Conceptual equivalence and health-related quality of life: an exploratory
study in Japanese and Dutch cancer patients
W. Chr. Kleijn1,2, K. Ogoshi3, K. Yamaoka4, T. Shigehisa5, Y. Takeda6, C. L. Creutzberg7,
J. W. R. Nortier7& A. A. Kaptein1
1Leiden University Medical Center, Medical Psychology, P. O. Box 9555, 2300 RB, Leiden, The Netherlands,
3Department of Surgery, School of Medicine, Tokai University, Tokyo, Japan;4Department of Technology
Assessment and Biostatistics,National Institute of Public Health, Saitama, Japan;5Department of Psychology,
Tokyo Kasei Gakuin University, Tokyo, Japan;6Department of Pulmonary Medicine, International Medical
Center of Japan, Tokyo, Japan;7Department of Clinical Oncology, Leiden University Medical Center, Leiden,
2Centrum’ 45, Rijnzichtweg 35, 2342 AX, Oegstgeest, The Netherlands;
Accepted in revised form 1 February 2006
Research into the equivalence of Western and Japanese conceptualizations of health-related quality of life
(HR-QOL) is scarce. We used the Western (European Organization for Research and Treatment of Cancer,
EORTC-QLQ-C30) and the Japanese (HRQoL-20) questionnaire in order to analyze the conceptual
similarity of HR-QOL factors, and the associations between specific symptom items with overall HR-QOL
in Japanese (n=265) and Dutch (n=174) patients with various types of cancer. Both populations com-
pleted both instruments. In both patient groups, the overall health scale of the EORTC-QLQ-C30 corre-
lated highly (r=0.59; p<0.001) with the HRQOL-20 composite average score, indicating substantial
conceptual comparability. Relationships between all EORTC-QLQ-C30 symptom items with HR-QOL
were examined by ranking their correlations with the two overall measures of HR-QOL. Comparable
patterns in the Japanese and Dutch samples were observed. The results suggest a considerable conceptual
equivalence of HR-QOL in Japanese and Dutch cancer patients, and indicate a satisfactory structural and
cross-cultural equivalence for the EORTC-QLQ-C30 with regard to items measuring functioning and
specific symptoms. Longitudinal studies are needed to examine the impact of specific symptoms on general
quality of life.
Key words: Cancer, Health-Related Quality of Life, Conceptual equivalence, EORTC-QLQ-C30,
The growing number of internationally organized
clinical trials and comparative studies in cancer
patients corroborate the importance of cross-
cultural validation and standardization of health-
related quality of life (HR-QOL) measurements
[1–3]. In order to examine the HR-QOL of cancer
patients in a cross-culturally valid way, the Euro-
pean Organization for Research and Treatment of
Cancer (EORTC) has developed various QOL
questionnaires that have become standard HR-
QOL instruments. The EORTC-QLQ-C30 is a
generic questionnaire in this set of measures .
Quality of Life Research (2006) 15: 1091–1101
? Springer 2006
This questionnaire has been used in more than
3000 studies worldwide. As a copyrighted instru-
ment, it has been translated and validated into
more than 50 languages, including Dutch and
According to Naito et al. , researchers in
English-speaking countries have pioneered the use
and standardization of HR-QOL measurements.
These authors state that clinicians in Japan report-
ing on QOL research have generally been interested
in the use of methods that primarily measure the
physical aspects of QOL. However, some Japanese
researchers have evaluated the multidimensional
QOL of cancer patients, employing either Japanese
translations of Western measures, or HR-QOL
instruments developed in Japan such as the
Kobayashi et al.  translated the EORTC-
QLQ-C30 into Japanese and compared the scores
of Japanese lung cancer patients with European
patients. They stated that ‘‘both the Japanese
language and culture are vastly different from
those of English speaking countries’’, but con-
firmed the cross-cultural functional equivalence
and concluded that this instrument can be used in
international phase III studies. In his comments on
the results of Kobayashi et al.’s study, Aaronson
 draws attention to the particular importance of
the relatively high correlations observed in the
Japanese sample between scales assessing physical
and psychological health domains, stating that ‘‘...
one cannot rule out ... the cross-cultural explana-
tion ... that the Japanese culture adheres less
strongly to a Cartesian view of health than is the
case in the West; that the illness experience of
Japanese patients is of a more holistic nature.’’ [7,
p. 768]. To study cross-cultural equivalence,
several dimensions were proposed by Hui and
Triandis : functional equivalence (adequacy of
translation), scale equivalence (comparability of
scales), operational equivalence (standardization
of testing procedure), and metric equivalence
(transferability of scoring results from one culture
However Fayers et al. [9, 10] have pointed to two
peculiarities that are characteristic of many health
used ones like the EORTC-QLQ-C30, which com-
plicate the study of cross-cultural equivalence.
Theseinstrumentstypically measureconcepts using
a combination of subscales and scale types. HR-
QOL subscales generally consist of a number of
items measuring concepts such as emotional, phys-
ical, or social well-being, plus single-item scales
measuring specific symptoms. Factor analytic
methods are appropriate to examine structural
characteristics of instruments that measure certain
concepts. But in case of QOL instruments such as
the EORTC-QLQ-C30, such methods are only
useful for a subset of the items: only for the items
defining certain constructs. Disease-specific QOL
assess symptoms or treatment side effects. These
patients experiencing them, but the reverse rela-
tionship need not apply: a poor level of QOL does
not imply that the patient suffers from that specific
analytic strategy proposed by Fayers et al. and use
their method of distinguishing between different
of structural equation modelling, these authors
named the items belonging to subscales that pri-
marily reflect the level of general QOL ‘effect indi-
of HR-QOL. Items which typically assess specific
disease or treatment symptoms (such as ‘pain’ or
‘vomiting’) are called a-priori, ‘causal indicators’,
because the occurrence of these symptoms could
cause a change in the level of general QOL.
Apart from this distinction between ‘effect
indicators’ and ‘causal indicators’, we make a
distinction between ‘overall’ QOL and ‘domain
specific’ QOL in this study. Both QOL instruments
used in this study measure domain specific QOL
by using sets of conceptually related items mea-
suring cognitive, emotional or social functioning.
In addition, overall QOL is measured by one single
item in the EORTC-QLQ-C30, while overall QOL
is measured by the HRQoL-20 by calculating a
mean score based on all 20 items. Our primary
objective was to use the distinction in types of
indicators and the different measures of HR-QOL
in analyses of conceptual equivalence, a strategy
which has not been used in previous cross-cultural
As mentioned before few studies have focused
on the issue of equivalence of conceptualizations
of HR-QOL between Western and Japanese
cancer patients. Japanese HR-QOL instruments
have not yet been tested in Western populations of
cancer patients. So their potential cultural bias or
equivalence in measuring QOL is not known.
However, the use of instruments from both cul-
tures in a cross-cultural investigation can be highly
informative. Therefore, in our study both Western
and Japanese QOL instruments were employed to
examine HR-QOL in both populations.
To summarize, this study had three aims: (1) to
compare the relations of the specific symptom
items of the EORTC-QLQ-C30 to measures of
overall HR-QOL in Japanese and Dutch patients
with the aim of analyzing conceptual equivalence;
(2) to examine associations of quality of life with
background variables such as country of origin,
gender, age and illness stage; and (3) to describe
and compare quality of life in Japanese and Dutch
groups of cancer patients, using two different QOL
scales developed in Japan and Europe.
The data for an exploratory analysis on HR-QOL
were selected from two ongoing Japanese and
Dutch studies, both of which included patients
with cancer [11, 12]. Because these studies used
translated versions of the same HR-QOL instru-
ments, and because both the HRQoL-20 and the
EORTC-QLQ-C30 were used in both study sam-
ples, it was appropriate to combine and compare
the data on these common variables. Patient illness
characteristics were also examined. This combined
dataset was used to analyze cross-cultural simi-
larities and differences in the patterns of HR-QOL
The data collection procedure was the same in
both the Netherlands and Japan: during a consul-
tation visit at the hospital, patients who had
already undergone at least some form of treatment
(surgery, chemotherapy, etc.) were invited to par-
ticipate in the study by their physician. Other
inclusion criteria were cancer site (lung, colorectal
and stomach in Japan; lung, colorectal, breast and
prostate in The Netherlands) and the absence of
any apparent psychiatric disorder (as judged by
the consulted physician). After informed consent
procedures were completed, patients received a
booklet including the Japanese HRQoL-20 as well
as the European EORTC-QLQ-C30 plus various
other instruments not reported on in this study.
All consecutive patients from the participating
oncology departments that fulfilled the inclusion
criteria were invited to participate into the study,
reflecting the incidence rates of the most common
cancer sites in Japan and the Netherlands. The
Japanese (n=265) and Dutch (n=176) patient
groups differed with respect to cancer site, gender
(the Dutch group included more females), and
Tumor Node Metastasis (TNM) stage (the Japa-
nese group included more TNM stage IV patients)
Two self-report instruments measured HR-QOL:
the 20-item Health Related Quality of Life ques-
tionnaire (HRQoL-20) , and the 30-item Euro-
pean Organization for Research and Treatment of
The HRQoL-20 was developed in Japan and
measures generic HR-QOL. It comprises 20 items
from the physical, psychological, and social
domains (see Appendix). Items can be scored by
marking one of three available categories (coded 1,
2 or 3, relating to a negative, intermediate, or
positive response to the item, respectively). In
addition to the 3 subscales measuring physical (8
items), psychological (8 items) and social (4 items)
QOL, a measure of general QOL is derived by
calculating an overall composite HR-QOL score
based on the average of all 20 items . Results
for each scale are presented as the mean score of
the items of the scale, with higher scores repre-
senting better QOL. The HRQoL-20 has been used
in research with healthy subjects , and with
cancer and non-cancer patients . The Dutch
adaptation was constructed after a forward and
backward translation procedure (from English).
The EORTC-QLQ-C30 is a 30-item question-
naire developed to assess the QOL of cancer
patients. It has been translated and validated into
many languages, among which are Dutch and
Japanese [2, 4]. It comprises two overall scales (one
on general health and one on overall QOL), five
functional subscales (measuring role, emotional,
cognitive and social functioning), and nine symp-
tom subscales (measuring specific symptoms). All
21 items specifying certain symptoms are listed in
Table 4. The overall QOL and general health
scales are scored using a 7-point scale, and all
other subscales are scored using a 4-point Likert
scale. The overall QOL scale is comprised of the
single EORTC-QLQ-C30 item 30 (‘‘How would
you rate your overall quality of life during the past
week?’’). The complete questionnaire can be
obtained at the EORTC website (http://www.
Although both instruments are similar in that
they measure certain aspects of HR-QOL, how-
ever they differ according to their item content,
construction and lay-out, the scoring, the number
of items and scales.
calculated for the two samples with respect to all
HR-QOL scales. To analyze conceptual similari-
ties, Pearson correlation coefficients were com-
putedbetween the European
HR-QOL scales, separately for the Japanese and
Dutch patient groups.
Differences in HR-QOL between the Japanese
and Dutch patient groups were analyzed by a
MANOVA followed by independent ANOVA’s
(2 ? 2 ? 4), using the various QOL scales as
dependent variables, the country of origin, gender
and TNM-stage as factors, and age as a covariate.
Partial g (a partial correlation coefficient) was
applied as an effect-size parameter to examine the
influence of the factors and covariate, in explain-
ing differences in QOL between the Japanese and
Dutch patients (squared g can be used as a mea-
sure of percentage of the variance explained by the
Similarity of the HR-QOL construct between
Dutch and Japanese patient groups on the level of
individual symptoms was analyzed by following
the approach of Fayers and Hand . Each specific
EORTC-QLQ-C30 symptom item (a-priori defined
as a potential ‘causal indicator’ of HR-QOL) was
ranked according to the Pearson correlation coef-
ficient of each item with the overall measure of
Table 1. Demographic and illness characteristics of Japanese and Dutch patients
Age (mean; SD)
av2indicates differences between Japanese and Dutch patients.
bTNM stage: Tumor Node Metastasis stage.
cDue to missing values, N=207 (Japanese) and 153 (Dutch).
HR-QOL: EORTC-QLQ-C30 item 30. The same
procedure was followed for the composite mean
score of the HRQoL-20. This was done separately
for the Japanese and Dutch patient groups. These
results were compared to those of Fayers and Hand
as a means of reference.
Finally, similarities between the different rank-
ings of the different patient groups were examined
using Spearman’s q rank correlations.
As shown in Table 1 distributions of gender,
cancer site and TNM stage were significantly
related to country of origin, with relatively more
male patients in the Japanese group. Stomach and
lung cancer were the dominant types of cancer in
the Japanese group, while breast and prostate
cancer were the dominant types of cancer in the
Dutch group. The Japanese group consisted of
more TNM stage IV patients. Mean age was not
significantly different between the two groups.
Reliability coefficients were calculated for the
overall composite scale and the subscales of the
Japanese HRQoL-20, and for the overall QOL
scale and the functioning scales of the European
EORTC-QLQ instrument. Table 2 presents the
Cronbach’s a coefficients showing comparable
values, with only a few exceptions. The Cron-
bach’s a coefficients for the overall HRQoL-20
and overall EORTC-QLQ scales were good for
both samples. Most a values of the subscales were
also satisfactorily high, but there were some
exceptions: the a coefficient of the HR-QoL-20
social subscale was low in both samples (range
0.38–0.16). Also, the values of the social and
cognitive subscales of the EORTC-QLQ were
rather low in both patient groups (range 0.47–
0.63), only reaching marginal values of reliability.
Conceptual similarities were analyzed by com-
puting the correlations between the Japanese
HRQoL-20 scales and the scales of the Western
EORTC-QLQ-C30, separately for the Japanese
and Dutch patient groups (see Table 2). The pat-
terns of the various correlation coefficients are
remarkably similar between the two groups. The
EORTC-QLQ-C30 overall QOL scale correlated
HRQoL-20 scale in both groups, indicating sub-
stantial conceptual equivalence. Similar values
were found when comparing the correlations be-
tween the physical and emotional EORTC-QLQ-
C30 subscales with the physical and psychological
Table 2. Cronbach’s a coefficients of quality of life scales, and correlations between Japanese and European quality of life scales,
in Japanese and Dutch patients
Japanese patientsDutch patients
Overall quality of life
EORTC-QLQ-C30 overall scalesb
Quality of life
EORTC-QLQ-C30 functioning scalesb
*p<0.05; **p<0.01; ***p<0.001.
aHRQoL-20: Health Related Quality of Life questionnaire-20.
bEORTC-QLQ-C30: European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire C-30.
HRQoL-20 scales. An exception to these observed
coherent patterns is the correlation between
the subscales assessing the social domain of the
Japanese and European instruments. As might be
expected from the low reliability values of these
subscales, there was no substantial correlation
between these subscales of the two instruments.
Not shown in Table 2 are the correlations be-
EORTC-QLQ-C30, but they proved to be low:
r=0.17 (p<0.001) in the Japanese and r=0.36
Differences in quality of life scores
All the scores of the HR-QOL scales (see Table 3)
country, gender, and TNM-stage as factors and age
as a covariate. Country (F=3.80; df=19,292;
p<0.001), gender (F=3.08; df=19,292; p<0.001)
and age (F=4.03; df=19,292; p<0.001) were all
the interaction between country, gender, and
TNM-stage proved to be significant, although to a
much lesser degree (F=1.48; df=57,882; p<0.05).
Table 3 shows the mean values of all the HR-QOL
scales (as used in the MANOVA) for the Japanese
the results of the univariate ANOVA’s on these
As Table 3 shows, the absolute values on the
Japanese HRQoL-20 and on the Western EO-
RTC-QLQ-C30 overall and functioning scales
were all higher (meaning a better QOL) in the
Table 3. Differences in quality of life scores between Japanese and Dutch patients
Japanese patientsDutch patientsEffect size (Partial g, F-test)d
MSDM SD CountryGenderAgeStage
Overall quality of life
EORTC-QLQ-C30 overall scalesb
Quality of life
EORTC-QLQ-C30 functioning scalesb
EORTC-QLQ-C30 symptom scalesc
HRQoL-20: Health Related Quality of Life questionnaire-20; EORTC-QLQ-C30: European Organization for Research and Treatment
of Cancer Core Quality of Life Questionnaire C-30.
aScores can range from 1 to 3 with higher scores indicating better QOL.
bScores can range from 0 to 100 with higher scores indicating better QOL functioning.
cScores can range from 0 to 100 with higher scores indicating more severe symptoms.
dOnly significant partial correlations are shown (*p<0.05, **p<0.01, ***p<0.001).
Dutch sample. For all the HRQoL-20 scales and
for the EORTC-QLQ-C30 overall QOL scale,
these differences were significantly related to
country. In case of the physical, psychological and
overall subscales of the HRQoL-20 and also of
the EORTC-QLQ-C30’s overall scales and role
functioning subscale, these differences were also
significantly but negatively related to TNM-stage
(if the illness invasion level was worse, then QOL
values were lower).
All scores on the EORTC-QLQ-C30 symptom
scales were higher (meaning more symptoms) in
the Japanese patient group (except for nausea
and insomnia). Pain, dyspnea and appetite loss
were significantly related to TNM-stage, and
appetite loss, constipation and financial difficulties
were significantly related to country. A few
statistically significant interaction effects were
found, but they were rather small: one Coun-
try ? TNM-stage interaction for the EORTC-
QLQ-C30 dyspnea scale (p<0.05), and four
Country ? Gender ? TNM-stage
for the HRQoL-20 social scale (p<0.01), and the
EORTC-QLQ-C30 scales of overall health, emo-
tional well-being and constipation (all p<0.05).
Causal indicators’ of quality of life
Information on the possible (theoretical) causal
relationship between each specific a-priori indica-
tor with overall HR-QOL can be found in Table 4.
Pearson correlations of each of the 21 individual
EORTC-QLQ-C30 symptom items with the two
different measures of overall QOL (EORTC-QLQ-
C30 item 30 and average HRQoL-20) are pre-
sented separately for the Japanese and Dutch
samples. In addition, the rankings of these symp-
tom items are presented. The rankings of items in
the first column were based on their correlations
with the overall EORTC-QLQ-C30 scale as was
Table 4. Correlations and rankings (in parentheses) between specific EORTC-QLQ-C30 symptom items and two overall measures
of quality of life
Ranked EORTC items Correlation with overall
Correlation with overall
Correlations with overall
JapaneseDutch JapaneseDutch Fayers and Handc
Fayers and Handd
Lack concentration (Q20)
Need rest (Q10)
Family life (Q26)
Pain interference (Q19)
Social activities (Q27)
Shortness of breath (Q8)
Financial problems (Q28)
aOverall QOL scale EORTC as defined by European Organization for Research and Treatment of Cancer Core Quality of Life
Questionnaire C-30 item 30.
bOverall QOL scale HRQoL-20 as a composite score derived from the Health Related Quality of Life questionnaire-20.
cData of Norwegian head and neck cancer patients as reported by Fayers Nd Hand (10).
dData of Danish primary breast cancer patients as reported by Fayers and Hand (10).
found in the Japanese patients group. Relatively
high correlations suggest that an item belongs to
the group of indicators that define core aspects of
general HR-QOL. Relatively low correlations may
point to a ‘causal’ relationship between the specific
item and overall HR-QOL. Although such an item
can ‘influence’ the level of QOL, the reverse is not
necessarily true. Taken as a single symptom such
items are not suitable to be included in the defi-
nition of QOL as a construct.
Fayers and Hand  published similar correla-
tions in a study on Norwegian patients with head
and neck cancer and Danish women with breast
cancer. Therefore, the correlations from their study
can be used as a reference and they are also shown
in Table 4. In the Japanese patient group, EORTC-
QLQ-C30 items (‘lack of concentration,’ ‘the need
to rest,’ ‘insomnia,’ ‘interference with family life,’
‘tiredness,’ and ‘depression’) showed relatively high
correlations with the overall EORTC-QOL-C30
score. Because of these substantial associations,
these items can be conceptualized (at least for the
Japanese group)as belongingto the core concept of
HR-QOL. The items with low rankings in this
sample (‘diarrhea,’ ‘financial problems,’ ‘vomiting,’
‘constipation,’ and ‘nausea’) showed relatively low
associations with the overall HR-QOL, and can
therefore be conceived as belonging to the category
of causal indicators of QOL.
In comparing the six rankings in Table 4 on the
Japanese patient group, the Dutch group and the
groups of Fayers and Hand, many similarities
emerge but also some discrepancies can be
observed. Table 5 shows Spearman’s rank corre-
lations of all the combinations of the six available
rankings. In general, the rank correlations are
statistically significant and mostly very high.
Correlations obtained with the EORTC-QLQ-C30
ranged from q=0.63 to q=0.93, an indication of a
rather good cross-cultural structural equivalence
for this instrument with regard to the symptom
items. Using the average HRQoL-20 score as a
measure of overall QOL, some lower rank corre-
lations were found especially in the Japanese
patient group (q=0.41–0.64) compared with the
Dutch group (q=0.56–0.84).
Internationally organized clinical trials on the
treatment of cancer patients have stimulated the
development of cross-culturally valid QOL instru-
ments, such as the EORTC-QLQ-C30. The use of
these instruments facilitates the comparison of
outcomes on QOL from different countries and
increases their accuracy. Although the integration
of information on the various indicators of
Table 5. Spearman’s q correlations between rankingsaof EORTC-QLQ-C30 symptom items in different groups of patients
EORTC- QLQ-C30 HRQoL-20 EORTC- QLQ-C30
Japanese Dutch Japanese Dutch Fayers and Handd
Fayers and Hande
Fayers and Handd
Fayers and Hande
aSee ranked correlations as depicted in Table 4.
bOverall QOL scale EORTC as defined by European Organization for Research and Treatment of Cancer Core Quality of Life
Questionnaire C-30 item 30.
cOverall QOL scale HRQoL-20 as a composite score derived from the Health Related Quality of Life questionnaire-20.
dData of Norwegian head and neck cancer patients as reported by Fayers and Hand (10).
eData of Danish primary breast cancer patients as reported by Fayers and Hand (10).
intercultural equivalence of QOL instruments is
important and valuable, it is often a slow process
and seldom the main objective of QOL studies.
Comparing the conceptual
HR-QOL in Japanese and Dutch cancer patients
was a major goal of the present study.
Examining the associations of HR-QOL out-
comes with variables such as country of origin and
illness characteristics was also an important part
of the study. Most of the overall, functioning and
domain scales showed internal consistency values
that were sufficiently high, regardless of the origin
of the instrument or country of origin of the
patient group. Exceptions to this rule were the a
coefficients found for the social and cognitive
scales. Values of the overall scales were compara-
ble to results from other studies [3, 15].
Although the overall EORTC-QLQ-C30 scale
and the overall HRQoL-20 measure general HR-
QOL quite differently, the cross-correlations of
these overall scales showed a substantial and
robust cross-cultural equivalence of the QOL
concept. We can only partly confirm the observa-
tions of Kobayashi et al.  of relatively high
correlations in Japanese patients between the
somatic and emotional scales of the EORTC-
QLQ-C30 and Aaronson et al.’s  relatively
lower correlations of these scales in European
patients. We found a high correlation between the
Physical and Psychological scales of the HRQoL-
20, but we did so for both patients groups, irre-
spective of country of origin. Comparison of our
results with research findings from the field of
general subjective well-being (SWB) shows some
similarities and some differences. As Diener et al.
[16, 17] reported in a review of measurement
validity across cultures, that although cross cul-
tural comparisons of SWB have certainly some
degree of validity, some authors report evidence of
a slightly different structure for SWB items across
cultures. Our results, while using measures of well-
being in a health-related setting, did only partly
support these observations. Particularly measure-
ments in the physical and psychological domains
indicated a good degree of conceptual equivalence,
but some problems were encountered in the social
Ranking the 21 EORTC QLQ-C30 symptom
items according to their correlations with overall
measures of HR-QOL has enabled the evaluation
of the a-priori classification of symptom items as
‘causal indicators’. The various rankings showed
substantial similarities, but some discrepancies
could also be observed.
A consequence of the procedure of this study
was the use of consecutive samples of patients.
This limited our possibilities to untangle some of
the confounded factors. Some patient character-
istics like TNM stage or disease site are related to
country of origin. A study using a controlled
design is needed to corroborate our main conclu-
sions. In addition, in order to be able to evaluate
the impact of these groups of symptoms in cancer
clinical trials, studies with a longitudinal design are
Dean  pointed out how cultural differences in
explanatory models of cancer experience shape the
responses of patients to their illness, participation
in screening, compliance with treatment, and their
HRQoL-20 developed in Japan, clearly contains
items that reflect Japanese explanatory models
about patients’ responses to cancer. For example,
the HRQoL-20 item that asks about whether the
respondent has ‘‘taken up gardening’’ or ‘‘taking
care of a pet’’ refers to social isolation in the Jap-
anese culture, as it reflects turning away from
humans. In North American or European culture,
taking care of pets or taking up gardening, as a
response to facing cancer would be viewed as a
constructive coping response to a taxing situation.
The EORTC QLQ-C30 has been used in a re-
cent study that examined QOL in a multicultural
population of cancer patients from Asian and non-
Asian Pacific Islanders . The authors report
comparable results and conclude that the EORTC-
QLQ-C30 seems viable for the multi-cultural
samples examined. Factor analytic techniques ap-
plied to the patients’ scores on the EORTC-QLQ-
C30 items supported the dimensions as defined by
its developers, especially regarding factors reflect-
ing the concept of social support. In our samples
(Dutch and Japanese), we too found support for
the viability of the EORTC QLQ-C30, with some
reservations about dimensions that purport to
assess social functioning. This social domain cer-
tainly necessitates further research. Ishikawa et al.
 and Kaplan et al.  studied physician–pa-
tient communication patterns in a Japanese cancer
setting. They found that these patterns have a
strong biomedical emphasis compared to Western
patient–physician communication patterns ,
illustrating this issue.
In summary, the present study underlines the
cross-cultural viability of the EORTC-QLQ-C30
in samples of Dutch and Japanese patients with
various types of cancer. The concepts of ‘causal
indicators’ and ‘effect indicators’, introduced by
Fayers and Hand , were also found to be rele-
vant in these patient groups. Future research
assessing QOL in Japanese and Dutch samples
would benefit from using additional questionnaires
(e.g., IPQ-R, see http://www.uib.no/ipq) in order
to shed light on explanatory models or illness
representations and patients’ communication style
preferences in relation to QOL. This additional
information could also be of help in shaping po-
tential interventions on illness perceptions of these
patients . Longitudinal designs would be
instrumental in clarifying the issue of causal and
effect indicators in this type of research.
Part of this study was made possible by financial sup-
port of Pfizer Foundation of Japan.We like to thank
Dr. H. Vos-Westerman, Dr. R. Bevers, Dr. L.
Willems, Dr. A. Dingemans, Dr. R.A.E.M. Toll-
enaar, Dr. G.G. Kenter and Dr. F.W. Dekker for
their support, and Josefine Pouw, Femke Blonk,
Elske van den Berg, Emma Tiesma and Wil den
Heyer for their assistance in data acquisition.
1. Naito M, Nakayama T, Fukuhara S. Quality of life
assessment and reporting in randomized controlled trials:
A study of literature published from Japan. Health Qual
Life Outcomes 2004; 2: 31–37.
2. Aaronson NK, Ahmedzai S, Bergman B, et al., for the
EORTC Study Group on QOL. The European Organiza-
tion for Research and Treatment of Cancer QLQ-C30: A
quality-of-life instrument for use in international clinical
trials in oncology. J Natl Cancer Inst 1993; 85: 365–376.
3. Kobayashi K, Green J, Shimonagayoshi M, et al. Valida-
tion of the care notebook for measuring physical, mental
and life well-being of patients with cancer. Qual Life Res
2005; 14: 1035–1043.
4. Kobayashi K, Takeda F, Teramukai S, et al. A cross-
validation of the European Organization for Research and
Treatment of Cancer QLQ-C30 (EORTC QLQ-C30) for
Japanese with lung cancer. Eur J Cancer 1998; 34: 810–
5. Yamaoka K, Ogoshi K, Haruyama K, et al. Validity of the
Japanese version of the questionnaire for quality of life
measurement (QOL20). Int Med J 1998; 5: 23–29.
6. Yamaoka K, Hayashi F, Hayashi C, et al. A Japanese
version of the questionnaire for quality of life measure-
ment. Ann Cancer Res Ther 1994; 3: 45–53.
7. Aaronson NK. Assessing the quality of life of patients with
cancer: East meets west. Eur J Cancer 1998; 34: 767–769.
8. Hui C, Triandis HC. Measurement in cross-cultural
psychology: A review and comparison of strategies. Cross
Cult Psychol 1985; 16: 131–152.
9. Fayers PM, Hand DJ. Factor analysis, causal indicators
and quality of life. Qual Life Res 1997; 6: 139–150.
10. Fayers PM, Hand DJ, Bjordal K, Groenvold M. Causal
indicators in quality of life research. Qual Life Res 1997; 6:
RelatedQualityof Lifequestionnaire items
1 PhHow would you describe your
state of health?
How often do you feel tired?
How would you describe your mood?
Are you able to do the physical activities
you would like to do?
Do you have worries in everyday live?
Can you handle stress?
How often do you have swollen legs?
How often are you in physical pain?
How often do you wonder if your
illness is incurable?
How often do you feel irritated?
How often do you feel ill?
How often do you feel as if you
How often do you feel lonely?
Do you have a good appetite?
Do you sleep well?
Do you feel that members of your family
or other people need you?
How often are you worried about
the cost of living?
How would you describe your
relationship with your family,
friends, and neighbors?
Do you have pets or do you like gardening?
Are you satisfied with your current life?
Ph, Ps or So refer to items belonging to the Physical, Psycho-
logical and Social subscales.
11. Yamaoka K, Hayashi F, Hayashi C, Shigehisa T, Watan-
abe M. Influence of personality on quality of life measure-
ment. Jap J Health Psychol 1996; 9: 11–20.
12. Kaplan RM, Shigehisa T, Oldenburg B. Health policy,
international variation and doctor–patient interaction. Jap
J Health Psychol 2003; 4: 23–49.
13. Shigehisa T. Eysenckian personality traits and health-
related quality of life in patients with stomach, colorectal or
breast cancer. Jap J Health Psychol 2003; 10: 19–32.
14. Yamaoka K, Shigehisa T, Ogoshi K, et al. Health-related
quality of life varies with personality types: A comparison
among cancer patients, non-cancer patients and healthy
individuals in a Japanese population. Qual Life Res 1998; 7:
15. Yamaoka K, Takeda F, Shigehisa T, et al. Health-related
quality of life in Japanese lung cancer patients as deter-
mined by two questionnaires: The HRQOL-20 and the
EORTC QLQ-C30. Ann Cancer Res Ther 2003; 11: 31–46.
16. Diener E, Oishi S, Lucas R. Personality, culture, and
Subjective well-being: Emotional and cognitive evaluations
of life. Annu Rev Psychol 2003; 54: 403–425.
17. Diener E, Suh E, Smit H, Shao L. National differences in
subjective well-being: Why do they occur?. Soc Indic Res
1995; 34: 7–32.
18. Dean S. Explanatory models and attitudes towards cancer
in different cultures. Lancet Oncol 2004; 5: 119–124.
19. Gotay CC, Blaine D, Haynes SN, Holup J, Pagano IS.
Assessment of quality of life in a multicultural can-
20. Ishikawa H, Takayama T, Yamazaki Y, Seki Y, Katsu-
mata N. Physician–patient communication and patient
satisfaction in Japanese cancer consultations. Soc Sci Med
2002; 55: 301–311.
21. Kaptein AA, Kleijn WC, Nortier H. Doctor-patient
communication: A European perspective. Ann Cancer
Res Ther 2003; 11: 7–14.
22. Petrie KJ, Cameron LD, Ellis CJ, Buick D, Weinman J.
Changing illness perceptions after myocardial infarction:
An early intervention trial. Psychosom Med 2002; 64: 580–
Address for correspondence: Wim Chr. Kleijn, Leiden University
Medical Center, Medical Psychology, P.O. Box 9555, 2300 RB
Leiden, The Netherlands
Phone: +31-71-527-5242; Fax: +31-71-527-3668