Rheumatology 2012;51:1906–1915

RHEUMATOLOGY doi:10.1093/rheumatology/kes165
Advance Access publication 22 July 2012

Original article
LupusQoL-FR is valid to assess quality of life in
patients with systemic lupus erythematosus
Hervé Devilliers1,2, Zahir Amoura3,4, Jean-Francois Besancenot1,
Bernard Bonnotte5,6, Jean-Louis Pasquali7,8, Denis Wahl9,10,
Francois Maurier11, Pierre Kaminsky12, Jean-Loup Pennaforte13,
Nadine Magy-Bertrand14, Laurent Arnaud3, Christine Binquet2 and
Francis Guillemin15,16

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
Abstract
Objective. To cross-culturally adapt the LupusQoL into French, to test its measurement properties and to
further investigate its domain structure.
Methods. The cultural adaptation process according to guidelines and pre-testing resulted in the
LupusQoL-FR. SLE patients completed the LupusQoL-FR at baseline, 15 days, 3 months and 6
months. Validity was studied through content and construct validity (factorial and Rasch analysis for
structural validity, Spearman’s correlation and Mann–Whitney tests for external validity). Cronbach’s a
and intra-class correlation coefficients were computed for reliability. The standardized response mean
was computed to evaluate responsiveness.
Results. In all, 182 patients, age 39.6 (10.6) years, mostly outpatients [mean SELENA-SLEDAI 2.6 (3.5)]
were recruited. Factor analysis with eight imposed factors was very close to the original LupusQoL. A
screeplot with parallel analysis showed that LupusQoL domains could be aggregated in two physical and
mental scales. Both eight- and two-factor structures showed a good Rasch fit, internal consistency
(Cronbach’s a: 0.85–0.95), and test–retest reliability (intra-class correlation coefficient 0.79–0.95).
External convergent (correlation with SF-36, r = 0.59–0.78) and divergent validity (according to
CLINICAL
SCIENCE

SELENA-SLEDAI) were also satisfactory.

Conclusion. The LupusQoL-FR is valid to assess quality of life in SLE patients. A two-factor structure of
physical and mental aggregated scales is a valid alternative to the original eight-domain structure.
Key words: systemic lupus erythematosus, quality of life, outcome assessment, psychometrics.

Introduction However, recent studies showed that SLE-related

physical, psychological, emotional and social burdens
Recent advances in medical care have considerably cause considerable deterioration of patients’ everyday
extended life expectancy in patients with SLE [1]. life [2–7]. In past decades, the health-related quality of

Medicine and Clinical Immunology, Metz, 12Department of

1
Department of Internal Medicine and Systemic Diseases, University
Hospital of Dijon, Dijon Cedex, 2Clinical Investigation Centre - Clinical
Epidemiology/Clinical Trials, INSERM CIE1, 3Department of Internal
Medicine, University Hospital Pitié-Salpêtrière, AP-HP, National
Referral Center for Systemic Lupus Erythematosus and
Antiphospholipid Antibody Syndrome, 4Université Pierre et Marie Curie
Univ Paris 06, Paris, 5Department of Internal Medicine and Clinical
Immunology, University Hospital of Dijon, Dijon, 6INSERM U1098,
University Hospital of Dijon, 7Clinical Immunology Department,
University Hospital of Strasbourg, Nouvel Hopital Civil, 8CNRS
UPR9021, 9Department of Vascular Medicine, University Hospital of
Nancy, Strasbourg, 10INSERM U961, Vandoeuvre-lès-Nancy,
11
Department of Internal Medicine, Ste Blandine Hospital, Vascular
Internal Medicine, Systemic Diseases and Rare Diseases, University
Hospital Nancy, Vandoeuvre-lès-Nancy, 13Department of Internal
Medicine, University Hospital Robert Debré, Reims, 14Department of
Internal Medicine, University Hospital Jean Minjoz, Besançon,
15
University Hospital, Clinical Epidemiology and Evaluation
Department, INSERM, CIC-EC CIE6, Vandoeuvre-lès-Nancy and
16
Public Health School, Nancy University, Paris Descartes, Nancy,
France.
Submitted 24 November 2011; revised version accepted 10 May 2012.
Correspondence to: Hervé Devilliers, Department of Internal Medicine
and Systemic Disease, Dijon University Hospital, 3 rue Faubourg
Raines, 21000 Dijon Cedex, France. E-mail: hervedevilliers@free.fr

! The Author 2012. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com

life (HRQOL) of SLE patients has been assessed with
generic questionnaires, such as the Medical Outcome
Study Short-Form 36-item Health Survey (SF-36). Data
from these studies have shown that even with inactive
disease, SLE patients had a poorer HRQOL compared
with healthy subjects [8]. Furthermore, most HRQOL
domain scores did not correlate with disease activity
and damage assessed by physicians [8], which shows
that HRQOL questionnaires provide different and comple-
mentary information. One limitation of previous studies
was that some important issues for SLE patients (such
as sleep, sexual function and body image) were absent
from the generic measures used in SLE [9]. Recently,
lupus-specific HRQOL questionnaires have been de-
veloped to describe the specific impact of SLE on pa-
tients’ everyday life, of which two were available at the
start of this study [10–12]. Of these, LupusQoL was de-
veloped in the UK, in patients that are culturally close to
French patients, although there are substantial differences
between health care systems and ethno-cultural mixes in
the two countries. It showed good psychometric proper-
ties in English-speaking SLE patients in both the UK
[11,13] and the USA [14,15]. To date, no lupus-specific
HRQOL questionnaire has been validated in French. Our
objectives were to cross-culturally adapt the LupusQoL
into French, to test its measurement properties and to
further investigate its domain structure.
Patients and methods
Study design
We conducted a multicentre prospective cohort study in
seven French university hospitals. We included consecu-
tive SLE patients between March and November
2009. The inclusion criteria were age between 18 and
75 years, diagnosis of SLE according to the 1997 modified
ACR criteria [16] and the ability to understand and com-
plete self-report questionnaires. Follow-up comprised two
visits with the patient’s usual SLE specialist (mostly intern-
ists) at baseline and 6 months later and two mail surveys
15 days and 3 months after the baseline visit. The study
was approved by the Research Ethics Committee (Comité
de Protection des Personnes EST 1), and all of the par-
ticipants gave informed consent.
Data collection
Self-report questionnaires, including the LupusQoL-FR,
the Systemic Lupus Activity Questionnaire (SLAQ), the
SF-36 and a self-administered visual analogue scale
(VAS) from 0 to 100 about SLE activity were collected
during each of four assessments. The Safety of Estrogens
in Lupus Erythematosus National Assessment–Lupus
Systemic Lupus Erythematosus Disease Activity Index
(SELENA-SLEDAI) and a Physician Global Assessment
on a 0- to 3-point VAS where 3 represents the most
severe were scored by investigators at day 0.
Investigators were trained to complete the SELENA-
SLEDAI. The first author, who had received certified train-
ing in SELENA-SLEDAI (H.D.) reviewed all medical
www.rheumatology.oxfordjournals.org
LupusQoL-FR in patients with SLE
records and checked the SELENA-SLEDAI. The ethnicity
of the patients was anonymously obtained by asking
each investigator to rate the proportion of each ethnic
group among patients they included in the study.
The LupusQoL is a lupus-specific HRQOL questionnaire
consisting of 34 items grouped in eight domains: physical
health, pain, planning, intimate relationships, burden to
others, emotional health, body image and fatigue.
A score from 0 to 100 (best HRQOL) is calculated for
each domain (the method of scoring is presented in the
supplementary data, available at Rheumatology online).
The SLAQ is a validated patient self-assessed activity
questionnaire with a score ranging from 0 to 48 (greatest
activity). It includes 24 questions related to disease symp-
toms assessed by the Systemic Lupus Activity Measure
Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
that are amenable to self-report [17, 18]. The SF-36 is
a 36-item generic QOL questionnaire consisting of eight
domain scores ranging from 0 to 100 (best HRQOL) [19].
The SELENA-SLEDAI instrument is a score that aims
to rate SLE activity from 0 to 105 [20].
Cross-cultural adaptation process
The cross-cultural adaptation was conducted according
to published international guidelines [21, 22]. The devel-
oper gave her permission to use and translate the
instrument and provided a preliminary French translation.
Two persons (a rheumatologist and one French-mother-
tongue English teacher) independently translated the
LupusQoL into French. An agreed-on version was written
based on the consensus of the two translators and was
back-translated into English by a bilingual native English
speaker, blinded to the original English version. Then,
a multidisciplinary consensus committee comprising the
two translators, an internist, two epidemiologists, includ-
ing one specialized in cross-cultural adaptation met in
order to carry out a review of all the available material
(original version, preliminary translation, two independent
and agreed-on translations and one back-translation) and
to conduct the harmonization step. During this meeting,
the committee checked that the translation was perfectly
understandable, verified cross-cultural semantic, idio-
pathic and conceptual equivalence of the source and
target languages, and produced a pre-final version by
consensus. This version was pre-tested in 10 SLE
French outpatients. Then, cognitive debriefing in a collect-
ive meeting assessed whether they fully understood all
items or had problems with the formulation of the
French items version. The results of the cognitive debrief-
ing helped modify the pre-final version and produce the
final version of the LupusQoL-FR, which was approved
by the developer.
Validation study
We used recent international consensus-based definitions
of measurement properties [23]. We used SAS software
(version 9.1.3, SAS Institute, Inc., Cary, NC, USA), LISREL
8.80 (Scientific Software International, Inc., Stokie, IL,
USA) for confirmatory factor analysis (CFA) and
1907
Hervé Devilliers et al.
Winsteps, version 3.60 (Winsteps, Chicago, IL, USA) for
item response theory analysis.
Validity
Content validity
Patients and SLE experts of the multidisciplinary consen-
sus committee were asked whether all of the items
adequately referred to relevant aspects of the construct
to be measured, comprehensively reflected the construct
and were a representative sample of the total domain for
the study population in which the measurement properties
were evaluated. Next, we asked the pre-test sample
of patients whether the items were understandable for
the study population. They were individually asked to
reformulate the items of the questionnaire. Patient’s
reformulation of the items and free comments on the
LupusQoL-FR were reviewed by the experts to ensure
that item content was correctly captured and for any
residual difficulty.
Construct validity
Structural validity. First, we conducted an exploratory
factor analysis (EFA) [principal factor analysis (PFA) fol-
lowed by oblique promax rotation]. We next studied item
loading on each factor after rotation, considering 0.5 as
significant loading. We computed an eight-factor item
loading matrix, according to the a priori hypothesis of an
eight-factor structure supported by the original LupusQoL
dimensionality. Kaiser’s criterion (eigenvalues > 1) [24],
Horn’s parallel analysis [25] and graphical analysis of the
screeplot [26] were also used to generate hypotheses
about the number of factors to be extracted.
Finally we performed a CFA using the UK structure of the
LupusQoL. In this analysis, each item was defined to rep-
resent only one domain, but the domains were allowed to
correlate with each other. Since item responses were not
normally distributed, unweighted least-squares estimation
was used. As stated in CFA reporting recommendations
[27], we reported (1) the ratio of 2 and degrees of freedom
(chisq/df); (2) an index to describe incremental fit, the
Comparative Fit Index (CFI); and (3) a residual-based
measure, the standardized root mean square residual
(SRMR). To indicate an acceptable fit, chisq/df, CFI and
SRMR should be 43, 50.95 and 40.08, respectively [28].
Rasch analysis. Analyses were conducted separately for
each LupusQoL domain. We used a Rasch-derived model
for polytomous items in which the distance between re-
sponse categories was constrained to be equal: the rating
scale model [29]. In the Rasch analysis, persons’ abilities
(quality of life) and difficulties of items are calculated on
the same linear scale expressed in log-odds units (logits)
[30–32]. Detailed methods for IRT analysis are available in
supplementary data, available at Rheumatology online.
Item fit was assessed using the infit and outfit statistics
for each item. Acceptable ranges were set according to
Smith’s recommendations [33], considering a sample size
of 182 patients, at 0.85–1.15 and 0.56–1.44 for infit and
outfit, respectively. Principal component analysis (PCA) of
the residual [34] and residual correlation matrix were
1908
examined to ensure unidimensionality and local independ-
ence of items.
External validity. Convergent validity was determined by
comparing the results of the LupusQoL-FR with those of
the SF-36, SLAQ and patient VAS using Spearman’s rank
correlation. Correlation coefficients >0.50, 0.35–0.50 and
<0.35 were considered strong, moderate and weak, re-
spectively. We hypothesized that the SLAQ score and pa-
tient VAS would correlate strongly with the LupusQoL-FR
domains. In particular, we expected better correlations
with domains related to physical activities than other do-
mains. Common person equating was performed by plot-
ting the person’s ability estimated by the Rasch model
from the LupusQoL-FR and the corresponding SF-36
Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
scale against each other. SF-36 Rasch estimates were
adjusted for the mean difference between SF-36 and
LupusQoL. The linear regression line should have a
slope close to 1 and an intercept close to 0 to demon-
strate that the scales are measuring close concepts [35].
Next, we compared LupusQoL-FR domain scores
between groups expected to have different HRQOL
using Mann–Whitney non-parametric tests (divergent
validity). We hypothesized that LupusQoL-FR domains
would be significantly altered in patients with active
disease with a SELENA-SLEDAI score cutoff of 4.
Reliability
Internal consistency
Cronbach’s a coefficient computed for each dimension
of the LupusQoL-FR was considered satisfactory when
5 0.7.
Test–retest reliability was assessed by computing
intra-class coefficients, comparing LupusQoL-FR do-
mains scores at baseline and 15 days later in patients
whose self-assessed quality of life on a seven-item
Likert scale was rated as no change.
Results
Among the 182 patients who agreed to participate,
complete data for all study visits were available for
162 patients. We included 160 women, most of whom
had inactive disease [SELENA-SLEDAI mean (S.D.) 2.6
(3.5), range 0–18] (Table 1). The patients were European
(Caucasian) (74%), Afro-Caribbean (9%), North African
(13%) and Asian (3%).
Cross-cultural adaptation process
Cultural adaptation in the translation process concerned
the wording of heading items: for grammatical reasons,
the French translation of ‘How often’ (A quelle fréquence)
was omitted from the items and reflected in the answer
choices. The translation of answer choices resulted in
choosing ‘assez souvent’ instead of ‘une bonne partie
du temps’ to represent ‘a good bit of the time’, because
the literal translation was considered to be too close to
adjacent response choices. The word lethargic (item 32),
for which the literal translation, lethargique, would not be
understood by many patients, was replaced by ‘ralenti’
www.rheumatology.oxfordjournals.org
 LupusQoL-FR in patients with SLE

(slowed). ‘I am unable to’ was initially translated to ‘je suis
incapable de’, which was considered too negative by the
pre-test patients. It was replaced by ‘je n’arrive pas à’
(I don’t manage) or ‘je ne peux pas’ (I can’t) in items 13
and 31.
Validity
Content validity
Both experts from the translation committee and patients
agreed that items comprehensively referred to relevant
aspects of SLE patients’ quality of life. Reformulation of
the 34 items by the pre-test sample of patients showed no
major distortion of the item’s meaning. The mail response
rate exceeded 90% at day 15 and 3 months. Patients’
Seventeen patients (9.3%) did not respond to at least
one LupusQoL-FR item at the baseline visit. Descriptive
statistics for LupusQoL-FR are presented in Table 2.
At least one dimension could not be scored because of
missing responses for three patients (1.7%). No significant
floor effect was observed, but ceiling effects occurred
ranging from 10.3% to 52.3%.
Construct validity
Structural validity. PFA resulted in eight first components
explaining 96% of the total variance (four with an eigen-
value > 1). Furthermore, parallel analysis led to the reten-
tion of an eight-factor structure. Interestingly, screeplot
analysis suggested a two-factor structure, accounting
for 76% of the total variance.

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
feedback was positive both in the pre-test group of
Rotation with eight factors resulted in items aggregating
patients and in written comments from the mail survey.
in a very similar way to the original LupusQoL as shown on
the rotated factor pattern (Table 3). Physical health no. 8
correlated with pain factor (loading 0.78 vs 0.02), planning
no. 2 had a higher correlation with physical health (loading
TABLE 1 Patients’ baseline characteristics (n = 182) 0.42 vs 0.34), fatigue no. 1 had a higher correlation with
emotional health (loading 0.30 vs 0.20), and burden to
Characteristic Mean (S.D.) N (%) others no. 3 had a higher correlation with emotional
health factor (loading 0.46 vs 0.35). The three remaining
Age, years 39.6 (10.5)
problematic items had a loading >0.3 but <0.5 with their
Female 160 (88)
own factor: physical health no. 7 (loading 0.45), pain no. 1
Disease duration, years 9.4 (7.5)
Antimalarial treatment 142 (78) (loading 0.47) and pain no. 3 (loading 0.3).
Daily prednisone use 125 (69) The Rasch fit statistics are shown in Table 3.
Immunosuppressive drugsa 62 (34) Three items underfitted according to both outfit and infit
SLAQ score 10.2 (7.4) statistics: physical health no. 8 (item 8), burden to others
Patient VAS 26.7 (26.9) no. 3 (item 19) and fatigue no. 1 (item 31). These
PGA 0.5 (0.7) items were also found to be problematic with factor
SLEDAI 5 4 57 (35) analysis. Category probability curves showed that the
Diploma (European equivalence)
answer choices were correctly ordered for all domains.
Entry level certificate or none 19 (11)
Certificate level 1 44 (25) PFA of residuals showed no major violation of the
Diploma level 3 31 (18) unidimensionality assumption. No significant local depen-
Diploma level 5 27 (15) dencies were found by the computed residual correl-
Diploma level six or more 50 (28) ation matrix. CFA showed an acceptable fit to the
UK-LupusQoL structure (chisq/df = 2.85, CFI = 1.00,
a
Excluding antimalarials and prednisone. SLAQ: maximum SRMR = 0.053). Standardized coefficients for all items in
activity = 48. their domains were >0.7. Applying various modification

TABLE 2 LupusQoL-FR descriptive data: missing responses, scores, floor and ceiling effects

Number of Number (%) Number (%)

LupusQoL missing responses Mean of patients with of patients with
domain (unable to score domain) (S.D.) Median, IQR minimum score maximum score

Physical health 0 77.4 (20.6) 81.3, 31.3 0 (0) 25 (13.7)

Pain 1 74.0 (24.3) 75.0, 41.7 1 (0.6) 49 (27.1)
Planning 1 80.7 (23.2) 91.7, 33.3 1 (0.6) 76 (42)
Intimate relationship 33a 82.2 (25.4) 100.0, 25.0 5 (3.4) 78 (52.3)
Burden to others 0 68.4 (26.6) 75.0, 33.3 4 (2.2) 28 (15.4)
Emotional health 0 73.1 (22.2) 77.1, 29.2 0 (0) 19 (10.4)
Body image 25a 81.1 (23.2) 90.0, 30.0 1 (0.6) 55 (35)
Fatigue 2 64.5 (26.1) 68.8, 43.8 2 (1.1) 23 (12.8)

a
Scores for 33 and 25 patients in the intimate relationship and body image domains, respectively, were not scored because
items were reported as non-applicable. IQR: interquartile range.

www.rheumatology.oxfordjournals.org 1909
1910
TABLE 3 Factor loading and Rasch fit statistics for the eight-factor structure LupusQoL-FR

Rotated factor patterna MNSQ Rasch fit

Hervé Devilliers et al.

Factor Factor Factor Factor Factor Factor Factor Factor

Items and dimensions 1 2 3 4 5 6 7 8 Infit Outfit

Physical health 1 I need help to do heavy physical jobs such as digging the 0.88 0.01 0.05 0.07 0.10 0.06 0.00 0.18 0.91 0.81
garden, painting and/or decorating, moving furniture
2 I need help to do moderate physical jobs, such as vacuuming, 0.92 0.10 0.05 0.01 0.00 0.02 0.05 0.03 0.88 0.71
ironing, shopping, cleaning the bathroom
3 I need help to do light physical jobs, such as cooking/preparing 0.57 0.24 0.03 0.06 0.11 0.01 0.05 0.24 1.41 1.39
meals, opening jars, dusting, combing my hair or attending to
personal hygiene
4 I am unable to perform everyday tasks, such as my job, 0.83 0.04 0.00 0.02 0.15 0.03 0.06 0.14 0.84 0.74
childcare, housework as well as I would like to
5 I have difficulty in climbing stairs 0.63 0.11 0.02 0.10 0.30 0.03 0.02 0.13 0.97 0.94
6 I have lost some independence and am reliant on others 0.70 0.05 0.06 0.07 0.09 0.14 0.01 0.20 0.69 0.66
7 I have to do things at a slower pace 0.45 0.17 0.02 0.18 0.14 0.11 0.04 0.02 0.78 0.8
8 My sleep pattern is disturbed 0.02 0.02 0.09 0.08 0.78 0.08 0.04 0.07 1.74 1.83
Pain 9 I am prevented from performing activities the way I would like to 0.34 0.08 0.03 0.12 0.47 0.03 0.01 0.05 0.7 0.67
because of pain due to lupus
10 The pain I experience interferes with the quality of my sleep 0.07 0.06 0.08 0.03 0.77 0.12 0.12 0.08 1.29 1.19
11 The pain due to my lupus is so severe that it limits my mobility 0.28 0.00 0.04 0.10 0.30 0.09 0.12 0.27 1.09 0.98
Planning 12 I avoid planning to attend events in the future 0.10 0.07 0.10 0.16 0.00 0.08 0.06 0.62 0.96 0.96
13 Because of the unpredictability of my lupus, I am unable to 0.42 0.16 0.05 0.10 0.15 0.08 0.07 0.34 1.27 1.21
organize my life efficiently
14 My lupus varies from day to day, which makes it difEcult for me 0.37 0.02 0.08 0.12 0.04 0.09 0.06 0.53 0.72 0.74
to commit myself to social arrangements
Intimate 15 Because of the pain I experience due to lupus, I am less 0.07 0.03 0.05 0.08 0.03 0.01 0.78 0.14 0.76 0.58
relationship interested in a sexual relationship
16 I am not interested in sex 0.00 0.09 0.03 0.04 0.08 0.00 0.85 0.05 0.99 0.97
Burden to 17 I am concerned that my lupus is stressful for those who are 0.06 0.11 0.04 0.09 0.10 0.81 0.03 0.06 0.63 0.64
others close to me
18 I am concerned that I cause worry to those who are close to me 0.06 0.05 0.04 0.06 0.11 0.82 0.00 0.01 0.76 0.74
19 I feel that I am a burden to my friends and/or family 0.17 0.46 0.05 0.06 0.16 0.35 0.08 0.05 1.63 1.58
Emotional 20 I have found my lupus makes me resentful 0.08 0.63 0.11 0.11 0.00 0.01 0.20 0.12 1.42 1.18
health 21 I have found my lupus makes me so fed up nothing can cheer 0.06 0.74 0.10 0.05 0.03 0.06 0.03 0.12 1 0.9
me up
22 I have found my lupus makes me sad 0.02 0.86 0.05 0.08 0.06 0.00 0.05 0.11 0.79 0.85
23 I have found my lupus makes me anxious 0.12 0.69 0.05 0.05 0.02 0.14 0.14 0.02 0.82 0.96
24 I have found my lupus makes me worried 0.15 0.78 0.04 0.04 0.04 0.09 0.09 0.03 0.55 0.54
25 I have found my lupus makes me lacking in self-confidence 0.07 0.79 0.01 0.18 0.09 0.00 0.04 0.05 1.51 1.44

www.rheumatology.oxfordjournals.org
(continued)

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
 LupusQoL-FR in patients with SLE

suggested by the EFA did not significantly improve the fit

Outfit
MNSQ Rasch fit

1.01

0.65

0.76
1.52
1.13

1.52
0.48

0.69
1.26
indices.
Rotation with two factors resulted in a first factor
(aggregated physical) comprising items from physical
health, pain, planning and fatigue dimensions. The

1.02

0.65

0.77
1.67
0.97

1.64
0.47

0.72
Infit

1.3
second factor (aggregated mental) comprised intimate
relationship, burden to others, emotional health and

Promax rotation after principal factor analysis. Loadings >0.5 are reported in bold characters. Infit statistics > 1.15 and outfit >1.5 are reported in bold characters.
body image. Seven of 34 items had a loading <0.5 with
Factor

0.12

0.05

0.07
0.09
0.23

0.06
0.02

0.05
0.15
both factors, but the loadings were stronger with the
8

factor containing the item’s original dimension (Table 4).

Item Rasch fit statistics of each of the two aggregated
Factor

0.10

0.03

0.02
0.05
0.15

0.01
0.04

0.04
0.15
domains were acceptable, with three items underfitting
7

with both statistics (physical health no. 3, body image

no. 4 and fatigue no. 3). The physical aggregated dimen-
Factor

0.01

0.01

0.06
0.13
0.09

0.05
0.07
0.04
0.14
sion was found to be unidimensional. The PCA residual-

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
6

based unidimensionality test was significant for the mental

Rotated factor patterna

aggregated domain (9.34% of t-tests comparing highest

Factor

0.04

0.07

0.03
0.10
0.08

0.15
0.10
0.02
0.09 positive loaded items vs highest negative loaded items
5

were significant, with 95% CI 5.1, 11.6 excluding 5%).

Significant local dependencies were found by computed
Factor

0.07

0.05

0.05
0.23
0.14

0.20
0.67
0.75
0.75

residual correlation matrix between seven and five pairs of

4

items in the physical and mental aggregated domains,

respectively.
Factor

0.58

0.79

0.89
0.58
0.57

0.18
0.02
0.00
0.05

External validity
3

Convergent validity. The LupusQoL-FR correlated strongly

with the corresponding SF-36 domains (Table 5). As ex-
Factor

0.22

0.15

0.07
0.09
0.05

0.30
0.16
0.14
0.19

pected, Spearman’s correlation with the SLAQ was high

2

for domains of the aggregated physical factor (r = 0.69,

0.77, 0.60, 0.75 and 0.69 for physical health, pain, planning,
Factor

0.05

0.03

0.15
0.21
0.03

0.19
0.18
0.03
0.06

fatigue and aggregated physical, respectively). Domains of

1

the mental aggregated factor also correlated significantly

but less strongly with the SLAQ (r = 0.43, 0.50, 0.62, 0.5 and
The hair loss I have experienced makes me feel less attractive

0.65 for intimate relationship, emotional health, burden to

My appearance (e.g. rash, weight gain/loss) makes me avoid

The weight gain I have experienced during treatment makes

others, body image and aggregated mental, respectively).

My physical appearance due to lupus interferes with my

Similarly, patient VAS of disease activity correlated more

Lupus-related skin rashes make me feel less attractive

closely with physical domains (r = 0.62, 0.65, 0.56 and

0.65 for physical health, pain, fatigue and aggregated
physical, respectively). Common person equating using
I cannot concentrate for long periods of time

patients’ ability Rasch estimates demonstrated that

SF-36 and LupusQoL-FR physical domains measured
a close concept, with a slope of the regression line of
I am often exhausted in the morning

0.91 (95% CI 0.80, 1.03) and intercept 0.27 (95% CI

Items and dimensions

0.14, 0.70). This was not the case for pain/bodily pain,
I feel worn out and sluggish

fatigue/vitality and emotional health/mental health.

I need to have early nights
me feel less attractive

Divergent validity. As expected, LupusQoL-FR domains

correlated weakly with the Physician Global Assessment
enjoyment of life

social situations

(Spearman r = 0.18 to 0.37). LupusQoL-FR was signifi-

cantly lower for physical health, pain, intimate relationship
and aggregated physical domains in active patients
(SELENA-SLEDAI 5 4, mean difference range for these
domains 7.2–8.1, Table 6).

Reliability
26

27

28
29
30

31
32
33
34

Cronbach’s a ranged from 0.85 (fatigue scale) to

TABLE 3 Continued

0.92 (emotional health scale), which indicates good

Body image

internal consistency. Aggregated physical and mental

domains had higher a (0.95 and 0.94, respectively).
Fatigue

Test–retest reliability was also good to excellent among

www.rheumatology.oxfordjournals.org 1911
Hervé Devilliers et al.

TABLE 4 Factor loading and Rasch fit statistics for the two-factor structure LupusQoL-FR

Rotated factor pattern MNSQ Rasch fit

Item
Factor 1 Factor 2 Infit Outfit

1 Physical 1 0.78 0.03 1.14 1.12

2 Physical 2 0.91 0.25 1.06 0.86
3 Physical 3 0.73 0.25 1.37 1.7
4 Physical 4 0.82 0.04 0.93 0.8
5 Physical 5 0.74 0.04 1.05 1.05
6 Physical 6 0.82 0.01 0.64 0.6
7 Physical 7 0.67 0.19 0.73 0.68
8 Physical 8 0.58 0.12 1.46 1.42
9 Pain 1 0.72 0.17 0.57 0.6
10 Pain 2 0.65 0.13 1.03 0.89

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
11 Pain 3 0.64 0.14 0.85 0.81
12 Planning 1 0.50 0.34 1.16 1.27
13 Planning 2 0.64 0.17 0.92 0.73
14 Planning 3 0.68 0.16 0.94 0.79
15 Intimate 1 0.28 0.40 1.42 1.4
16 Intimate 2 0.03 0.52 1.72 1.38
17 Burden 1 0.35 0.38 1.14 1.17
18 Burden 2 0.33 0.39 1.1 1.16
19 Burden 3 0.23 0.55 1.15 0.96
20 Emotional 1 0.03 0.69 0.92 0.82
21 Emotional 2 0.15 0.71 0.66 0.61
22 Emotional 3 0.14 0.70 0.54 0.56
23 Emotional 4 0.00 0.82 0.67 0.7
24 Emotional 5 0.00 0.86 0.53 0.63
25 Emotional 6 0.02 0.73 1.15 1.08
26 Body 1 0.07 0.67 1.12 0.84
27 Body 2 0.09 0.82 1.04 0.76
28 Body 3 0.19 0.77 1.37 1.34
29 Body 4 0.10 0.45 1.94 1.56
30 Body 5 0.03 0.58 1.24 1.13
31 Fatigue 1 0.34 0.46 1.41 1.34
32 Fatigue 2 0.52 0.34 0.7 0.78
33 Fatigue 3 0.36 0.24 1.55 1.59
34 Fatigue 4 0.45 0.33 1.13 1.1

Promax rotation after principal factor analysis. Loadings >0.5 are reported in bold characters. Infit stat-
istics > 1.15 and outfit > 1.5 are reported in bold characters.

99 patients who reported no change on a 7-point Likert
scale between baseline and day 15 (intra-class correlation
coefficient range 0.79–0.95).
Discussion
This study provides evidence that the LupusQoL-FR is a
valid tool to assess the quality of life of SLE patients.
Moreover, we showed that a two-factor structure could
be an interesting and complementary way to aggregate
LupusQoL-FR domains. This alternative item grouping is
conceptually different from the mental component sum-
mary (MCS) and physical component summary (PCS)
from SF-36 items described by Ware et al. [36] because
it was based on the items themselves rather than a higher-
order factor analysis on the dimensions. When we sought
an optimal number of factors to be extracted, we did not
formulate an a priori hypothesis of a two-factor structure.
This structure was suggested by statistical arguments and
was consistent with the combination of some of the pre-
viously described eight LupusQoL dimensions and with
the content of the item.
Structural validity was acceptable insofar as the factor
structure resulting from the EFA with eight imposed
factors was very close to the structure of the original
LupusQoL. Indeed, McElhone et al. [11] have already
reported high loading for item 8 (physical health no. 8) in
pain (0.51 vs 0.50 in the physical health component) and
for item 31 (fatigue no. 1) in emotional health (0.31 vs
0.37 in the fatigue component). These two findings were
also confirmed in the paper by Jolly et al. [14] about the
US version of the LupusQoL. Therefore these problems
are probably intrinsic to the LupusQoL rather than being
due to the cultural adaptation and should be addressed in
a future international study, especially as items 8 and
31 were found to exhibit a poor fit in the Rasch model.

1912 www.rheumatology.oxfordjournals.org
 LupusQoL-FR in patients with SLE

TABLE 5 Convergent validity of LupusQoL-FR

Correlation with the SF-36

Spearman’s Spearman’s
Spearman’s correlation correlation
LupusQoL domain SF-36 domain correlation with the SLAQ with patient’s VAS

Physical Physical 0.78 0.69 0.62

health function
Role physical 0.63
Pain Bodily pain 0.74 0.77 0.65
Planning Physical function 0.59 0.60 0.53
Social functioning 0.60
Intimate relationship Social functioning 0.45 0.43 0.45
Mental health 0.47
Burden to others Social functioning 0.63 0.50 0.42

Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
Emotional health Mental health 0.74 0.62 0.50
Role emotional 0.59
Body image Social functioning 0.55 0.50 0.38
Role emotional 0.54
Fatigue Vitality 0.72 0.69 0.56
Aggregated physical Physical function 0.73 0.75 0.66
Role physical 0.65
Aggregated mental Mental health 0.71 0.65 0.55
Role emotional 0.63

TABLE 6 External divergent validity

Disease activity according to SELENA-SLEDAI score

LupusQoL domain SLEDAI 4 4 (n = 108)a SLEDAI > 4 (n = 57)a Pb

Physical health 80.3 (19.1) 72.4 (23.6) 0.021

Pain 77.7 (23.1) 69.6 (25.9) 0.041
Planning 83.9 (20.4) 76.5 (26.7) 0.05
Intimate relationship 85.1 (25.2) 77.2 (27.1) 0.10
Burden to others 69.4 (25.3) 68.1 (28.8) 0.78
Emotional health 74.6 (20.9) 70.0 (24.2) 0.21
Body image 84.4 (20.6) 79.5 (24.3) 0.20
Fatigue 66.9 (25.2) 63.4 (27.2) 0.42
Physical aggregated 77.6 (19.2) 70.4 (22.6) 0.033
Mental aggregated 67.9 (16.9) 64.7 (18.8) 0.26

a
Mean domain score (S.D.); bMann–Whitney non-parametric test.

Item 8 (physical health) refers to sleep pattern, while
item 10 (pain dimension) refers to the extent that the
quality of sleep is affected by pain. It is therefore not sur-
prising that the two items were loaded together, which
could in part explain the overlap of the physical health
and pain dimension in the French, US and UK samples.
Item 11 (pain limits my mobility) refers to both physical
capacity and pain and was also highly loaded in physical
health in the UK sample (0.46 vs 0.49 in the pain dimen-
sion), which could explain its poor loading in the French
version. Items 13 and 19 had a loading <0.50 in their re-
spective dimension (planning and burden to others) and
higher loading in another dimension (physical health for
item 13 and emotional health for item 19), suggesting
that these items could have dual concepts referred to in
the same item in the French version.
An acceptable Rasch fit was found for both the eight
domain and two aggregated domain structures.
Test–retest reliability, internal consistency and conver-
gent/divergent validity were also satisfactory and similar
to previous LupusQoL studies considering that validation
against a known body image tool was not available in
either the US or French version.
Quality of life in our sample was much better than in the
US sample, maybe because the two samples differ in
terms of ethnicity, health care systems and disease
activity patterns. Interestingly, the pattern of domain
scores was nearly the same in the two cohorts (the

www.rheumatology.oxfordjournals.org 1913
Hervé Devilliers et al.
mean difference of all domain scores between French and
US patients was close to 30 on the 0–100 scale).
Compared with the UK sample, our patients had a
global impairment in LupusQoL domains, but the pattern
of domains was noticeably different. Emotional health was
comparable (mean difference between UK and French
scores 0.3), whereas the intimate relationship and physi-
cal health domains were disproportionally higher in the
French patients (mean difference 21.3 and 17.3,
respectively). In other domains the mean difference was
close to 10 on the 0–100 scale. A study of item differential
functioning in an international study could further clarify
this issue.
As a first limitation, our study recruited mostly outpa-
tients with inactive disease. This could therefore limit the
generalization of our results to active SLE patients.
Nonetheless, during the 8-month period between March
and November 2009, 15% of patients had a moderate/
severe SLE flare, which corresponds to the proportion of
active SLE patients seen in the national French care net-
work for lupus in clinical practice. Of note, we observed
more ceiling effects, especially in the intimate relationship
and planning scales, than in the UK and US studies.
Further, international studies should determine whether
this difference is related to disease activity or to ethnic,
social or cultural features.
Rasch fit statistics was somewhat better with outfit than
with infit statistics, with which we encountered mostly
overfit problems, which implies that items were less pro-
ductive for measurement and may produce misleadingly
good reliabilities and separations without degrading the
measurement system. Less restrictive published rules
[37] of 0.5–1.5 for infit and outfit cutoffs would result in
underfit of only four items in both eight- and two-factor
structures.
Conclusion
The LupusQoL-FR (supplementary data, available at
Rheumatology Online) is reliable and valid for the assess-
ment of SLE patients’ heath-related quality of life.
Perspectives for future research include testing of the cul-
tural invariance across international studies and confirm-
ation of the accuracy of the two-factor structures.
Rheumatology key messages
. The LupusQoL-FR is valid for use in the eight- or
two-scale format.
. The LupusQoL-FR shows good measurement
properties through Rasch analysis and classical
test theory.
Acknowledgements
The authors wants to thank the French national society of
internal medicine (SNFMI) and patient support associ-
ations LupusFrance and Max de Vie for financial support,
Mary Gawlicki (Corporate Translations, Inc.) for providing
1914
preliminary translation, Philip Bastable for manuscript
editing, Professor Jean-Francis Maillefert for his participa-
tion in the multidisciplinary consensus committee,
Professor Lee-suan Teh for her kind authorization to
work on the LupusQoL and the French East Internists
College (CIEST) for patient recruitment. A list of associate
investigators who recruited patients in the study can be
found as supplementary data, available at Rheumatology
Online.
Funding: This study was funded by the patient support
association LupusFrance through a tender launched
by the French National Society of Internal Medicine
(10 000E).
Disclosure statement: Z.A. has received consultant fees
Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
from GSK. All other authors have declared no conflicts
of interest.
Supplementary data
Supplementary data are available at Rheumatology
Online.
References
1 Wallace DJ, Hahn B, Dubois EL. Dubois’ lupus
erythematosus, 7th edn. Philadelphia: Lippincott Williams
& Wilkins, 2006.
2 Alarcón GS, McGwin G Jr, Uribe A et al. Systemic lupus
erythematosus in a multiethnic lupus cohort (lumina). XVII.
Predictors of self-reported health-related quality of life
early in the disease course. Arthritis Rheum 2004;51:
465–74.
3 Haupt M, Millen S, Janner M et al. Improvement of coping
abilities in patients with systemic lupus erythematosus: a
prospective study. Ann Rheum Dis 2005;64:1618–23.
4 Kuriya B, Gladman DD, Ibañez D, Urowitz MB. Quality
of life over time in patients with systemic lupus
erythematosus. Arthritis Rheum 2008;59:181–5.
5 Sanchez ML, McGwin G Jr, Durán S et al. Factors
predictive of overall health over the course of the disease
in patients with systemic lupus erythematosus from the
lumina cohort (LXII): use of the SF-6D. Clin Exp Rheumatol
2009;27:67–71.
6 Trupin L, Tonner MC, Yazdany J et al. The role of
neighborhood and individual socioeconomic status in
outcomes of systemic lupus erythematosus. J Rheumatol
2008;35:1782–8.
7 Lau CS, Mak A. The socioeconomic burden of SLE.
Nat Rev Rheumatol 2009;5:400–4.
8 McElhone K, Abbott J, Teh LS. A review of health related
quality of life in systemic lupus erythematosus. Lupus
2006;15:633–43.
9 Yee CS, McElhone K, Teh LS, Gordon C. Assessment
of disease activity and quality of life in systemic lupus
erythematosus—new aspects. Best Pract Res Clin
Rheumatol 2009;23:457–67.
10 Leong KP, Kong KO, Thong BY et al. Development and
preliminary validation of a systemic lupus
www.rheumatology.oxfordjournals.org

erythematosus-specific quality-of-life instrument
(SLEQOL). Rheumatology 2005;44:1267–76.
11 McElhone K, Abbott J, Shelmerdine J et al. Development
and validation of a disease-specific health-related quality
of life measure, the lupusqol, for adults with systemic
lupus erythematosus. Arthritis Rheum 2007;57:972–9.
12 Doward LC, McKenna SP, Whalley D et al. The
development of the L-QoL: a quality-of-life instrument
specific to systemic lupus erythematosus. Ann Rheum Dis
2009;68:196–200.
13 McElhone K, Castelino M, Abbott J et al. The LupusQoL
and associations with demographics and clinical
measurements in patients with systemic lupus
erythematosus. J Rheumatol 2010;37:2273–9.
14 Jolly M, Pickard AS, Wilke C et al. Lupus-specific health
outcome measure for us patients: the LupusQoL-US ver-
sion. Ann Rheum Dis 2010;69:29–33.
15 Jolly M, Pickard SA, Mikolaitis RA et al. LupusQoL-US
benchmarks for US patients with systemic lupus
erythematosus. J Rheumatol 2010;37:1828–33.
16 Hochberg MC. Updating the American College of
Rheumatology revised criteria for the classification of
systemic lupus erythematosus. Arthritis Rheum 1997;40:
1725.
17 Karlson EW, Daltroy LH, Rivest C et al. Validation of a
systemic lupus activity questionnaire (SLAQ) for popula-
tion studies. Lupus 2003;12:280–6.
18 Devilliers H, Guillemin F, Besancenot JF et al. Validation
transculturelle en langue française du systemic lupus
activity questionnaire. Rev Med Interne 2009;30:S376.
19 Leplège A, Ecosse E, Verdier A, Perneger TV. The French
SF-36 health survey: translation, cultural adaptation and
preliminary psychometric evaluation. J Clin Epidemiol
1998;51:1013–23.
20 Petri M, Kim MY, Kalunian KC et al. Combined oral
contraceptives in women with systemic lupus
erythematosus. N Engl J Med 2005;353:2550–8.
21 Beaton DE, Bombardier C, Guillemin F, Ferraz MB.
Guidelines for the process of cross-cultural adaptation of
self-report measures. Spine (Phila Pa 1976) 2000;25:
3186–91.
22 Guillemin F, Bombardier C, Beaton D. Cross-cultural
adaptation of health-related quality of life measures:
literature review and proposed guidelines. J Clin Epidemiol
1993;46:1417–32.
23 Mokkink LB, Terwee CB, Patrick DL et al. The COSMIN
study reached international consensus on taxonomy,
terminology, and definitions of measurement properties
www.rheumatology.oxfordjournals.org
LupusQoL-FR in patients with SLE
for health-related patient-reported outcomes. J Clin
Epidemiol 2010;63:737–45.
24 Kaiser H. The application of electronic computers to factor
analysis. Educ Psychol Meas 1960;20:141–51.
25 Horn JL. A rationale and test for the number of factors in
factor analysis. Psychometrika 1965;30:179–85.
26 Cattell RB, Vogelmann S. A comprehensive trial
of the scree and kg criteria for determining the
number of factors. Multivariate Behav Res 1977;12:
289–325.
27 Jackson DL, Gillaspy JA, Purc-Stephenson R. Reporting
practices in confirmatory factor analysis: an overview
and some recommendations. Psychol Methods 2009;14:
6–23.
Downloaded from http://rheumatology.oxfordjournals.org/ at East Tennessee State University on June 18, 2015
28 Schreiber JB, Nora A, Stage FK, Barlow EA, King J.
Reporting structural equation modeling and confirmatory
factor analysis results: a review. J Educ Res 2006;99:
323–38.
29 Andrich D. Application of a psychometric rating model to
ordered categories, which are scored with successive
integers. Appl Psychol Meas 1978;2:581–94.
30 Hays RD, Morales LS, Reise SP. Item response theory
and health outcomes measurement in the 21st century.
Med Care 2000;38:II28–42.
31 Rasch G. Probabilistic models for some intelligence
and attainment tests. Chicago: University of Chicago
Press, 1960.
32 Tennant A, Conaghan PG. The Rasch measurement model
in rheumatology: what is it and why use it? When should it
be applied, and what should one look for in a Rasch
paper? Arthritis Rheum 2007;57:1358–62.
33 Smith RM, Schumacker RE, Bush MJ. Using item mean
squares to evaluate fit to the Rasch model. J Outcome
Meas 1998;2:66–78.
34 Smith EV Jr. Detecting and evaluating the impact of
multidimensionality using item fit statistics and principal
component analysis of residuals. J Appl Meas 2002;3:
205–31.
35 Bond TG, Fox CM. Applying the Rasch model:
fundamental measurement in the human sciences,
2nd edn. Mahwah, NJ: Lawrence Erlbaum Associates,
2007.
36 Ware JE, Kosinski M, Keller SD. SF-36 physical and
mental summary scales: a user’s manual. Boston, MA:
Health Institute, New England Medical Center, 1996.
37 Wright BD, Linacre JM. Reasonable mean-square fit
values. Rasch Meas Trans 1994;8:370.
1915