Short-Term Effects of
Population-Based Screening
for Prostate Cancer on
Health-Related Quality of Life
Marie-Louise Essink-Bot, Harry
J. de Koning, Hubert G. T. Nijs,
Wim J. Kirkels, Paul J. van der
Maas, Fritz H. Schröder*
Background: Population-based screening for prostate cancer is currently being evaluated in randomized clinical
trials in the United States and in Europe. Side effects arising from the process of screening and from the earlier
treatment of screen-detected prostate
cancer may be important factors in the
evaluation. To examine health-related
quality of life (or health status) among
men screened for prostate cancer, we
conducted a longitudinal study of 626
attenders to the Rotterdam (The Netherlands) prostate cancer screening program and of 500 nonparticipants.
Methods: Attenders of the screening
program and nonparticipants completed self-assessment questionnaires
(SF-36 [i.e., Medical Outcomes Study
36-Item Short-Form Health Survey]
and EQ-5D [i.e., EuroQol measure for
health-related quality of life] health
surveys) to measure generic health status, as well as an additional questionnaire for anxiety and items relating to
prostate cancer screening. Results:
Physical discomfort during digital rectal examination and during transrectal
ultrasound was reported by 181 (37%)
of 491 men and by 139 (29%) of 487
men, respectively; discomfort during
prostate biopsy was reported by 64
(55%) of 116 men. Mean scores for
health status and anxiety indicated that
the participants did not experience relevant changes in physical, psychological, and social functioning during the
screening procedure. However, high
levels of anxiety were observed
throughout the screening process
among men with a high predisposition
to anxiety. Similar scores for anxiety
predisposition were observed among
attenders and nonparticipants. Conclu-
sions: At the group level, we did not
find evidence that prostate cancer
screening induced important shortterm health-status effects, despite the
short-lasting side effects related to the
biopsy procedure. However, subgroups
may experience high levels of anxiety.
The implication is that unfavorable
health-status effects of prostate cancer
screening occur mainly in the treatment phase. [J Natl Cancer Inst 1998;
90:925–31]
Prostate cancer screening is currently
being evaluated in two large randomized
trials, i.e., the European Randomized
Study of Screening for Prostate Cancer
and the screening trial for prostate, lung,
colorectal, and ovarian cancers in the
United States (1–4). Provided that a decrease in prostate cancer mortality as a
result of screening will ultimately be
shown, the crucial question is whether
this benefit outweighs the inevitable unfavorable side effects of population-based
cancer screening.
Unfavorable side effects on healthrelated quality of life (or health status)
may occur during the screening procedure
itself (e.g., pain, discomfort, and feelings
of uncertainty and distress). Furthermore,
health-status effects of cancer screening
can be found in the treatment phase. Subjects with screen-detected cancer are labeled as cancer patients and treated earlier
than if there had been no screening. At the
population level, cancer screening causes
an increase in the number of life-years
lived with side effects of primary treatment (5). However, if prostate cancer
mortality decreases as a result of screening, the incidence of end-stage disease decreases accordingly. At the population
level, the decrease in prostate cancer mortality will result in a decrease in the time
lived in the bad health states associated
with end-stage cancer. This result constitutes a favorable effect on health status of
population-based screening (6). All of the
health-status effects mentioned above
contribute to the ultimate balance of advantageous and disadvantageous public
health effects of a cancer screening program (7).
This report describes an empirical
study on the short-term health-status effects of the prostate cancer screening process itself. Three research questions will
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
be answered: 1) To what extent is the generic health status of participants in the
short term affected by either the initial
screening tests or the biopsy procedure?
2) Are there any specific effects relating
to emotional distress? 3) Is there any support for psychological self-selection
among nonparticipants?
Subjects and Methods
The Rotterdam Screening Trial for
Prostate Cancer
Six European centers have participated in the European Randomized Study of Screening for Prostate
Cancer. The trial will provide an empirical answer to
the question of whether screening in men in the 55to 69-year-old age group reduces prostate cancer
mortality. The ongoing Rotterdam trial was begun in
June 1994 and was preceded by a number of pilot
studies (8). The screening protocols in the centers
participating in the European Randomized Study of
Screening for Prostate Cancer are slightly different.
In Rotterdam (P6 protocol), men in the 55- to 74year-old age group were identified in municipal
population registries (with 100% coverage) and invited by mail to participate in the trial. Their letter of
invitation included general information regarding
prostate cancer screening and the trial as well as an
informed consent form. After providing written informed consent to the screening bureau, the men
were randomly and individually assigned by computer to either the screening arm or the control arm
of the trial. The men in the control arm were informed about their control status. Since the start of
the P6 protocol, 8639 men (43% of eligible men)
have participated in the trial.
Men who were randomly assigned to the screening arm were invited for screening. At the screening
center, they provided a venous blood sample for
prostate-specific antigen (PSA) testing. Subsequently, digital rectal examination and transrectal
ultrasound were performed (both by the same person
per attender). The PSA result was not available
when the other tests were performed. An additional
appointment for prostate biopsy was scheduled if the
PSA level equaled 4 ng/mL or more or if either one
*Affiliations of authors: M. L. Essink-Bot, H. J.
de Koning, P. J. van der Maas, Department of Public
Health, Erasmus University Rotterdam, The Netherlands; H. G. T. Nijs, Department of Health Policy
and Management, Erasmus University Rotterdam,
and Department of Health Promotion, Municipal
Health Services, Rotterdam Area; W. J. Kirkels, F.
H. Schröder, Department of Urology, Erasmus University Rotterdam, and Academic Hospital, Rotterdam.
Correspondence to: Marie-Louise Essink-Bot,
M.D., Ph.D., Department of Public Health, Erasmus
University Rotterdam, P.O. Box 1738, 3000 DR
Rotterdam, The Netherlands. E-mail: ESSINK@
MGZ.FGG.EUR.NL
See ‘‘Notes’’ following ‘‘References.’’
© Oxford University Press
REPORTS 925
of the other tests gave a suspicious result. If applicable, attenders were requested to discontinue aspirin or oral anticoagulant therapy 10 days before the
prostate biopsy was performed. All candidates for a
biopsy took preventive antimicrobial therapy 2
hours before the procedure and 4 hours after its
completion. A sextant transrectal prostate biopsy
was performed (and at the lesion, if visible by transrectal ultrasound). The urologist informed each man
about the result approximately 2 weeks after the biopsy was performed.
A comprehensive cost-effectiveness analysis has
been conducted alongside the randomized screening
trial. This analysis included health-status assessments in the relevant phases of the screening and the
treatment path.
The screening trial, the cost-effectiveness analysis, and the health-status study were approved by the
Medical Ethical Committee of the Academic Hospital in Rotterdam, The Netherlands. Approval for
the prostate screening program and the screening
trial was obtained from the Minister of Health and
the Health Council, as required by the Population
Screening Act (The Netherlands, 1992).
Hypotheses and Design of the Study
on the Short-Term Health-Status
Effects of Prostate Cancer Screening
A number of hypotheses were formulated at the
start of the study. First, we did not expect to find, on
average, relevant effects of the screening procedure
itself on the generic health status. Second, an increased level of anxiety (above the age- and sexadjusted reference scores) among attenders to the
screening was expected to be found immediately
preceding the initial screening tests, whereas the average anxiety level was expected to return to normal
after the subject was informed that all was well.
Furthermore, the prostate biopsy was expected to be
a stressful event. Even higher levels of anxiety were
expected to be measured in the subjects who had
undergone a biopsy procedure recently and were
waiting to be informed about the results. The high
anxiety levels of these subjects were expected to
return to normal after they received the message that
they did not have cancer. Third, we assumed that
men with a predisposition to react with high anxiety
levels to a stressful event (such as cancer screening)
would be more likely not to respond to the invitation
to the screening trial. Such behavior could then tentatively be explained as a method employed to reduce the stress of the receipt of the invitation and
prevent the anticipated levels of distress induced by
the screening itself. If it could be shown that nonparticipants differ from attenders with respect to the
predisposition to anxiety, this could lend support to
the hypothesis of psychological self-selection
among nonparticipants in a prostate cancer screening program.
Two groups of men were approached for empirical health-status measurement: attenders and nonparticipants. The first group (attenders) had provided written informed consent to participate in the
screening trial and had been randomly assigned into
the screening arm. Health-status data were longitudinally collected (Fig. 1). A sample of 626 men who
were consecutively invited for screening during the
period from September 1995 through March 1996
received the first questionnaire by mail, approxi-
926 REPORTS
mately 3 weeks before the scheduled visit to the
screening center (time 1 or T1). The second assessment (time 2 or T2) took place in the waiting room
at the screening center directly preceding the screening (venous blood sampling for PSA testing, digital
rectal examination, and transrectal ultrasound). Men
for whom the results of the initial screening tests
were unsuspicious and for whom a prostate biopsy
was deemed to be unnecessary were asked to complete a health-status questionnaire at time 3 (T3), 1
week after they were informed about this result.
Men with a PSA level of 4 ng/mL or more and/or
suspicious results of either of the other tests were
offered a biopsy procedure, and they were asked to
complete a questionnaire at time 4 (T4), i.e., during
the 2-week period between the biopsy and the visit
to the outpatient clinic in which the man was informed about the biopsy result. The men who received the message that prostate cancer was not confirmed by the biopsy result (i.e., false-positive initial
screening result) completed another questionnaire at
time 5 (T5), i.e., 1 week after they received the
message that they did not have cancer. Patients with
screen-detected cancer did not receive the T5 questionnaire of the present study because the healthstatus impact of prostate cancer and its various treatment modalities are the subjects of a separate ongoing longitudinal study. Reminder cards were sent
to nonrespondents at T3 and T5, respectively, 2
weeks after distribution of the initial questionnaire.
Reminder procedures were not feasible because of
time constraints at the other assessments.
In summary, the attenders completed questionnaires at T1 (baseline) and T2 (waiting room) and
either at T3 (when nothing suspicious was found in
the initial screening tests) or at T4 (during the waiting period for the biopsy result) and T5 (after being
informed that the biopsy result did not confirm a
prostate cancer diagnosis).
The second group of men approached for empirical health-status assessment (nonparticipants) consisted of men who did not respond to the initial
invitation to participate in the trial. A random selection of 500 nonparticipants who had been invited
during the period from September 1995 through
March 1996 received a questionnaire by mail in July
1996. Reminders were sent to all nonrespondents at
2 and 4 weeks after the initial mailing.
Health-Status Measures
Health-related quality of life (or health status) was
defined as the functioning of subjects in the physical, psychological, and social domains. Two generic,
i.e., comprehensive and non-disease-specific,
health-status measures (SF-36 [i.e., Medical Outcomes Study 36-Item Short-Form Health Survey]
and EQ-5D [i.e., EuroQol measure for health-related
quality of life]) were supplemented by a specific
measure for anxiety (the State–Trait Anxiety Inventory [STAI]) and self-developed items on specific
consequences and the experiences of the various
screening tests. More complete descriptions of these
measures are given below.
The SF-36 was developed in the United States in
the late 1980’s as part of the Medical Outcomes
Study (9–11), a longitudinal investigation of the
self-reported health status of patients with a range of
chronic conditions. It consists of 36 items, organized
into eight scales (physical functioning [10 items],
role limitations due to physical problems [role-
physical; four items], bodily pain [two items], general health perceptions [five items], vitality [four
items], social functioning [two items], role limitations due to emotional problems [role-emotional;
three items], and mental health [five items]). One
item on health change that is scored separately is not
reported on in this paper. The number of response
choices per item ranges from two to six. Each scale
yields a score in the range from 0 (the worst possible
score) to 100 (the best possible score) (9). Differences of 6.5 points in the physical dimensions and
7.9 points in the mental dimensions of the SF-36 are
thought by its authors to be clinically meaningful
(10). Furthermore, two summary measures can be
computed that aggregate scales: the physical component summary and the mental component summary. These summary scales were constructed so
that they have a mean score of 50 (25th–75th percentile score interval, 43–56) for the general population in the United States (10,11). The Dutch version of the SF-36 employed in the current study was
developed as a part of the International Quality of
Life Assessment Project, whose objective was to
translate, adapt, and validate the SF-36 for use in
some 15 countries (12). We used the version that
utilizes a time frame (reference period) of 1 week.
The EQ-5D measure for health-related quality of
life was developed by the EuroQol Group as a generic five-item classification (13). The EuroQol
Group is a multidisciplinary and originally European
group of researchers, with current members from
centers in the U.K, The Netherlands, Spain, Germany, Denmark, Finland, Sweden, Greece, the
United States, and Canada. The standard Dutch EQ5D was used (14). It consists of five items regarding
mobility, self-care, usual activities, pain/discomfort,
and anxiety/depression, with each item having three
levels: 1 4 no problems, 2 4 some problems, and
3 4 extreme problems. For reasons of brevity, these
items are not described in this report. Furthermore,
respondents are asked to rate their own health directly on a visual analogue scale ranging from 0
(labeled as ‘‘worst imaginable health state’’) to 100
(labeled as ‘‘best imaginable health state’’).
The STAI is a U.S. 40-item questionnaire, of
which a validated Dutch version is available (15–
17). It consists of two separate parts of 20 items
each. The first, the state part, measures temporary
anxiety as occurring in reaction to specific situations. The second, the trait part, intends to measure
a stable personality characteristic, i.e., a person’s
predisposition to react with high anxiety levels to
anxiety-provoking situations. A subject with a high
trait anxiety score will generally be more anxious in
a specified situation (e.g., prostate cancer screening)
than a subject with a low trait anxiety score in the
same situation. State anxiety scores range from 20
(lowest possible anxiety score) to 80 (highest possible anxiety score). Similarly, the score range of the
trait anxiety is limited by 20 (lowest possible predisposition to anxiety ) to 80 (highest possible predisposition to anxiety).
Screening-specific items. Subjects were retrospectively asked to grade pain and physical discomfort, respectively, experienced during the venous
blood sampling for PSA testing, digital rectal examination, transrectal ultrasound, and, if relevant,
prostate biopsy, respectively, on 3-point Likert
scales at T3 or T5. Biopsy-specific items relating to
symptoms experienced in the week after the biopsy
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
Fig. 1. Flow chart describing the cohort of men invited to participate in the
Rotterdam part of the European Randomized Study of Screening for Prostate
Cancer during the period from September 1995 through March 1996. Vertical
arrows show the timing of the health-status assessments. Health-status measures
used and numbers of usable response are indicated. SF-36 4 Medical Outcomes
Study 36-Item Short-Form Health Survey; EQ-5D 4 EuroQol measure for
health-related quality of life; STAI 4 State–Trait Anxiety Inventory; PSA 4
and their consequences were included in the questionnaire at T4. Limitations as a result of the biopsy
in daily activities, social activities, and sex life in the
week after the biopsy were graded on 5-point Likerttype scales with labeled end points, ‘‘no limitations
at all’’ and ‘‘extremely limited.’’
Demography and morbidity. Information on
age, marital status, educational level, and employment status was obtained. Educational level was
classified as low (primary school or lower technical
education), high (college/university degree), or intermediate. Morbidity was assessed by the standard
list of chronic conditions from the Dutch Bureau for
Statistics (‘‘CBS’’; located in Voorburg, The Netherlands) that enumerates 28 conditions in lay terms
(e.g., ‘‘asthma, chronic bronchitis, or chronic obstructive pulmonary disease’’ and ‘‘diabetes’’). Respondents were asked whether they had any of these
conditions, either currently or during the previous
year.
prostate-specific antigen. Timing of the health-status assessments: T1 4 baseline, approximately 3 weeks before screening; T2 4 in the waiting room before
screening; T3 4 1 week after being notified about the unsuspicious results of the
initial screening tests, no biopsy needed; T4 4 during the 2-week period between the biopsy procedure and the result while waiting for biopsy result; T5 4
1 week after being notified that prostate cancer was not confirmed by the biopsy
result.
Presentation of the Questionnaires
The questionnaires presented to the attenders at
various time points and to the nonparticipants are
shown in Fig. 1.
Statistical Analysis
Missing values for the items of the SF-36 were
imputed following the guidelines in the SF-36
Health Survey Manual (9). Missing values for the
items of the other instruments (EQ-5D, STAI, and
screening-specific items) were not imputed.
Mean values are presented as the measure of central tendency in the scores observed for the SF-36
scales, state anxiety, trait anxiety, EQ-5D self-rated
health, and demographic characteristics with interval properties (age and number of chronic conditions). The dispersion observed in these scores is
presented by the interval bounded by the 25th and
the 75th percentile scores.
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
The two-sided t test for dependent samples was
used to test differences between T3 (after unsuspicious results of initial prostate cancer screening
tests) and T1 (baseline) mean SF-36 scale scores and
EQ-5D self-rated health scores of attenders who underwent only the initial screening tests. Similarly,
the two-sided t test for dependent samples was used
to test differences between mean SF-36 scale scores
and mean EQ-5D self-rated health scores at T5 (after
notice that prostate cancer was not confirmed by
biopsy) and at baseline (T1) of attenders who underwent a screening including a biopsy procedure.
Differences in mean scores for the state anxiety, trait
anxiety, and EQ-5D self-rated health assessments at
various time points were tested by the same procedure.
Mean difference scores and 95% confidence intervals of the differences were computed for SF-36
scale scores and EQ-5D self-rated health scores of
the group who underwent screening without biopsy
REPORTS 927
(T3 – T1) and with biopsy (T5 – T1), respectively.
Similarly, mean difference scores and 95% confidence intervals of the differences were computed for
state anxiety and trait anxiety scores (T3 – T2, procedure without biopsy; T4 – T2 and T5 – T4, procedure with biopsy).
Differences between the attenders and the nonparticipants were assessed with descriptive statistics,
and no statistical tests were performed.
All P values resulted from the use of two-sided
statistical tests. Analyses were conducted with SPSS
6.1 for Windows (SPSS Inc., Chicago, IL) or GLIM
(Royal Statistical Society, London, U.K.).
Changes in Attenders’ Health-Related
Quality of Life
tionnaire. The attenders had a mean age of
63.2 years (25th–75th percentile score interval, 59.0–66.8 years); 88% were married; 63% were pensioners; 40% had a
low level of education, 41% had a medium level, and 19% had a high level; and
they reported 1.6 conditions on average
(25th–75th percentile score interval, zero
to two conditions).
Mean score results from the SF-36,
STAI, and self-rated health EQ-5D questionnaires completed by the attenders to
screening and by the nonparticipants are
shown in Table 1. Tables 2 and 3 show
longitudinal comparisons and statistical
testing results.
The numbers of usable questionnaires
returned from men eligible for participation in the study were as follows: at T1,
600 (96%) of 626 questionnaires; at T2,
555 (96%) of 581 questionnaires; at T3,
381 (86%) of 442 questionnaires; at T4,
160 (98%) of 164 questionnaires; and at
T5, 116 (89%) of 130 questionnaires. At
T2, T3, and T4, for unintentional, organizational reasons, 45, 15, and five eligible
men, respectively, did not receive a ques-
• SF-36 scores (screening without biopsy): The comparison of mean SF-36
scores at T1 and T3 to investigate the
effects of the initial screening tests indicated significantly better health status
at T3 for two of eight scales and for the
physical component summary. The actual size of the mean differences was
less than 2 points.
• SF-36 scores (screening procedure with
biopsy): The comparison of mean SF-
Results
36 scores at T1 and T5 to investigate
the effects of false-positive result of the
initial screening followed by biopsy
showed significant improvement for
two of eight scales and the physical
component summary. The mean differences are somewhat larger. The 95%
confidence intervals are wider here because of smaller numbers.
• Comparison of the scores for self-rated
health EQ-5D at T1, T2, and T3
(screening procedure without biopsy)
and at T1, T2, T4, and T5 (screening
procedure with biopsy) did not reveal
any statistically significant differences.
• Anxiety (screening without biopsy):
The mean state anxiety score at T3 was,
on average, 2.56 points lower than at
T2. However, both mean scores were at
or below the age- and sex-adjusted reference score for the Dutch general
population (mean score, 34; standard
deviation, ±11) (15). A similar result
was found for the STAI trait scores.
• Anxiety (screening procedure with biopsy): Differences in state anxiety
scores observed at T4 and T5 were statistically significant, with a mean dif-
Table 1. SF-36, STAI, and self-rated health EQ-5D scores (mean values [25th–75th percentile score intervals]) of attenders and nonparticipants to the
Rotterdam prostate cancer screening program*
Attenders†
T1
(n 4 600): mean
(25th–75th
percentile)
T2
(n 4 555): mean
(25th–75th
percentile)
T3
(n 4 381): mean
(25th–75th
percentile)
T4
(n 4 160): mean
(25th–75th
percentile)
T5
(n 4 116): mean
(25th–75th
percentile)
Nonparticipants
(n 4 235): mean
(25th–75th
percentile)
SF-36 (score 100–0)
Physical functioning
Role-physical‡
Bodily pain
General health perceptions
Vitality
Social functioning
Role-emotional§
Mental health
84 (78–100)
85 (100–100)
81 (62–100)
70 (59–82)
75 (65–90)
88 (88–100)
90 (100–100)
80 (72–92)
—
—
—
—
—
—
—
—
85 (80–100)
87 (100–100)
83 (72–100)
70 (59–82)
76 (56–90)
89 (88–100)
89 (100–100)
82 (76–92)
—
—
—
—
—
—
—
—
86 (80–100)
90 (100–100)
86 (74–100)
74 (62–87)
75 (65–90)
90 (88–100)
91 (100–100)
180 (72–92)
79 (65–95)
79 (75–100)
85 (74–100)
65 (52–77)
74 (60–85)
87 (78–100)
85 (100–100)
80 (72–92)
SF-36 summary scores
Physical component summary
Mental component summary
51 (48–56)
56 (53–60)
—
—
52 (49–56)
56 (53–60)
—
—
53 (50–57)
55 (53–60)
49 (44–55)
55 (52–61)
—
—
34 (27–40)
32 (24–37)
32 (24–37)
31 (23–36)
35 (26–43)
32 (23–38)
32 (24–36)
32 (24–36)
—
34 (28–39)
82 (75–90)
82 (75–90)
82 (75–93)
84 (75–95)
84 (75–95)
80 (70–95)
STAI (score 20–80)
State anxiety
Trait anxiety
EQ-5D (score 100–0)
Self-rated health today
*SF-36 4 Medical Outcomes Study 36-Item Short-Form Health Survey; STAI 4 State–Trait Anxiety Inventory; EQ-5D 4 EuroQol measure for health-related
quality of life.
†Timing of the health-status assessments: T1, baseline, approximately 3 weeks before screening; T2, in the waiting room before screening; T3, 1 week after being
notified about the unsuspicious results of the initial screening tests, no biopsy needed; T4, during the 2-week period between the biopsy procedure and the result
while waiting for biopsy result; T5, 1 week after being notified that prostate cancer was not confirmed by the biopsy result. Each attender completed the SF-36 twice,
i.e., at T1 and T3 or at T1 and T5, depending on whether a biopsy was performed; the STAI at T2 and at T3 or at T4 and T5; and the EQ-5D at all relevant
assessments.
‡Role-physical 4 role limitations due to physical problems.
§Role-emotional 4 role limitations due to emotional problems.
928 REPORTS
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
Table 2. Longitudinal comparisons of SF-36 scores and self-rated health EQ-5D scores of attenders to the Rotterdam prostate cancer screening program: mean
difference score, 95% confidence interval, and P value*
Without biopsy: T3 − T1† (n 4 381)
With biopsy: T5 − T1† (n 4 116)
Mean difference score
95% confidence interval
P‡
Mean difference score
95% confidence interval
P‡
0.21
1.67
1.68
0.25
0.78
0.83
−1.06
1.72
−0.78 to +1.19
−0.32 to +3.67
+0.29 to +3.07
−0.95 to +1.45
−0.32 to +1.91
−0.61 to +2.28
−2.72 to +0.60
+0.69 to +2.75
.6818
.1009
.0187
.6839
.1638
.2585
.2119
.0012
0.51
2.54
4.20
2.78
−0.80
1.01
−0.34
−0.37
−2.10 to +1.07
−1.77 to +6.86
+1.01 to +7.38
+0.56 to +5.00
−3.44 to +1.83
−2.03 to +4.06
−3.78 to +3.11
−3.06 to +2.31
.5277
.2504
.0111
.0155
.5512
.5151
.8486
.7859
SF-36 summary scores
Physical component summary
Mental component summary
0.51
0.35
+0.11 to +0.91
−0.20 to +0.90
.0127
.2122
0.87
−0.10
+0.13 to +1.62
−1.45 to +1.25
.0237
.8828
EQ-5D
Self-rated health
0.17
−0.65 to +0.99
.6825
−0.05
−1.86 to +1.76
.9538
SF-36
Physical functioning
Role-physical§
Bodily pain
General health perceptions
Vitality
Social functioning
Role-emotional\
Mental health
*SF-36 4 Medical Outcomes Study 36-Item Short-Form Health Survey; EQ-5D 4 EuroQol measure for health-related quality of life.
†Timing of the health-status assessments: T1, baseline, approximately 3 weeks before screening; T3, 1 week after being notified about the unsuspicious results
of the initial screening tests, no biopsy needed; T5, 1 week after being notified that prostate cancer was not confirmed by the biopsy result.
‡Two-sided t test for dependent samples.
§Role-physical 4 role limitations due to physical problems.
\Role-emotional 4 role limitations due to emotional problems.
Table 3. Longitudinal comparisons of STAI scores and self-rated health EQ-5D scores of attenders to the Rotterdam prostate cancer screening program: mean
difference score, 95% confidence interval, and P value*
Without biopsy (n 4 381)
With biopsy (n 4 116)
T3 − T2†
STAI
State anxiety
Trait anxiety
EQ-5D
Self-rated health
Mean
difference
score
95%
confidence
interval
−2.56
−1.08
0.27
T4 − T2†
P‡
Mean
difference
score
95%
confidence
interval
−3.51 to −1.60
−1.80 to −0.35
.0001
.0039
0.54
1.09
−0.60 to +1.14
.5384
−0.76
T5 − T4†
P‡
Mean
difference
score
95%
confidence
interval
P‡
−0.98 to +2.07
−0.07 to +2.25
.4877
.0670
−3.73
−1.24
−5.17 to −2.29
−2.14 to −0.33
.0001
.0086
−2.31 to +0.78
.3351
1.05
−0.31 to +2.42
.1336
*STAI 4 State–Trait Anxiety Inventory; EQ-5D 4 EuroQol measure for health-related quality of life.
†Timing of the health-status assessments: T1, baseline, approximately 3 weeks before screening; T2, in the waiting room before screening; T3, 1 week after being
notified about the unsuspicious results of the initial screening tests, no biopsy needed; T4, during the 2-week period between the biopsy procedure and the result
while waiting for biopsy result; T5, 1 week after being notified that prostate cancer was not confirmed by the biopsy result. Each attender completed the STAI at
T2 and at T3 or at T4 and T5 (depending on whether a biopsy was performed) and the EQ-5D at all relevant assessments.
‡Two-sided t test for dependent samples.
ference score of −3.73 points, indicating relief of anxiety after obtaining the
message that prostate cancer was not
confirmed by biopsy. Only the mean
score at T4 was marginally higher than the
age- and sex-adjusted reference scores.
• A more detailed look at the anxiety
scores by dividing the attenders into
two groups by the 75th percentile of the
trait anxiety score observed at T2
showed a level effect; that effect can be
illustrated by the results of this analysis
in the group of men who underwent
screening including a biopsy. Men with
trait anxiety scores greater than or equal
to 35 had a mean state anxiety score at
T2 of 43 (25th–75th percentile score interval, 39–48), at T4 of 45 (37–53), and
at T5 of 44 (32–53). The men with trait
anxiety scores below 35 had a mean
state anxiety score of 31 (25th–75th
percentile score interval, 26–36) at T2,
of 32 (24–37) at T4, and of 28 (21–32)
at T5. A STAI state anxiety score of
more than 2 standard deviations above
the average age- and sex-adjusted reference score of 34 (standard deviation,
±11), i.e., higher than 56, was observed
in five (1%) of the attenders at T2, 11
(3%) at T3, five (3%) at T4, and four
(3%) at T5.
• Six men (1% of 495 men who returned
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
usable response to this item; pooled
data from T3 and T5) reported that they
had experienced pain during the venous
blood sampling for the PSA test, and
eight (2%) of 494 men reported physical discomfort during the venous blood
sampling for the PSA test. Of 491 men,
181 (37%) reported that they had experienced physical discomfort during the
digital rectal examination. Furthermore,
139 (29%) of 487 men reported physical discomfort during the transrectal ultrasound procedure. The latter two tests
caused pain in 73 (15%) of 491 men
and 49 (10%) of 488 screened individuals. The biopsy procedure was reported
REPORTS 929
to have induced pain in 41 (35%) of 116
men and physical discomfort in 64
(55%) of 116 men.
• Relevant proportions of the men experienced one or more physical symptoms
in the week after prostate biopsy (Table
4). Only small percentages reported that
symptoms interfered with daily, social,
or sexual activity.
Comparison of Attenders and
Nonparticipants
The usable response rate from the nonparticipants was 47% (n 4 235). They
had a mean age of 64.8 years (25th–75th
percentile score interval, 59.5–70.0
years); 80% were married; 65% were pensioners; their attained level of education
was low in 78%, intermediate in 18%, and
high in 4%; and they reported 1.5 conditions on average (25th–75th percentile
score interval, 0–2). Nonparticipants were
less well educated than attenders. Healthstatus scores of the nonparticipants are
shown in Table 1. The SF-36 scores of the
nonparticipants were in the same range as
age- and sex-adjusted reference scores for
the SF-36 from the Dutch general population (Aaronson NK, Muller M, Cohen
PD, Essink-Bot ML, Fekkes M, Sander-
Table 4. Physical symptoms, medical consumption, and functional consequences in the week after
prostate biopsy, as reported by attenders to the Rotterdam prostate cancer screening program (n 4 160)
in the T4 questionnaire
Physical symptoms
Pain in prostate region
No
Yes, but <1 wk
Yes, ù1 wk
Fever >38 °C
No
Yes
Blood in urine
No
<1 wk
ù1 wk
Blood in feces
No
Yes
Blood in sperm
No
Yes, but <1 wk
Yes, ù1 wk
Not applicable/do not know
Medical consumption related to biopsy
Use of painkillers
Yes
No
Visit to general practitioner
Yes
No
Functional consequences
Interference with daily activities
Not at all
A little bit
Moderately
Quite a bit
Extremely
Interference with social activities
Not at all
A little bit
Moderately
Quite a bit
Extremely
Interference with sexual activities
Not at all
A little bit
Moderately
Quite a bit
Extremely
Not applicable
Absolute No.*
Percentage of total
98
56
3
62
36
2
153
7
96
4
60
82
10
40
54
7
118
39
75
25
9
48
48
52
6
31
31
33
6
153
4
96
4
155
3
97
130
22
3
1
3
82
14
2
1
2
140
11
4
1
3
88
7
3
1
2
60
30
5
0
1
62
38
19
3
0
1
39
*Absolute numbers per item may not add up to 160 because of incidental missing values.
930 REPORTS
man R, et al.: manuscript submitted for
publication). The SF-36 scores observed
among the attenders generally indicated a
better health status than among the general population. The mean trait anxiety
scores were slightly different between attenders and nonparticipants. The mean
score of 34 observed among the nonparticipants was not higher than the score
presented in the reference tables for
Dutch men older than 51 years (15).
Discussion
Despite the fact that the initial screening tests for prostate cancer and, in particular, the prostate biopsy caused physical discomfort in large proportions of the
attenders, a relevant impact on generic
health status was not found. Even falsepositive results of the initial screening
tests followed by prostate biopsy caused
little interference at the level of activities
despite the frequent occurrence of physical symptoms such as hematuria. Although the nonresponse rate to the T4
questionnaire was low (2%), it may, however, partly be caused by subjects having
severe complications, such as septicemia,
after biopsy. Rietbergen et al. (18) studied
1687 patients who underwent a prostate
biopsy procedure and reported that seven
(0.4%) had to be admitted to a hospital
after biopsy, with one patient having a
life-threatening sepsis.
The results do not provide strong support for an anxiety-raising effect of the
screening. The mean state anxiety score
observed in the subjects in the waiting
room before the screening was significantly higher than that observed in the
subjects after being informed that the results of the initial tests were unsuspicious.
The mean state anxiety score was highest
in the men who underwent a prostate biopsy procedure but who were not yet
aware of the results from the histologic
examination of their specimens. However, the changes in the anxiety scores
were small in absolute terms, and the
mean scores remained in the range of the
age- and sex-adjusted reference scores for
the Dutch population.
There are at least two possible interpretations for these results. First, prostate
cancer screening examinations may be
perceived by attenders as routine tests.
Separate analyses revealed that men with
a high predisposition to anxiety experi-
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
ence high anxiety levels throughout the
screening process. No information is
available as to whether the observed high
anxiety scores decrease further in the direction of normal after a longer period
after the screening. A second explanation
relates to the method of quantitative selfassessment using the STAI. Maybe this
approach is too general for measuring
anxiety in a screening situation. The results of our study are remarkably similar
to those of Sutton et al. (19), which did
not show anxiety-raising effects of breast
cancer screening. The use of a measure
directed specifically at worries related to
prostate cancer, similar to the approach
advocated by Lerman et al. (20) in breast
cancer screening, could provide additional insight. Other approaches could include qualitative or psychophysiologic research.
Nonparticipants appeared to differ
from the attenders, particularly with respect to educational level. The similarity
of the trait anxiety scores among nonparticipants and attenders challenges the hypothesis of psychological self-selection as
a cause of nonparticipation.
Results from other prostate cancer
screening studies for comparison purposes are not available in the literature.
Published empirical studies on the healthstatus effects of cancer screening programs remain scarce. There is evidence
that the initial screening test (mammography) in breast cancer screening generally
does not have important negative psychological consequences (without ignoring
the fact that mammography can be quite
painful). The recall for additional diagnostic work-up (biopsy) in women whose
biopsy specimens later proved to be falsepositives induced substantial distress that
lasted beyond receiving reassurance that
all was well; however, 6 months later, no
excess distress was demonstrated by these
women (20,21).
The results of our study can be seen as
negative. The other interpretation is that
we did not find empirical support for one
type of objection to population screening
programs in general, i.e., the assumption
that the availability of a screening program provokes emotional distress among
large population groups. If prostate cancer
mortality and the incidence of end-stage
disease are reduced by early detection
programs, unfavorable health-status effects of prostate cancer screening occur
mainly in the treatment phase.
(12)
(13)
(14)
References
(1) Schroder FH, Denis LJ, Kirkels W, De Koning
HJ, Standaert B. European randomized study
of screening for prostate cancer. Progress report of Antwerp and Rotterdam pilot studies.
Cancer 1995;76:129–34.
(2) Auvinen A. Screening for prostate cancer using serum prostate-specific antigen: a randomised, population-based pilot study in Finland. Br J Cancer 1996;74:568–72.
(3) Kramer BS, Gohagan J, Prorok PC, Smart C. A
National Cancer Institute sponsored screening
trial for prostatic, lung, colorectal, and ovarian
cancers. Cancer 1993;71:589–93.
(4) Auvinen A, Rietbergen JB, Denis LJ, Schroder
FH, Prorok PC. Prospective evaluation plan for
randomised trials of prostate cancer screening.
The International Prostate Cancer Screening
Trial Evaluation Group. J Med Screen 1996;3:
97–104.
(5) Stewart-Brown S, Farmer A. Screening could
seriously damage your health [editorial]. BMJ
1997;314:533–4.
(6) de Koning HJ, van Ineveld BM, de Haes JC,
van Oortmarssen GJ, Klijn JG, van der Maas
PJ. Advanced breast cancer and its prevention
by screening. Br J Cancer 1992;65:950–5.
(7) de Haes JC, de Koning HJ, van Oortmarssen
GJ, van Agt HM, de Bruyn AE, van der Maas
PJ. The impact of a breast cancer screening
programme on quality-adjusted life-years. Int J
Cancer 1991;49:538–44.
(8) Schroder FH, Damhuis RA, Kirkels WJ, de
Koning HJ, Kranse R, Nijs HG, et al. European
randomized study of screening for prostate
cancer—the Rotterdam pilot studies. Int J Cancer 1996;65:145–51.
(9) Ware JE, Snow KK, Kosinski M, Gandek B.
SF-36 health survey manual and interpretation
guide. Boston: New England Medical Center,
The Health Institute, 1993.
(10) Ware JE, Kosinski M, Keller SD. SF-36 physical and mental component summary measures—a users’ manual. Boston: New England
Medical Center, The Health Institute, 1994.
(11) Ware JE Jr, Kosinski M, Bayliss MS, McHorney CA, Rogers WH, Raczek A. Comparison
of methods for the scoring and statistical analy-
Journal of the National Cancer Institute, Vol. 90, No. 12, June 17, 1998
(15)
(16)
(17)
(18)
(19)
(20)
(21)
sis of SF-36 Health Profile and Summary measures: summary of results from the Medical
Outcomes Study. Med Care 1995;33(4 Suppl):AS264–79.
Aaronson NK, Acquadro C, Alonso J, Apolone
G, Bucquet D, Bullinger M, et al. International
Quality of Life Assessment (IQOLA) Project.
Qual Life Res 1992;1:349–51.
Brooks RG, EuroQol Group. EuroQol: the current state of play. Health Policy 1996;37:
53–72.
EuroQol Group. EQ-5D user guide. Rotterdam, The Netherlands: EuroQol Business
Management, 1996.
van der Ploeg HM, Defares PB, Spielberger
CD, editors. Manual of the state–trait anxiety
inventory (Handleiding bij de zelfbeoordelingsvragenlijst) (in Dutch). Leiden, The Netherlands: Swets & Zeitlinger, 1980.
van der Ploeg HM. Validation of the state–trait
anxiety inventory (Validatie van de zelfbeoordelingsvragenlijst) (in Dutch). Ned T
Psychol 1980;35:243–9.
Spielberger CD, Gorschuch RL, Lushene RE,
editors. STAI manual for the state–trait anxiety
inventory. Palo Alto (CA): Consulting Psychologists Press, 1970.
Rietbergen JB, Kruger AE, Kranse R, Schroder
FH. Complications of transrectal ultrasoundguided systematic sextant biopsies of the prostate: evaluation of complication rates and risk
factors within a population-based screening
program. Urology 1997;49:875–80.
Sutton S, Saidi G, Bickler G, Hunter H. Does
routine screening for breast cancer raise anxiety? Results from a three wave prospective
study in England. J Epidemiol Community
Health 1995;49:413–8.
Lerman C, Trock B, Rimer BK, Jepson C,
Brody D, Boyce A. Psychological side effects
of breast cancer screening. Health Psychol
1991;10:259–67.
Cockburn J, Staples M, Hurley SF, De Luise T.
Psychological consequences of screening
mammography. J Med Screen 1994;1:7–12.
Notes
Present address: M. L. Essink-Bot, Institute for
Medical Technology Assessment, Erasmus University Rotterdam, The Netherlands.
Present address: H. G. T. Nijs, Municipal Health
Services Zuid-Holland Zuid, Dordrecht, The Netherlands.
We thank the personnel of the Department of
Urology at the University Hospital Rotterdam for
administrative support and all respondents for participating in this study. We also thank Joanna B.
Madalinska for her valuable comments on an earlier
version of this manuscript.
Manuscript received November 26, 1997; revised
March 17, 1998; accepted April 7, 1998.
REPORTS 931