Psychological Distress among Prostate Cancer Patients: Fact Or Fiction?

Abstract

Although the detrimental effect upon psychological well-being of receiving a diagnosis of, or treatment for, cancer has been demonstrated across many different types of cancer, three recent reviews of the psychological health of prostate cancer patients have produced contradictory conclusions. In order to elucidate the reasons for these apparent different conclusions, each of these reviews is described, with principal methods and findings summarised. Actual data, methodology used to select/reject research studies for inclusion in reviews, plus the validity of strict methodological culling of some research studies are discussed. Several extra studies and commentaries are also described, and a resolution of the apparent contradictory review conclusions is offered.

Full text

Page 1

Clinical Medicine: Oncology 2008:2 563–572
563
REVIEW
Correspondence: Professor C.F. Sharpley, PO. Box 378, Coolangatta, Qld, 4225, Australia.
Email: csharpley@onthenet.com.au
Copyright in this article, its metadata, and any supplementary data is held by its author or authors. It is published under the
Creative Commons Attribution By licence. For further information go to: http://creativecommons.org/licenses/by/3.0/.
Psychological Distress among Prostate Cancer Patients:
Fact Or Fiction?
Christopher F. Sharpley1, Vicki Bitsika2 and David H.R. Christie3
1Centre for Bioactive Discovery in Health and Ageing, University of New England New South
Wales, Australia. 2Bond University, Queensland, Australia. 3Premion, Queensland, Australia.
Abstract: Although the detrimental effect upon psychological well-being of receiving a diagnosis of, or treatment for,
cancer has been demonstrated across many different types of cancer, three recent reviews of the psychological health of
prostate cancer patients have produced contradictory conclusions. In order to elucidate the reasons for these apparent different
conclusions, each of these reviews is described, with principal methods and fi ndings summarised. Actual data, methodology
used to select/reject research studies for inclusion in reviews, plus the validity of strict methodological culling of some
research studies are discussed. Several extra studies and commentaries are also described, and a resolution of the apparent
contradictory review conclusions is offered.
Keywords: prostate cancer, psychosocial distress, depression, anxiety
Introduction
Receiving a diagnosis of cancer has been linked with a reduction in psychological well-being, Quality
of Life, interpersonal relationships and optimism [1,2]. There are also data indicating that cancer patients
who are anxious or depressed have poorer treatment outcomes, perhaps via lower compliance with
medication [3], highlighting the need to address psychological functioning as an integral aspect of
overall cancer treatment protocols. In support of this viewpoint, Greenberg [4: p. 1036] stated that
“Inadequate treatment of major depressive disorders in this day and age is substandard oncology
treatment”, a sentiment echoed in Australia by the NH&MRC Clinical Practice Guidelines for the
Psychosocial Care of Adults with Cancer [5].
However, prior to administration of treatments aimed at alleviating the psychological sequelae of
receiving a diagnosis of, and treatment for, cancer, the reliable identifi cation of the incidence of
psychological distress needs to be determined. Although this has been reported for several cancer patient
groups (e.g. breast cancer [6]), there has been some debate as to the existence of elevated psychologi-
cal distress among prostate cancer patients (PCa), with one review paper arguing that there was no real
difference between the psychological distress of prostate cancer patients and their peers [7], but two
other reviews claiming that the data from many studies indicate the presence of elevated anxiety and
depression, plus lowered Quality of Life, among prostate cancer patients compared to their non-patient
peers [8,9]. In order to resolve this apparent contradiction, the present paper examined the fi ndings from
those three major reviews of the previous research which investigated the presence of psychological
distress among PCa patients. Each of these reviews will be described below, with some comments upon
their methodologies, fi ndings and conclusions. From that examination, overall fi ndings regarding the
presence (or not) of elevated psychological distress among prostate cancer patients will be discussed
and the aforementioned disagreement between the three reviews will be resolved.
While there have been many research reports on the issue of psychological distress among prostate
cancer patients, plus the three reviews mentioned above and described below, there has been no recent
overall summary of the fi eld to date, leaving the issue open to debate. Thus, this paper extends the previous
literature by bringing together the previous major reviews and synthesizing their fi ndings. The present
structure and style of ‘reviewing the reviews’ was chosen instead of an alternate methodology (e.g. meta-
analysis) because the process of gathering individual studies and reviewing them (as in the previous

Page 2

564
Sharpley
Clinical Medicine: Oncology 2008:2
three reviews) has not produced a clear outcome to
the questions of whether or not prostate cancer
patients have higher levels of psychological distress
than their peers. A further review of the original
literature was considered not to be required at this
juncture because performing another similar review
might simply add to the disagreement shown by those
three papers without advancing understanding of the
overall literature to date and thus resolving the appar-
ent disagreement between those previous reviews.
Methods
The construct “psychological distress” was opera-
tionalised by reference to the nomenclature com-
monly used in the DSM-IV-TR [10] for “anxiety”
and “depression” as representing the most common
and researched indices of unhappiness of a psy-
chological nature. In addition, and because they
may be used as indicators of the relative absence
of psychological distress, the global terms “psy-
chological health” and “Quality of Life” were used
to indicate the alternative “healthy” aspects of
psychological status. “Quality of life” has been
widely used in the health literature and may relate
to a range of patient factors, including mental well-
being, physical status and emotional health as
assessed by the patient [11].
From a search of Pub Med, PsychInfo, Google
Scholar to February 2008, using the descriptors
“prostate cancer”, “depression”, “anxiety”, “psycho-
logical health” and “Quality of Life”, plus follow-up
hand searches, three reviews of the literature were
identifi ed. Each of these three reviews was examined
in terms of the articles they described, tabulations of
data, measures of psychological distress used in the
literature and conclusions drawn by the reviews.
Each review was therefore “re-reviewed” by an
examination of the original reports which had been
reviewed and any discrepancies between the original
data reported in the literature and the tabulations,
descriptions or conclusions drawn by the three lit-
erature reviews were noted. On these bases, critical
evaluations of the three literature were compiled and
are presented below.
Results
Review 1: Bennet and Badger (2005)
Taken chronologically, Bennett and Badger [8]
conducted the first comprehensive review of
psychological distress among PCa patients from
studies reported between 1988 and 2004. Those
authors identifi ed 38 studies (27 of which were
descriptive and 11 that used experimental or quasi-
experimental designs) from 39 published articles;
12 studies had samples of PCa patients that were
less than 50 in number, eight studies reported on
samples of between 50 and 100 PCA patients and
18 papers had samples of more than 100 PCa
patients. Bennett and Badger focussed their review
upon the incidence of depression among PCa men
and noted that the Profi le of Mood States was the
most frequently used measure of depression, fol-
lowed by the Hospital Anxiety and Depression
Scale, the Beck Depression Inventory and the Cen-
ter for Epidemiological Studies-Depression scale.
Bennett and Badger organised their review
across fi ve categories of depression incidence:
those studies which reported on the prevalence and
correlates of depression or other mood disorders
(n = 9), papers that included depression in studies
of fatigue and/or pain (4), studies of Quality of Life
(QoL) and depression (12), the comparative inci-
dence of depression among PCA patients and their
partners (4), and intervention studies which used
mood as the principal dependant variable (9). It is
of note that, although Bennett and Badger criticised
some studies on methodological grounds, they still
included them in their review on the basis that these
studies contributed valuable information to the
discussion despite the presence of some method-
ological shortcomings. However, they did indicate
that issues needing to be addressed were: lack of
uniformity among measures used to tap depression,
assessment of mood (via the POMS) rather than
depression per se, low representation of some
ethnic and social-cultural subgroups within the
society where the largest number of research stud-
ies were conducted (U.S.A.); and that these short-
comings limited the generalisability of the overall
conclusions from their review to the wider popula-
tion of PCa patients. Bearing all these points in
mind, Bennett and Badger nevertheless drew the
conclusion that PCa patients experienced depressed
mood at levels “higher than those reported for older
men in the general population” (p. 554). Major risk
factors for depressed mood were advanced PCa,
prominent symptomatology and side effects from
treatment, plus a prior history of depression. PCa
patients in pain were also more likely to suffer
depressed mood, whereas being married, receiving
strong social support, being optimistic and having

Page 3

565
Psychological distress among prostate cancer patients: Fact or fi ction?
Clinical Medicine: Oncology 2008:2
greater physical functioning were buffers against
depressed mood in the samples studied. Of major
import for the development of effective treatment
options for PCa patients, Bennett and Badger noted
that “the state of the science for supportive care
interventions aimed toward men with prostate cancer
is limited” (p. 554), thus emphasising the comments
made by Greenberg (above) and the Guidelines of
the NH&MRC and other similar national bodies that
support provision of psychosocial care for all cancer
patients as part of basic treatment protocols. This
point is particularly relevant to the discussion of
whether PCA patients actually do experience
psychological distress at elevated frequencies
compared to their non-PCa peers.
Review 2: Katz (2007)
The second major review of PCa patients’ psycho-
logical well-being was reported by Katz [9], who
reviewed studies of PCa and Quality of Life (QoL)
from 1999 to 2005, a briefer period than that tapped
by Bennett and Badger, but also including more
recent data. Overall, Katz examined 68 reports of
QoL among PCA patients, and specifi cally evalu-
ated the effects of treatment modality (surgery,
radiation therapy, cryosurgery) and also the effects
of advanced PCa upon QoL. Katz’s conclusion
mirrored that reported by Bennett and Badger:
“Quality of Life is decreased in both the short and
long terms for men with prostate cancer” (p. 302).
However, one difference between the data reviewed
by Bennett and Badger and that which Katz
reviewed was the choice of the principal dependant
variable. As mentioned above, Bennett and Badger
included studies that used measures of mood
(including depression) and QoL, whereas Katz
restricted her review to QoL studies exclusively.
Despite this different database, the conclusions
drawn from these two reviews of PCa patients’
mood and QoL were congruent across these differ-
ent dependant variables and thus form a signifi cant
statement regarding the psychological state of PCa
patients.
It should, however, be noted that the use of QoL
as an indicator of psychological distress may be
fl awed because QoL does not assess the accepted
symptomatology of psychological maladjustment.
Chen [12] commented that QoL measures are
designed to assess the patient’s perception of the
effects of symptoms upon daily role function and
may/may not include the psychological impact of
those symptoms. Thus, QoL measures are not
designed to act as screening instruments for anxiety
or depression and are not validated on agreed
criteria for any disorder, leading to “much disagree-
ment and confusion” about the exact meaning of
QoL, with “researchers often using the same term
(Quality of Life) to mean very different things”
[12: p. 2695]. This means that the exact nature of
QoL data is open to conjecture because of a “lack
of standard criteria” [13, p. 626]. In terms of the
implications of QoL data for intervention, Whelan
et al. [14] noted that “Quality of life data are
diffi cult for clinicians to interpret” because they
do not relate to clinical symptoms (p. 6937). Thus,
it could be argued that the value of QoL data in
determining the psychological health of PCa
patients (and their need for additional treatment)
is limited. This does not invalidate Katz’s review
but does restrict the generalisability of her fi ndings
to the specifi c issue of PCa patients’ psychological
distress levels compared to their non-PCa
counterparts.
By contrast, Nordin et al. [15] commented that,
when investigating psychological distress and
adjustment problems in cancer populations,
“anxiety and depression are the most prevalent”
indices to be assessed (p. 376). Both anxiety and
depression are clearly defi ned in the ICD and
DSM nomenclature and both are of major clinical
signifi cance to patients and patient care. Thus, the
reviews by Bennett and Badger and Katz which
respectively focused on depressed mood and QoL
of PCa patients may be viewed as being limited
in demonstrating that this patient group is similar
to other cancer patients in experiencing increased
levels of psychological distress While mood and
QoL are of value when investigating the general
mental states of participants, they are also some-
what distanced from the key diagnostic criteria
of anxiety and depression [10] that constitute
mental illnesses in themselves (and thereby
represent a major treatment challenge for clini-
cians who deal with PCa patients). Therefore, it
may be that measuring the presence of symptoms
of anxiety and depression directly via ICD- or
DSM-based instruments or procedures could be
more informative than using patients’ refl ections
upon their mood or level of functioning and how
these make them feel. On this basis, the focusing
of research reviews upon those studies which
apply valid and reliable measures of anxiety and
depression to PCa patients is probably necessary

Page 4

566
Sharpley
Clinical Medicine: Oncology 2008:2
to adequately address the issue of PCa patients’
psychological distress.
Review 3: Bloch et al. (2007)
The third major review of PCa patients’ psycho-
logical distress is also of value because it adopted a
different approach to literature selection from that
followed by Bennett and Badger and Katz. Bloch
et al. [7] examined studies on the psychological
effects of PCa that were reported between 1988 and
June, 2006, noting that that “Few of the … investi-
gations have enlisted large samples or controls”
(p. 3), leading those authors to exclude studies that
were “methodologically inadequate” (p. 545).
Applying these criteria, Bloch et al. found just three
methodologically acceptable cross-sectional studies,
plus two longitudinal studies of PCa patients’ psy-
chological distress and concluded from these fi ve
studies that there were “few differences in levels of
psychological distress” (p. 11) between men with
and without PCa on the basis of “studies with rea-
sonably sized and unbiased samples with appropri-
ate designs, including adequate descriptions of
instruments used” (p. 2). Because of its relatively
different conclusions to those from Bennett and
Badger and Katz, it is worth examining the proce-
dures and fi ndings of the Bloch et al. review in some
greater detail to determine the validity of those
conclusions vis-à-vis those of Bennett and Badger
and Katz (even though those reviews were not
specifi cally focussed upon psychological distress as
defi ned in the DSM or ICD terms). As a fi rst step in
that process, it is of relevance to examine the three
cross-sectional and two longitudinal studies upon
which Bloch et al. based their conclusion (Bacon
et al. [16], Clark et al. [17], Helgason et al. [18],
Nordin et al. [15] and Visser et al. [19]).
Bacon et al. [15] reported that PCa patients had
signifi cantly worse scores on 7 out of 10 QoL
subscales than controls but not on the mental
subscale. Although the samples were restricted to
a particular occupational group (i.e. health profes-
sionals) and thus results are limited in their gener-
alisability to the population of PCa patients, data
do indicate that these PCa patients were experienc-
ing a signifi cantly lower overall quality of life than
non-PCa men. Similarly, the widespread negative
attitudes reported by PCa patients in Clark et al.’s
[17] study were highlighted by those authors, who
also commented on the unreliability of focussing
upon the lack of differences between PCa patients
and controls in global measures of health status.
Helgason et al. [18] described their PCa sample as
“significantly more likely to be distressed”
(p. 1418) than non-PCa men due to decreases in
sexual functioning. These three studies constituted
what Bloch et al. described as those which “stand
out since the samples are adequate and compared
with controls” (p. 2), but they do not support Bloch
et al.’s conclusion that “Men with PCa do not
appear to experience marked impairment of
adjustment” (p. 12). In addition, none of the three
cross-sectional studies accepted by Bloch et al. as
methodologically satisfactory actually used valid
and reliable measures of depression or anxiety.
Instead, these three papers used QoL, patient
“attitudes” and “distress” as their DVs. While these
data are valuable in describing the emotional states
of the PCa patients examined, they do not assess
anxiety or depression as defi ned in ICD or DSM
terms.
Of the two longitudinal studies of psychological
adjustment among PCa patients that met Bloch
et al.’s methodological criteria for inclusion,
Nordin et al. [15], collected HADS data from
118 PCa patients close in time to their diagnosis
and then again from 99 of these patients six months
later. Data were also collected from breast, colorec-
tal and gastric cancer patients. Unfortunately, the
HADS data reported were the group means for PCa
and other patient groups rather than the percent
who reached the cutoff score indicating an anxiety
or depression disorder. Thus, it is not possible to
determine the incidence of clinically depressed or
anxious PCa patients at either diagnosis or six
months later. Visser et al. [19], used a QoL instru-
ment as well as a shortened version of the Profi le
of Mood States (POMS). Although the shorter
POMS has a depression subscale, scores on the
latter were not reported.
Thus, the data from the three cross-sectional
and two longitudinal studies referred to by Bloch
et al. as indicating a lack of difference in the psy-
chological distress of PCA patients versus their
non-PC a peers may be of limited value because
four of those studies used QoL as the main DV
rather than measures of clinical anxiety or depres-
sion and also because the single study which did
use a standardised measure of anxiety and depres-
sion (the HADS) did not report incidence data for
the PCa sample which reached clinical levels of
anxiety or depression. In addition, there were, in
fact, several key datum which indicated that PCa

Page 5

567
Psychological distress among prostate cancer patients: Fact or fi ction?
Clinical Medicine: Oncology 2008:2
patients were less happy with their health than
non-PCa peers.
By contrast, of the three cross-sectional papers
that focussed upon depressive symptoms and
which were rejected by Bloch et al. on the basis of
methodological weaknesses (Heim and Oei [20],
Stone et al. [21], Balderson and Towell [22]), Heim
and Oei [20] reported that 32% of their sample of
PCa patients met the criteria for a clinical diagnosis
of depression on the BDI, many times more than
the 3% or less reported in national surveys of the
older male population in Australia, where Heim
and Oei collected their data. Stone et al.’s [21] data
were not directly relevant to this discussion because
PCa patients’ combined anxiety-depression mean
scores (from the HADS) were not reported
separately to other cancer patient groups, but
Balderson and Towell [22] found 38% of their PCa
sample met the criteria for combined anxiety-
depression on the HADS, well above the norms
for this age group.
Discussion
These three reviews demonstrate some uncertainty
on the issue of PCa patients’ psychological distress.
As well as reporting two different conclusions,
each of these reviews adopted a different stand on
selection of studies to be reviewed. Bennett and
Badger [8] included all 39 studies of depression
among PCa patients that they identifi ed via elec-
tronic search and concluded that, while PCa
patients were, on balance, more distressed than
their non-PCa peers, the predominant measures
used were focussed upon mood rather than ICD- or
DSM-based anxiety or depression; Katz [9]
restricted her review to those studies which used
QoL as the principal dependant variable rather than
psychological disorders, again including all studies
regardless of methodological weaknesses and also
concluding that PCa patients had relatively poorer
QoL than non-PCa men. By contrast, Bloch et al.
[7] excluded studies on the basis of strict method-
ological criteria and concluded (on the basis of
those few studies which remained after exclusion)
that there were no signifi cant differences in the
psychological distress levels of PCa patients and
their non-PCa peers.
As well as presenting contradictory conclusions,
the three reviews described above raise several
issues for further examination, principally those
of: choice of DV and measures, selection of control
groups or reference to normative data for purposes
of comparison, the effects of research design vari-
ables upon validity of data reported and sample
size. While the soundness of research methodology
is undeniably an important aspect of selection of
which studies to include in literature reviews, those
methodological criteria need to be carefully con-
sidered before implementation. Specifically,
although (as argued by Bloch et al.) cross-sectional
‘snapshots’ do not provide data on the persistence
of symptoms, they do tell us if there is a discrep-
ancy between the scores of a certain sample on a
selected variable at a point in time when they are
compared to another group or set of norms . While
it would be valuable to know the features of depres-
sion and anxiety over time among PCa patients,
the lack of such longitudinal knowledge does not
preclude clinicians recognising that such distress
exists and that it can hinder treatment and be det-
rimental to patient health and well-being. So,
although a temporal “map” of psychosocial distress
across all periods of PCa diagnosis, treatment and
recovery is clearly of value (and such data have
been reported [23,24]), simple snapshots are not
worthless—they indicate whether this population
ever experiences psychosocial distress that war-
rants intervention.
A major methodological aspect of cross-
sectional and longitudinal studies is the presence
of control groups, which Bloch et al. described as
“crucial” (p. 12). However, this is not necessarily
the case in research which involves psychometric
testing, where the accepted procedure is to refer to
established norms when comparing the scores of
a selected sample. For example, data from the
Australian National Survey of Mental Health and
Well-Being [25] showed that anxiety and depres-
sion decline with age [26], and that the prevalence
of anxiety and depression among men aged 55 to
64 was 6% and 3% respectively, dropping to 3.5%
and ?1.0% after age 65 [27]. Data collected from
a sample of Australian PCa patients (at any time
after their diagnosis) may be profi tably compared
with these norms when drawing conclusions about
the effects of PCa upon patients, thus obviating the
need for another ‘control’ group. Similarly, the
issue of adequate sample size is open to some
further interpretation than simply classifying stud-
ies as methodologically insuffi cient because they
use samples that are smaller than might be ideal.
Power analysis may be profi table where the null
hypothesis has been accepted, but ad hoc rejection