Psychological factors and preferences for communicating prognosis in esophageal cancer patients.

Abstract

Discussing prognosis is often confronting and complex for cancer patients. This study investigates how patients' psychological characteristics relate to their preferences concerning the disclosure of prognosis.
One hundred and seventy-six esophageal cancer patients participated in the study. They had undergone esophagectomy within the past 28 months and did not have evidence of cancer recurrence. Patients completed a questionnaire eliciting their preferences for prognostic information. Sociodemographic characteristics, involvement preferences, anxiety, depression, fear of recurrence, striving for quality of life (QOL) or quantity of life and trust in physicians were explored as predictors for (a) wanting to be informed about prognosis and (b) the initiation of discussion about prognosis.
Patients wanting all prognostic information had more fear for the disease to recur (p<0.05) and were inclined to be more actively involved during consultation (p<0.001). Post hoc analyses showed that patients with worse QOL scores reported more fear of recurrence. Anxiety, depression, trust and tendency to strive for QOL or quantity of life were not related to preferences concerning prognostic information.
The more fear patients have for esophageal cancer to recur, the more information they want about prognosis. Thus, patient's fear for recurrent disease is not a reason for withholding prognostic information. Results also suggest that there is no harm in asking patients what information they want.

Full text

Page 1

Psycho-Oncology
Psycho-Oncology 18: 1199–1207 (2009)
Published online 28 January 2009 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pon.1485
Psychological factors and preferences for communicating
prognosis in esophageal cancer patients
Sanne J. Franssen1Sjoerd M. Lagarde2Jochem R. van Werven2Ellen M. A. Smets1Khe T. C. Tran3
John Th. M. Plukker4J. Jan B. van Lanschot2and Hanneke C. J. M. de Haes1?
1Department of Medical Psychology, Academic Medical Center at the University of Amsterdam, Amsterdam, The Netherlands
2Department of Surgery, Academic Medical Center at the University of Amsterdam, Amsterdam, The Netherlands
3Department of Surgery, Erasmus Medical Center, Rotterdam, The Netherlands
4Department of Surgery, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands
Abstract
Objective: Discussing prognosis is often confronting and complex for cancer patients. This
study investigates how patients’ psychological characteristics relate to their preferences
concerning the disclosure of prognosis.
Methods: One hundred and seventy-six esophageal cancer patients participated in the study.
They had undergone esophagectomy within the past 28 months and did not have evidence of
cancer recurrence. Patients completed a questionnaire eliciting their preferences for prognostic
information. Sociodemographic characteristics, involvement preferences, anxiety, depression,
fear of recurrence, striving for quality of life (QOL) or quantity of life and trust in physicians
were explored as predictors for (a) wanting to be informed about prognosis and (b) the
initiation of discussion about prognosis.
Results: Patients wanting all prognostic information had more fear for the disease to recur
(po0.05) and were inclined to be more actively involved during consultation (po0.001). Post
hoc analyses showed that patients with worse QOL scores reported more fear of recurrence.
Anxiety, depression, trust and tendency to strive for QOL or quantity of life were not related to
preferences concerning prognostic information.
Conclusions: The more fear patients have for esophageal cancer to recur, the more
information they want about prognosis. Thus, patient’s fear for recurrent disease is not a
reason for withholding prognostic information. Results also suggest that there is no harm in
asking patients what information they want.
Copyright r 2009 John Wiley & Sons, Ltd.
Keywords: esophageal cancer; oncology; prognosis; information preferences; patient–provider
interaction
Introduction
Prognosis is one of the most delicate subjects to
discuss in the oncological consultation. There is an
ongoing debate whether it is always beneficial for
patients to be confronted with a bad prognosis
[1–3]. Patients may benefit from a realistic sense of
the future. It enables them to make treatment-
related decisions, prepares them for what is about
to happen, provides them with a time frame and
may alleviate uncertainty [4–7]. However, prog-
nostic information could deprive patients of hope,
which could negatively affect their psychological
well-being [2,8–11]. A recent study showed hope to
be associated with a better quality of life (QOL),
higher self-esteem, better coping with disease and
less anxiety and depression [12]. Although a
minority, a substantial number of patients does
not want to be fully informed or is ambivalent
[11,13–15]. The challenge for physicians is to
recognize whether patients are ready to receive
prognostic information and to what extent. For-
cing information on patients who would rather not
know might be as harmful as withholding it from
those who want it [16].
Some studies have specifically explored cancer
patients’ preferences for prognostic information,
but evidence is limited to the metastatic setting
[3,8,14]. This study focuses on prognostic informa-
tion preferences of patients who have been treated
curatively but who nevertheless have a bad prog-
nosis. It studies esophageal cancer patients who
underwent esophagectomy. Surgery is potentially
curative but has substantial morbidity and a
negative impact on QOL [16–18]. Besides, over
50% of patients develop recurrent disease within
*Correspondence to:
Department of Medical
Psychology, Academic
Medical Center,
Meibergdreef 9, 1105 AZ
Amsterdam, The
Netherlands. E-mail:
j.c.dehaes@amc.uva.nl
Received: 4 April 2008
Revised: 8 August 2008
Accepted: 22 September 2008
Copyright r 2009 John Wiley & Sons, Ltd.

Page 2

5 years [19–21] and 5-year survival rates rarely
exceed 25% [16].
Our earlier results indicated that most esophageal
cancer patients desired prognostic information, but
that this desire declined when information became
more specific and referred to the near future [22]. For
example, 93% of patients wanted to know their
average prognosis, whereas only 71% of patients
wanted information about the chance of recurrence
within 1/2 year. Interestingly, wanting to know
prognosis was not associated with clinical factors,
such as stage of disease and QOL, but rather with
socio-demographic factors, such as a higher educa-
tion level. This raised the question whether other
predictors are relevant, especially psychological ones.
The aim of the present paper is, therefore, to
investigate how psychological factors are related to
esophageal cancer patients’ preferences for the
communication of prognostic information.
We considered five psychological factors to be
relevant. First, patients’ anxiety and depression
were studied. Among patients with diverse types of
cancer it was shown that higher levels of anxiety
were related to a lower information need [2,23].
However, in other studies the opposite was found:
more anxiety or depression meant wanting more
detailed prognostic information [1,14]. We have
assumed that in order to prevent getting more
distress, patients would be likely to avoid probable
negative information. Therefore, we investigated
whether esophageal cancer patients with higher
anxiety and depression levels want less-detailed
prognostic information and do not want to initiate
discussion about it.
Secondly, fear of recurrence is a construct similar
to anxiety, but bound to a specific situation. By
nature a person may be anxious in general but fear
of recurrence can only occur in someone when
diagnosed with a disease. Fear of recurrence is
particularly relevant in our population as recurrent
disease is quite common. We therefore also
investigated whether patients with more fear of
recurrence want less-detailed prognostic informa-
tion and would not want to initiate discussion
about it.
Thirdly, trust in physicians (i.e. surgeons) may
be related to patients’ information preferences. The
relation between trust and information seeking
seems complex. Qualitative studies showed that
trust can either enhance information sharing
between patients and physicians, or impede it
[24,25]. Some trusting patients reported telling
more to, and asking more from their physicians,
because they feel at ease and secure. Other trusting
patients said they asked less information, because
they would rather leave things with their physi-
cians. We therefore studied whether patients high
on trust would want less or more information and
leave initiation of the discussion about prognosis
over to their physician or not.
Fourth, patients’ preference for striving for
quantity versus QOL was considered relevant. To
prolong survival, some patients are willing to
go to extreme lengths and accept treatments that
severely impair their QOL thus avoiding being
confronted with their limited prognosis [26]. Other
patients rather strive for QOL. How this affects
prognostic information preferences has not yet
been investigated, but one might expect that
patients striving for QOL are less interested in
prognostic information than those striving for
quantity of life and thus would not initiate the
discussion themselves.
Finally, patients’ desire to be actively involved
during a medical consultation, for example by
asking questions, was investigated. Patients with a
tendency to show more active behavior were found
to want more information earlier [14]. Therefore,
they were expected to have a preference for
receiving much prognostic information.
Methods
Patients and study design
The present study was performed in three medical
centers in the Netherlands (Academic Medical
Center in Amsterdam, Erasmus Medical Center
in Rotterdam and University Medical Center in
Groningen), as part of a larger study. These
hospitals perform over 50% of all esophagectomies
in the country [27]. The study was approved by the
medical ethics committees of the three hospitals. In
the present consecutive series, patients (18 years or
older) underwent potentially curative esophagect-
omy for cancer (i.e. no distant metastasis) within
the previous 24 months and were discharged from
the hospital. Questionnaires were sent to the
patients’ home. When a patient did not return the
questionnaire, he/she was contacted by telephone
once. Patients who had developed recurrent disease
during follow-up were excluded.
Questionnaire
Information preference
The questionnaire was used, first, to elicit patient
preferences concerning the extent to which they
want to know about prognosis and the way
physicians should communicate this [22]. This
instrument was originally developed to assess such
preferences in Australian patients with metastatic
cancer [14], but most items are applicable to
non-metastatic patients. Only items regarding
end-of-life issues were omitted as these were not
relevant to our patient sample: patients did not
have evidence of metastatic disease or loco-
regional recurrence. The instrument was translated
Copyright r 2009 John Wiley & Sons, Ltd.Psycho-Oncology 18: 1199–1207 (2009)
DOI: 10.1002/pon
1200S. J. Franssen et al.

Page 3

into Dutch using standard forward–backward
translation.
Desired prognostic information: Participants in-
dicated their preferences for receiving information
on a list of in total 15 items by ticking ‘yes’, ‘no
preference’ or ‘no’. This list included six items
specifically addressing prognosis (general prog-
nosis, average survival, individual chance for
survival,5-,1-,and
Figure 1)). From these six items a preference for
prognostic information scale was constructed
(Cronbach’s a50.89).
Preference for initiation for discussing prognostic
information: Patients indicated for each of the six
prognostic information items whether they would
like (1) ‘the specialist to just tell me’; (2) ‘the
specialist to check first if I want to know’; (3) ‘the
specialist to tell me only if I ask’; (4) ‘to find out on
my own, e.g. internet’ or (5) ‘not to discuss at all’.
Cronbach’s a for this scale was 0.95.
1/2-year survival;see
Psychological predictor variables
The
instruments to measure psychological factors.
Levels of anxiety and depression were measured
by the 14-item Hospital and Anxiety and Depres-
sion Scale [28]. Each item was rated on a 4-point
scale, ranging from ‘not at all’ (score50) to
‘mostly’ (score53), resulting in a sum score
ranging from 0 to 21 for both the seven-item
subscale anxiety and the seven-itemsubscale de-
pression. A high score represents a higher level of
anxiety and/or depression. In the present sample,
Cronbach‘s a was 0.83 for the anxiety scale and
0.87 for the depression scale.
Fear of recurrence of the disease was measured
by three items, adapted from De Haes et al.
[29–31]: (a) How often do you think about the
fact your disease might recur? (answers 15‘very
often’ to 45‘never’); (b) After having had treat-
ment, to what extent do you feel you are totally
cured? (answers 15‘very strongly’ to 45‘not at
all’); (c) Since your operation, to what extent have
you been afraid to die? (answer 15‘very much’ to
45‘not at all’). Items were recoded and a sum
score was computed (score 3–12, with a higher
score representing more fear). In the present study
the scale had a Cronbach’s a of 0.66.
Patients’ trust in their physician was measured
by the short version of the Wake Forest Physician
Scale (five items) [32–34]. First,
indicated how much faith they had in their
physician in general. Secondly, how much faith
when it comes to (a) putting their patients’ interests
first; (b) being precise and careful; (c) decision
making about medical treatment and (d) informing
their patients honestly. Answers ranged from (1)
‘totally disagree’ to (5) ‘totally agree’. Sum scores
were transformed to a 0–100 scale with higher
questionnairealso contained established
participants
scores
a50.80).
Attitudes toward striving for length of life
(quantity) (four items; a50.73) and striving for
QOL (four items; a50.86), were assessed by the
Quality–Quantity Questionnaire (QQQ) [26,35,36].
Patients were asked to what extent they agreed with
statements on 5-point Likert scales, ranging from
(1) ‘totally disagree’ to (5) ‘totally agree’. For
example, one of the quantity items is: ‘If a
treatment could prolong my life, I would always
accept it, whatever the side effects might be.’ One
of the quality items is: ‘If a life-prolonging
treatment would prevent me from leading a normal
life, then I would rather not have it.’ Scale scores
range from 1 to 5 with high scores indicating giving
high weights to striving for length or QOL,
respectively.
A patient’s desire to be actively involved during a
medical consultation, for example by being inclined to
ask questions and participate in decision making, was
measured by the seven-item ‘Information’ subscale of
the Krantz Health Opinion Survey [37]. Each item
was rated on a 5-point Likert scale, ranging from (1)
‘totally disagree’ to (5) ‘totally agree’. Higher scores
(range 1–5) represent a patients’ desire toward more
active behavior. The reliability was acceptable in the
present study (a50.69).
indicating highertrust(Cronbach’s
Demographic and clinical predictor variables
Patients reported their demographic characteristics
in the questionnaire (age, educational level, lan-
guage spoken at home, parents’ country of birth
and marital status). Clinical characteristics were
first collected in prospective hospital databases (the
use of (neo-) adjuvant chemo-radiation therapy,
hospital stay, Intensive Care Unit stay, tumor stage
(depth of tumor penetration and presence of lymph
node metastasis [38]) and grade of complications
[39]). Secondly, (QoL) was assessed by the QLQ-
C30 cancer-specific questionnaire (version 3.0) that
includes five functional scales (physical, role,
emotional, cognitive and social) and one global
QoL scale, three general symptom scales (fatigue,
nausea and vomiting and pain) and six single-item
general symptoms (dyspnea, insomnia, appetite,
constipation, diarrhea, financial difficulties) [40].
Also, the esophageal cancer-specific module QLQ-
OES18 was used, which comprises four subscales
(eating, reflux, esophageal pain, dysphagia) and six
single items (cough, dry mouth, taste, choking,
speech and enjoyment) [41]. These data were
analyzed and described before [22].
Statistical analyses
Information preferences were analyzed using des-
criptive statistics. Missing Items from multi-item
scales were imputed if at least half of the items
Preferences for communicating prognosis1201
Copyright r 2009 John Wiley & Sons, Ltd. Psycho-Oncology 18: 1199–1207 (2009)
DOI: 10.1002/pon

Page 4

from the scale were answered [42]. Missing values
from single items were not imputed.
Psychological predictors of patients’ information
preferenceswere explored
Whitney-U, T-test, and logistic regression analyses.
Predictive factors that had a p-value less than 0.1 in
univariate analysis were entered in multivariate
binary logistic regressions, as were the demo-
graphic and clinical predictors that had proven to
be predictive and were described previously [22].
Results of multivariate analyses are given as odd
ratios with 95% confidence intervals. All data
analyses were performed by using SPSS
12.0 (Statistical Package for the Social Sciences,
Chicago, IL, USA).
using
w2, Mann–
sversion
Results
Patients
Of 204 consecutive patients invited to participate,
176 patients (86%) returned the questionnaire.
Nineteen (10%) patients did not return the ques-
tionnaire and nine patients (4%) refused participa-
tion. Non-responders did not differ significantly in
age, gender, neoadjuvant chemo-radiation therapy,
hospital, type of tumor, tumor stage, time since
discharge and the presence of complications.
Sample characteristics of participating patients
are shown in Table 1. The respondents’ median
age was 63 years (range 30–80) and 82% was male.
The mean time since esophagectomy was 11
months (range 1–28). Scores on the psychological
scales are shown in Table 2.
Desired prognostic information
A large majority of patients wanted information
about their average prognosis (93%) and more
specific information about
(71–85%) (see Figure 1). The number of patients
that wanted prognostic information declined when
information became more specific and referring to
the near future.
survivalduration
Predicting information preferences
Because of the skewed distribution, patients were
categorized into those who wanted maximum
information (maximum score: 100) and those
who wanted less than maximum information
(scoreo100).
Several predictors for wanting prognostic infor-
mation were found (see Table 3). In univariate
analysis, patients with more fear of recurrence and
a greater wish to be actively involved during the
consultation were more likely to want maximum
prognostic information.
After controlling for demographic and clinical
variables, a desire for active involvement remained
the only significant psychological predictor of
prognostic information preference. No relation
was found between the patients’ information
preference and anxiety, depression, trust and
striving for QOL or quantity of life. At the same
time, married and higher educated patients as well
as those whose time since discharge was shorter,
were more likely to want maximum prognostic
information.
Table 1. Demographics and clinicopathological parameters of
participating patients (N5176)
DemographicsN (%)
Age [years] (median, range)
Gender
Male
Female
Neoadjuvant therapy
No
Yes
Time since discharge in months (median, range)
Hospital
AMC (Amsterdam)
Erasmus MC (Rotterdam)
UMCG (Groningen)
Education level
o12 years
412 years
Partner
Yes
No
Relationship status
Married
Other
Ethnicity
Dutch
Other
Language spoken at home
Dutch
Other
63 (30–85)
145 (82%)
31 (18%)
128 (73%)
48 (27%)
11.30 (1–28)
93 (53%)
57 (32%)
26 (15%)
90 (51%)
86 (49%)
154 (87%)
22 (13%)
151 (86%)
25 (14%)
173 (98%)
3 (2%)
167 (95%)
9 (5%)
Clinical parametersN (%)
Tumor type
Adenocarcinoma
Squamous cell carcinoma
Other
Stage
Stage I
Stage IIa
Stage IIb
Stage III
Stage IVa
Grading of complications
Grade 0
Grade I
Grade II
Grade III
Grade IV
Hospital stay in days (median, range)
Intensive Care Unit stay in half days (median, range)
Chemoradiation
Yes
No
131 (74%)
39 (22%)
6 (3%)
33 (19%)
52 (30%)
16 (9%)
58 (30%)
17 (10%)
53 (30%)
12 (7%)
43 (24%)
34 (19%)
34 (19%)
16 (8–113)
3 (0.5–63)
48 (27%)
128 (73%)
1202 S. J. Franssen et al.
Copyright r 2009 John Wiley & Sons, Ltd. Psycho-Oncology 18: 1199–1207 (2009)
DOI: 10.1002/pon

Page 5

Who should initiate the discussion of prognostic
information?
Most participants (63–81%) preferred their specia-
list to initiate the discussion about prognosis (see
Table 4). However, of these a significant propor-
tion wanted their specialist to first ask whether they
wanted to be informed before doing so (12–22%).
At the same time, 26 patients (15%) did not want
their doctor to initiate the discussion about
prognosis for at least one of the six prognostic
items. They rather wanted the initiative on their
own side or did not want to discuss prognosis at all.
Predicting who should initiate the discussion of
prognostic information
Patients were categorized into two groups: (A) patients
who wanted their physician to initiate (with or without
asking the patient first) and (B) patients who wanted
the initiative on their own side (asking for prognosis
themselves or not wanting to discuss it at all).
Patients who did not want the doctor to initiate
the discussion of prognosis were identified as those
who had less desire to be actively involved during
the consultation (see Table 5). However, in multi-
variate analysis this predictor variable did not
remain significant. In other words, none of our
expectations concerning initiation of discussing
prognosis were verified. Of the clinical factors,
having a worse social functioning and a better taste
remained independently related to patients prefer-
ring to initiate the discussion themselves.
Table 3. Factors related to the amount of prognostic information desired. Patients were categorized into those who wanted
maximum information (those who received 100 points) and those who desired less than maximum information for at least one of the
items (less than 100 points)
Factor
Maximum prognostic
information desired N5115 (66%)
Less than maximum
prognostic information
desired N560 (34%)
P-value
Univariate
P-value
Multivariate
Odds ratio
(95% CI)
Demographic and clinicopathological variables
Age [years]a
Education level
o12 years
Z12 years
Relationship status
Married
Other
Time since discharge [months]b11 (7)
Dyspnea [EORTC-QoL]b,c
Psychological variables
Anxietyb
Depressionb
Fear of recurrenceb
Trustb
Striving for quality of lifeb
Striving for quantity of lifeb
Active involvementb
60 (30–81)66 (37–85)0.070 0.6951.01 (0.97–1.05)
50 (43%)
65 (57%)
39 (65%)
21 (35%)
0.007 0.046 0.46 (0.22–0.99)
103 (90%)
12 (10%)
47 (78%)
13 (22%)
13 (8)
17 (27)
0.0440.0050.23 (0.08–0.65)
0.079
0.074
0.035
0.126
0.95 (0.90–1.00)
1.01 (1.00–1.02) 24 (30)
4.02 (3.82)
3.86 (3.94)
5.71 (1.90)
83.13 (21.39)
3.85 (0.96)
3.27 (0.90)
3.14 (0.78)
4.25 (4.08)
4.48 (4.14)
5.02 (1.69)
85.68 (22.31)
3.66 (1.11)
3.35 (1.05)
2.71 (0.75)
0.941
0.346
0.013
0.198
0.414
0.883
o0.001
0.0711.23 (0.98–1.53)
0.005 2.05 (1.24–3.40)
p-Values smaller than 0.1 using w2, students-T, Mann–Whitney-U, and logistic regression analyses are shown.
aMedian (range).
bMean (SD).
cA higher score represents more symptoms.
00 10 1020 20303040 4050 50 606070 70 808090 90 100100
Average prognosis; 93%Average prognosis; 93%
Factors which influence individual prognosis positively or negatively; 85%Factors which influence individual prognosis positively or negatively; 85%
Chance of recurrence within 5 years; 82% Chance of recurrence within 5 years; 82%
Chance of recurrence within 1 year; 73%Chance of recurrence within 1 year; 73%
Chance of recurrence within 1/2 year; 71%Chance of recurrence within 1/2 year; 71%
Average time of survival; 74%Average time of survival; 74%
Figure 1. Desired information: the percentage of patients who
want information about specific topics (N5176)
Table 2. Psychological parameters of participating patients
(N5176)
Psychological parameters Mean score (SD)
Anxiety and depression (HADS)
Anxiety scale
Depression scale
Fear of recurrence
Trust
Striving for Life (Q-Q Questionnaire)
Quality scale
Quantity scale
Active involvement
4.10 (3.89)
4.08 (4.00)
5.47 (1.85)
84.09 (21.65)
3.79 (1.02)
3.30 (0.95)
3.00 (0.80)
Preferences for communicating prognosis1203
Copyright r 2009 John Wiley & Sons, Ltd. Psycho-Oncology 18: 1199–1207 (2009)
DOI: 10.1002/pon