and with functional impairments and management of colorectal cancer. Guideline number 126. SIGN,
and communication difficulties because they are more likely Edinburgh
to have anxiety or depression, as are those with symptoms 6. Phipps E, Braitman LE, Stites S, Leighton JC (2008) Quality of life
and symptom attribution in long-term colon cancer survivors. J Eval
(constipation, diarrhoea and dyspnoea). From this cross-
Clin Pract 14:254258. doi:10.1111/j.1365-2753.2007.00842.x
sectional study, we cannot say how much these factors cause 7. Denlinger CS, Barsevick AM (2009) The challenges of colorectal
anxiety and depression or are a consequence. Impaired cogni- cancer survivorship. J Natl Compr Canc Netw 7:883894
tive functioning and difficulties communicating with others, 8. Arndt V, Merx H, Stegmaier C, Ziegler H, Brenner H (2006) Restric-
tions in quality of life in colorectal cancer patients over three years
for example, are common manifestations of depression, but
after diagnosis: a population based study. Eur J Cancer 42:1848
may also contribute to it. A holistic approach to treatment, 1857. doi:10.1016/j.ejca.2006.01.059
tackling not just the anxiety and depression but also their 9. Harrington CB, Hansen JA, Moskowitz M, Todd BL, Feuerstein M
associated symptoms and functional impairments, would ap- (2010) Its not over when its over: long-term symptoms in cancer
pear then to be most likely to be effective. Future research survivorsa systematic review. Int J Psychiatry Med 40:163181
10. Tsunoda A, Nakao K, Hiratsuka K, Yasuda N, Shibusawa M, Kusano
should evaluate formalised approaches to targeted screening M (2005) Anxiety, depression and quality of life in colorectal cancer
and early holistic intervention. patients. Int J Clin Oncol 10:411417
11. Strong V, Waters R, Hibberd C, Rush R, Cargill A, Storey D, Walker
J, Wall L, Fallon M, Sharpe M (2007) Emotional distress in cancer
Conclusion patients: the Edinburgh Cancer Centre symptom study. Br J Cancer
96:868874. doi:10.1038/sj.bjc.6603626
People with colorectal cancer are at particular risk of depres- 12. Alacacioglu A, Binicier O, Gungor O, Oztop I, Dirioz M, Yilmaz U
(2010) Quality of life, anxiety, and depression in Turkish colorectal
sion, especially if they have received neo-adjuvant radiother-
cancer patients. Support Care Cancer 18:417421. doi:10.1007/
apy. Depression is associated with impairments of physical, s00520-009-0679-2
social and cognitive functioning, difficulties with personal 13. Medeiros M, Oshima CT, Forones NM (2010) Depression and anx-
care and constipation, all of which could benefit from iety in colorectal cancer patients. J Gastrointest Cancer 41:179184.
doi:10.1007/s12029-010-9132-5
treatment.
14. Ramsey SD, Berry K, Moinpour C, Giedzinska A, Andersen MR
(2002) Quality of life in long term survivors of colorectal cancer. Am
Acknowledgments The research team would like to thank all of the J Gastroenterol 97:12281234
clinicians who granted access to their clinics. We are deeply grateful to 15. Satin JR, Linden W, Phillips MJ (2009) Depression as a predictor of
those patients who gave their time to help with this research at what may disease progression and mortality in cancer patients. Cancer 115:
have been a very difficult and distressing time. Thanks to our Steering 53495361. doi:10.1002/cncr.24561
Group members Professor Mandy Ryan (University of Aberdeen), Sue 16. Spiegel D, Giese-Davis J (2003) Depression and cancer: mechanisms
Ziebland (University of Oxford), Professor David Weller (University of and disease progression. Biol Psychiatry 54:269282
Edinburgh) and Professor Sally Wyke (University of Glasgow) for their 17. Die Trill M (2012) Psychological aspects of depression in cancer
support and help with this project. The study was funded by a project patients. Ann Oncol 23(Supplement 10):x302x305. doi:10.1093/
grant from Cancer Research UK. The University of Aberdeen is the study annonc/mds350
sponsor. 18. Salvo N et al (2012) Frequency of reporting and predictive factors for
anxiety and depression in patients with advanced cancer. Clin Oncol
Conflict of interest There is no conflict of interests to declare. We (the 24:139e148e
authors) have had full control of all the primary data and are happy for the 19. Vodermaier A, Linden W, MacKenzie R, Greig D, Marshall C (2011)
journal to review this data as required. Disease stage predicts post-diagnosis anxiety and depression only in
some types of cancer. Br J Cancer 105:18141817
20. Rhondali W, Perceau E, Berthiller J, Saltel P, Trillet-Lenoir V, Tredan
O, Coulon JP, Bruera E, Filbet M (2012) Frequency of depression
References among oncology outpatients and association with other symp-
toms. Support Care Cancer 20:27952802. doi:10.1007/s00520-
012-1401-3
1. Maddams J, Utley M, Mller H (2012) Projections of cancer preva- 21. Simon AE, Thompson MR, Flashman K, Wardle J (2009) Disease
lence in the United Kingdom, 20102040. Br J Cancer 107:1195 stage and psychosocial outcomes in colorectal cancer. Colorectal Dis
1202. doi:10.1038/bjc.2012.366 11:1925. doi:10.1111/j.1463-1318.2008.01501.x
314 Support Care Cancer (2014) 22:307314

22. Pereira MG, Figueiredo AP, Fincham FD (2012) Anxiety, depression, 29. Moss-Morris R, Weinman J, Petrie K, Horne R, Cameron L, Buick D
traumatic stress and quality of life in colorectal cancer after different (2002) The Revised Illness Perception Questionnaire (IPQ-R).
treatments: a study with Portuguese patients and their partners. Eur J Psychol Health 17:116. doi:10.1080/08870440290001494
Oncol Nurs 16:227e232e 30. Wright P, Smith A, Roberts K, Selby P, Velikova G (2007) Screening
23. Bang S, Park SH, Kang HG, Jue JI, Cho IH, Yun YH, Cho EK, Shin for social difficulties in cancer patients: clinical utility of the social
DB, Lee JH (2005) Changes in quality of life during palliative difficulties inventory. Br J Cancer 97:10631070
chemotherapy for solid cancer. Support Care Cancer 13:515521 31. Wright EP, Kiely M, Johnston C, Smith AB, Cull A, Selby PJ (2005)
24. Hodges K, Winstanley S (2012) Effects of optimism, social support, Development and evaluation of an instrument to assess social diffi-
fighting spirit, cancer worry and internal health locus of control on culties in routine oncology practice. Qual Life Res 14:373386. doi:
positive affect in cancer survivors: a path analysis. Stress Heal 28: 10.1007/s11136-004-5332-4
408415. doi:10.1002/smi.2471 32. Crawford JR, Henry JD, Crombie C, Taylor EP (2001) Normative
25. Husson O, Thong MS, Mols F, Oerlemans S, Kaptein AA, van de data for the HADS from a large non-clinical sample. Br J Clin
Poll-Franse LV (2013) Illness perceptions in cancer survivors: what is Psychol 40:429434
the role of information provision? Psycho-Oncology 22(3):490498 33. Speck RM, Courneya KS, Masse LC, Duval S, Schmitz KH (2010)
26. Zigmond AS, Snaith RP (1983) The Hospital Anxiety and Depres- An update of controlled physical activity trials in cancer survivors: a
sion Scale. Acta Psychiatr Scand 67:361370. doi:10.1111/j.1600- systematic review and meta-analysis. J Cancer Surviv 4:87100. doi:
0447.1983.tb09716.x 10.1007/s11764-009-0110-5
27. Bjelland I, Dahl AA, Haug TT, Neckelmann D (2002) The validity of 34. Dieterich M, Irving CB, Park B, Marshall M (2010) Intensive case
the hospital anxiety and depression scale: an updated literature review. management for severe mental illness. Cochrane Database Syst Rev
J Psychosom Res 52:6977. doi:10.1016/S0022-3999(01)00296-3 10, CD007906. doi:10.1002/14651858.CD007906.pub2
28. Groenvold M, Klee MC, Sprangers MAG, Aaronson NK (1997) 35. Strong V, Waters R, Hibberd C, Murray G, Wall L, Walker J,
Validation of the EORTC QLQ-C30 quality of life questionnaire McHugh G, Walker A, Sharpe M (2008) Management of depression
through combined qualitative and quantitative assessment of for people with cancer (SMaRT oncology 1): a randomised trial.
patient-observer agreement. J Clin Epidemiol 50:441450 Lancet 372:4048. doi:10.1016/S0140-6736(08)60991-5

You might also like